Variable Current Control for DC Power Supply?

[Dave Hederich]

Can anyone tell me what is required to add variable current control to a 5V/20A DC power supply?

The challenge here is to build a 4.2V/0-20A power supply for charging Li (Li-ion and Li-poly) cells in parallel.

The backround is that many people (including myself) are successfully using variable voltage/current power supplies to charge series Li packs, as they act as CCCV chargers. Parallel Li charging is considered safer by some because a fixed voltage 4.2V charger cannot be set to exceed the critical 4.2V charging limit for a single (or parallel) Li cell(s). In series Li charging, when cells get out of balance, individual cells can be charged past 4.2V and suffer catastrophic failure, venting with flame.

An internet search shows that 5V/20A DC power supplies are readily available at low cost.

A key component that is not yet available is a 1-cell version of the Kokam Safety Guard, a device that limits charging voltage to Li cells. Kokam currently makes 2-cell (8.4V), 3-cell (12.6V) and 4-cell (16.8V) SGs. We are trying to convince Kokam to add a 4.2V SG to the lineup for single-cell or parallel charging. All the other SGs are rated to a max of 20A, so that's why the 20A limit has been selected for the power supply.

The other missing component is the variable current control that would allow the 20A power supply to function from 0-20A for different size Li cells to be charged at 1C. I'm thinking of something similar to the control knob on the Astro 109 Li charger that goes from 50mA-8A. I assume this controls a pot of some kind.

And that brings us to the final issue. I'm electronics-challenged. I'm just an idea guy here, and not a technical guy. I proposed, based on ~$30 for a 5V/20A power supply and a $13 Kokam Safety Guard (assuming a 4.2V version would be offered), that this 4.2V/0-20A Li charger could be built for <$75. That assumes a variable current control for this device could be added for ~$30.

Am I hallucinating, or am I on to something here?

[IanS]

Hi Dave,

I like to propose a slightly different version of this one. Instead of using a separate 5V/20A PS why do not use the a computer PSU. Many of us are using these to power our chargers at home. They have a very well build and regulated 5V output that is usually only partially loaded (5 -10 % with a resistor or an auto bulb) in order to use the 12V output and the remaining 90-95% of available current (10 - 15 or more Amperes) is just sitting there waiting for something useful....

The current regulated 0-20 A/4.2V is doable but probably will be expensive for DIY.

Why do not use any single cell dedicated ICs (available from any major IC manufacturer) and build as many single cell chargers as you like. At 5V input you can get away with a linear solution - for example a single cell 2A charger based on TI bQ2057CSN ($3 @ Digikey) will dissipate max 6W at the beginning of the charge process.

Best Regards,
IanS

[Dave Hederich]

Thanks, IanS. I had thought about computer PSUs being an option for the power supply. The 5V rail on many PSUs should be capable of providing >20A.

As for the dedicated single cell ICs, there are already many people working on these lower current solutions for single cell charging. Some people are more interested in parallel charging many cells at higher current levels. That's what this proejct is all about.

It's the 0-20A current regulator that is the biggest question mark (assuming Kokam will expand its Safety Guard lineup to include a 4.2V version). Anything anyone can tell me about the current regulator will be most appreciated.

[Dave Hederich]

Well, this is disappointing. When I didn't get a response to this question in the Batteries and Chargers forum, I thought surely someone in the DIY Electronics forum would know something about adjustable current controls. If nothing develops here, I guess I'll try Modeling Science next.

[Comatose]

Dave, it depends on how you're looking to do it. It would be really simple to make a controller that looks like a buck converter, but current controlled instead of voltage controlled. A buck regulator IC that's for adjustable voltage would do the trick, if you could massage the current sense resistor's output to feed the controller what its looking for. An op amp filter network would do. Alternatively, it might be easier to run a PIC and do the control in software. It depends mainly on what input and outputs you need from the device.

This seems like the hard way to go about it, though. The functionality really shouldn't be separate, as the same structure with a difference in feedback can provide current limiting and voltage limiting. If what you really want is a CCCV charger that only outputs 4.2V max at a user settable charge rate, that's easy enough to do.

But, then what you're basically making is a charger. At that point, why not just have the circuit accept a 12V input and spit out a CCCV waveform that peaks at a user definable current and 4.2V. The 12V input is technically no harder (easier, in fact) and 100W from a 12V supply is only 8A.

Assuming a run of 100 of these were made, a board level product was acceptable and the only user definable variable was the charge current (maybe with a three digit LED displaying the current) they could be sold retail for between $40 and $50. A larger run could be made more cheaply, of course. This is a 5 minute back of the envelope calculation, but it'd be in that range.

[Dave Hederich]

Now we're getting somewhere. A little technical for me, though. What you've laid out is a formula for someone with a lot more electronics experience than I have to build from scratch. RD Blakeslee has been trying to talk someone into offering something like this for a long time with no success. What I was hoping for was to be able to point to existing components that could be strung together to create a 4.2V/0-20A CCCV Li charger.

I take it you know of no variable current controls just sitting around for purchase that could be plugged into a 5V/20A power supply and a 4.2V Kokam Safety Guard (presuming we can convince Kokam to offer one).

[RD Blakeslee]

... what you're basically making is a charger. At that point, why not just have the circuit accept a 12V input and spit out a CCCV waveform that peaks at a user definable current and 4.2V. The 12V input is technically no harder (easier, in fact) and 100W from a 12V supply is only 8A.

Assuming a run of 100 of these were made, a board level product was acceptable and the only user definable variable was the charge current (maybe with a three digit LED displaying the current) they could be sold retail for between $40 and $50. A larger run could be made more cheaply, of course. This is a 5 minute back of the envelope calculation, but it'd be in that range.

Comatose (anybody), do you have any knowledge of what outfit might be interested in manufacturing a run of 100 of these?

- RD

[Andy W]

Well, this is disappointing.

Now we're getting somewhere. A little technical for me, though.

The problem is that you're looking to adapt something to a function that it was not designed for. Having current controls in the voltage regulating side would be much simpler than trying to add them after the fact.

As stated, you're trying to build a charger, which would be MUCH simpler than trying to build a CC regulator which can modify a previously regulated output voltage.
..a

[Dave Hederich]

The problem is that you're looking to adapt something to a function that it was not designed for. Having current controls in the voltage regulating side would be much simpler than trying to add them after the fact.

As stated, you're trying to build a charger, which would be MUCH simpler than trying to build a CC regulator which can modify a previously regulated output voltage.
..a

Well, I figured my lack of detailed understanding of electronics would create problems! Variable voltage/current power supplies have already proven to work well as Li-ion/LiPo chargers. But higher current models get expensive fast, with $250 being about the minimum on a 0-20A model. With parallel charging, you really don't need variable voltage. In fact, a fixed voltage of 4.2V is desirable from a safety standpoint.

Maybe I am coming at this from the wrong way. Maybe the way I should present this to those who really understand electronics is, what are the options for creating a fixed voltage 4.2V power supply with variable current from 0-20A that is safe, not too difficult to put together and represents significant cost savings over a $250 0-20A/0-30V power supply?

Is it reasonable to expect that an electronics novice can find modules to piece together, or does it require advanced electronics work? If advanced, is it reasonable to expect that someone with good knowledge could design one that could be built from a schematic at reasonable parts cost? Or, is it possible that someone would be interested in building these for a fairly small group of Li users who want to be able to charge parallel cells at higher current up to 20A without spending $250?

[hwhall]

> what are the options for creating a fixed voltage 4.2V power supply with variable current from 0-20A <

Perhaps I'm misreading the intent here, but thinking of Ohm's Law, I'm not sure you can have a fixed voltage across the charging battery and also have a fixed current. At any instant in a charging process, a battery represents some value of resistance, and the current through the battery is controlled at that instant by the voltage across that resistance. Perhaps what I'm missing is that the current you are thinking of is an average current, like a pulsed charge, so that while the instantaneous current is still ruled by Ohm's Law, the average current is controlled by the ratio of pulse on & off times? --Wayne

[dkselw]

>Perhaps I'm misreading the intent here, but thinking of Ohm's Law, I'm not sure you can have a fixed voltage across the charging battery and also have a fixed current.
.
.
.

CCCV chargers have a current limit and a voltage limit.
Initially the current will be be limiting factor and current will be regulated until the voltage limit is reached. Once the voltage limit is reached the voltage will be regulated.
It is not trying to regulate current and voltage at the same time.

[Dave Hederich]

Wayne, not sure if you know, but here's how a variable voltage/current power supply charges Li cells:

If you want to charge a 3-cell series pack of, say, 1000mAh Li cells at 1C, you set the voltage on the power supply to 12.6V and the current to 0.0A. You connect the pack to the power supply and turn the current adjustment knob up to 1.0A. The voltage immediately drops on the power supply as it starts charging the cells, and gradually climbs back up to 12.6V. This is the constant current portion of a CCCV charge.

After the voltage hits 12.6V, it remains there as the current starts falling from 1.0A to 0.0A. This is the constant voltage part of the CCCV charge. When the voltage readout is 12.6V and the current readout is 0.0A, the pack is 100% charged.

This is the process that I want to replicate on a 4.2V fixed voltage power supply for charging parallel Li cells.

[jeffs555]

I would be very afraid of promoting a homebuilt 20Amp 4.2volt power supply as a charger for assembly and general use by novices. A simple current and voltage limited supply is missing one of the most important safety features of a proper LiPoly charger. Any proper LiPoly charger will check the initial cell voltage, and if the cell voltage is below 3 volts either refuse to charge or apply a very low trickle charge. If you always check the voltage manually before charging, it could be safe, but I can just picture the massive fireball when someone who doesn't know better connects 10 or more deeply discharged cells to a 20 amp supply.

[Dave Hederich]

Good point. Most of the people who are looking at parallel charging right now are fairly savvy about measuring cell voltage. But if parallel charging became popular, the less savvy would want to jump in without understanding all the implications. For general consumption you'd want a well-engineered device with safety mechanisms in place. I'd like to see RD Blakeslee comment on these issues.

[Comatose]

Comatose (anybody), do you have any knowledge of what outfit might be interested in manufacturing a run of 100 of these?

- RD

Well, since you asked...

A friend of mine and I from college are setting up a short run circuit shop doing things like this. Power and motion control modules for the student/hobbyist/research market. We're producing in-house (we have a stencil printer, pick place and a reflow oven) and this is relatively up our alley, so if the demand was there I could see us doing this. The current active development pipeline is for a switch-mode 7805 drop-in replacement module first (6-35V in, 5V 1.25A out max continuous) the a parkflyer bec based around the same, then a 1.5-13V adjustable version of the same, then a pair of universal step-up or step down converters (3-30V in, 3-24V out) in 2A and 5A max (greater of input and output currents, depending on if you're stepping up or down) with LED voltage/current display. So I'd certainly be willing to make a run of 1 cell high current chargers if the demand was there. If demand was less than 100 the cost of a solder stencil and PCBs gets pretty high on a per unit basis.

If i was doing it, I think the minimum protections that it could be sold with (as a charger not a power supply) in good conscience would have to be overvoltage (in case someone gets funny ideas about that 3S pack they bought) undervoltage, reverse pack voltage, reverse input voltage, short circuit, overcurrent and a timeout (at, say 2 hours). Anyone have some to add to that?

[Terry S]

I just seen this thread and its something I have planned for the summer. I intend to convert a pc psu for use as a lipo charger. My first thoughts were to adjust the psu to 4.2v putput and then add a current limit circit but the current limiter loosees some voltage even when full current is asked for so this is not possable. Now my plan is to keep the psu at 5v or even tweek it up a bit if nessasary and use voltage regulators or transistors as current limiters for each of my 12 outputs. As you can see I am not charging in parralell but creating 12 single cell chargers. I will follow this thread closely and see if anyone comes up with a better low cost solution before I start building mine !

Terry

[hwhall]

dkselw & Dave, thanks for the info! I figured there just had to be something that I didn't understand about the problem. --Wayne

[GlowFly]

My first thoughts were to adjust the psu to 4.2v putput and then add a current limit circit but the current limiter loosees some voltage even when full current is asked for so this is not possable.
This isn't a problem. If you're prepared to break into the PSU and insert your current sensing resistor inside the 5v feedback path, you can still have your CCCV characteristic. With a little more care you can even add continuously variable current control, AND trickle charge at 3v and below. It's probably still simpler than building something from scratch, but does require a little more knowledge. Also each PSU is slightly different in the design of the regulation feedback.
--
Steve

[navkan]

Back at the beginning of this thread, IanS proposed a solution using commercial low cost battery management ICs. From my limited investigation they will handle the trickle charge, CC and CV requirements and are designed for charging Li-Ion and Li-Poly voltages. Although they probably can't control 20A, a multiple setup of 2-4A chargers might be an answer, right?

[Dave Hederich]

Although they probably can't control 20A, a multiple setup of 2-4A chargers might be an answer, right?

I think that's an answer to single cell charging, which several commercial sources are working on right now.

The purpose of this thread is to generate interest in parallel charging by making available at a reasonable cost the equipment required to parallel charge Li cells at 4.2V and high current.

[Terry S]

Thanks glowfly I haddnt thought of that, I will give it some thought though but it may be harder than it first appears we will see. I was just thinking of the easy and cheap way of doing the job, I always take the path of least resistance (no pun intended).

Terry

[Art Horne]

Comatose,

Greetings. I may be needing a small production run for the charger described in this link: http://www.rcgroups.com/forums/showthread.php?t=269638&page=1&pp=15
Drop me an email at ahorne1 AT comcast DOT net if you are interested.

Thanks,
Art Horne

[Andy W]

CCCV chargers have a current limit and a voltage limit.
Initially the current will be be limiting factor and current will be regulated until the voltage limit is reached. Once the voltage limit is reached the voltage will be regulated.
It is not trying to regulate current and voltage at the same time.

Some people feel that the PSU will be "forcing" that amount of current and voltage thru the load (in this case, the cell). In fact, current regulation is more about "limiting"..
..a

[flieslikeabeagle]

Dave, maybe I'm misunderstanding something, but I am wondering if building a 20A, 4.2V current-limited PSU is in fact a safe way to charge multiple cells in parallel. It seems to me that if one or two of the paralleled cells are more deeply discharged than others, those cells can draw excessive amounts of current and thus be damaged, while the total current still remains within the limits of the CC device. For instance, if you put ten 1200 mAh cells in parallel and set your charger to 12 amps, what if one of those cells draws 5 A while the other nine together draw the remaining 7 A?

It seems to me that the only foolproof way to use such a charger would be to start out with the current limited to 1C for a single cell - in our example, 1.2 A. After running at this low current long enough to guarantee all the cells were balanced, the current limit could be tweaked up to 12 A.

I think building a number of cheap, single-cell chargers may be a safer approach. This approach guarantees that no cell will get more than 1C current, under any circumstances.

Up till now I have not used the parallel charge/series discharge approach (I bought 2S battery packs with no taps), but I really like the idea and have been thinking of ways to implement it (or something equally good, i.e. balancers and taps for series charging) with my next set of packs.

I have been wracking my brain for a good way to do this. One idea floating around in my thoughts is to buy cheap single-cell surplus cellphone chargers, and use one per cell. It might even be possible to connect several of their outputs in series (I think some of them have floating outputs) and charge a pack with taps in this way.

At the suggestion of another RC Groups member, I checked a local 99-cent store and found some "Qualcomm Rapid Charger" cellphone chargers, model GXM 978G, and some Belkin "mini folding travel charger" cellphone chargers, for 99 cents apiece (what else? :) ).

Don't rush out to buy these just yet, though, because there's a BIG problem. Read on...

The first of these (Qualcomm) is designed to run of +12 V (cigarette-lighter plug) and the wording on the package suggests it puts out a full ampere, though I haven't measured it. The second (Belkin) comes with a wall-wart and says nothing about fast-charging, so I suspect it puts out very little current.

Here's the catch: I measured both chargers, and they put out well beyond 4.4V, in other words, if connected to an unprotected Lipo cell, these chargers would create an extreme likelihood of a destroyed Lipo cell, and/or a possible Lipo fire. :eek:

Sadly, the RC Groups member who first informed me of the availability of the Belkin chargers in the 99 cent store damaged some expensive Lipos by trying to use these chargers. So my warning statement above is not just hypothetical. Anyone reading this: DO NOT try to use these chargers on unprotected Lipo cells as-is.

I can't help wondering, though, if tweaking a single resistor might not set the voltage to a safe level...these are simply single-chip switching regulators, as far as I can tell. I have not had time to put into researching this, but it seems like a distinct possibility. If this works, it might be the ultimate economical, safe, Lipo charging solution: multiple single-cell, CC/CV, 1A chargers for ninety nine cents apiece, running off a large cheap 12V power supply (computer power supply, for instance).

I know my way around linear voltage regulators, but know little about switching voltage regulators. Anyone who knows more than I do care to offer suggestions? I'd be happy to crack open one of my 99c chargers and post any available chip part numbers, etc.

-Flieslikeabeagle

[flieslikeabeagle]

Andy W wrote:

Some people feel that the PSU will be "forcing" that amount of current and voltage thru the load (in this case, the cell). In fact, current regulation is more about "limiting"..

Exactly!

Another way of looking at it is that Ohm's law says current = voltage/resistance. When unloaded or lightly loaded, a CC/CV charger puts out a constant voltage and allows the cell to draw whatever current Ohm's law dictates, based on the cells resistance, i.e. state of charge. However if the cell has too low a resistance (i.e., it is discharged), it will try to draw too much current, and the CC/CV charger says "Oh no, you don't!" and drops its output voltage in such a way that the current is limited to a safe value. Ohm's law still applies - but the available voltage is decreased, thus keeping the currrent to a safe value despite the lower battery resistance.

Note that in CC mode, the charger *drops* its available voltage. It does not try to force extra current through the cell, it actually holds back the cell from drawing too much current.

-Flieslikeabeagle

[Dave Hederich]

I'll let RD Blakeslee defend the safety of parallel charging when he returns to this thread. I'm just trying to figure out what possibilities exist for a parallel charger.

As I understand those surplus Li cell phone chargers, the only part you're buying is the 5V power supply. They are intended to interface with the circuit board on the battery pack designed specifically to work with that charger. The circuit board reduces the 5V from the power supply to 4.2V for proper charging of the Li cell in the pack. So, yes, when you directly hook up a Li cell without a circuit board to a 5V power supply, you will have a major problem.

If Kokam would ever get around to producing a 1-cell, 4.2V Safety Guard, they could be used with the surplus cell phone chargers to limit charging current to 4.2V.

[flieslikeabeagle]

Dave, just to clarify, I think parallel charging done right is definitely safer than the prevailing "no taps, no balancers, series charging" scheme.

I do suspect that care has to be taken to ensure all the cells are in a similar state of charge when paralleled, otherwise some will draw more current than others, possibly to a dangerous degree. And it seems to me that using a dedicated CC/CV charger per cell is the safest solution of all, since it requires almost nothing from the user, beyond ensuring the polarity isn't reversed when the connectors are wired up.

I'll have to do some measuring on those surplus cellphone chargers to see what's going on. The open circuit voltage I measured from those surplus chargers was not 5V, but more like 4.4 - 4.5 V, IIRC. My thought is that it may only take a single resistor change to cut that to 4.2V - usually all you have to do to change the voltage from a regulator chip is vary the feedback amount by changing one resistor in a potential divider...

I do not know if current limiting is built into these switching regulators or not. I definitely need to crack one open to see whats in there, and make some measurements on what it does. If I learn anything useful, I will post the results.

While Kokam does not yet make 1-cell charge guards, I think Dan Baldwin's Lipo balancer circuit does exactly the same thing. The only catch is that you have to build it yourself, not hard if you are familiar with electronics construction, but daunting otherwise.

By the way, if you think this is not the thread for this topic, I will gladly start a different one. Just let me know.

-Flieslikeabeagle

[Comatose]

From a purely cost standpoint its far cheaper to make one big charger than half a dozen little ones. The major cost drivers in making a charger like this are the connectors, the interface, the controller chip, the inductor, the switches, and the capacitors. In a single charger, you need one controller, one interface, one set of connectors. You might have to go with parallel switching FETs, (for 20A I'd probably use two) and more expensive inductors and capacitors, but the cost increase to go from 5A to 20A is likely to be under 50%. This is especially true if the charger is running from a microcontroller and has an LCD panel style interface. Going from 1 to four chargers is probably a 3x increase in cost (not 4x because when volume goes up, prices come down)

Surplus chargers can probably be modified to do the trick, but unless there's a reliable source, they're unlikely to make a major impact. For example, the tutorials for the cdrom retrofits existed for a year or so before they really took off, which was largely due to the gobrushless kits which ensured a reliable and uniform supply. With the CDrom kits, a realiable supply chain exists because the motors are still in production. With surplus cell phone chargers, a reliable surplus source seems unlikely at the deeply discounted rates needed to make them worth considering.

[flieslikeabeagle]

Comatose wrote:

Surplus chargers can probably be modified to do the trick, but unless there's a reliable source, they're unlikely to make a major impact.

Quite true for the market at large, but something along these lines will make a major financial impact on any individual hobbyists who get chargers for 99 cents apiece :)

At present I charge all my 2S packs with older surplus Qualcomm Li-ion cellphone chargers that cost me $4.50 each. Those chargers are CC/CV, regulated to 300 mA and about 8.3V respectively. It takes a few hours to charge my big 2000 mAh pack, but for $4.50 I have no complaints. And since the chargers are cheap, I own three, one for each 2S Lipo pack I have...so they all charge simultaneously and I don't have to wait for one to finish to start charging the next.

I will take a look inside one of the 99c surplus cellphone chargers; I think there is a good chance I'll find a generic switching power supply IC in there. There may be no sustainable source of the assembled chargers, but there may well be a sustainable source for the components in one. Not for 99 cents, though, unfortunately.

-Flieslikeabeagle

[rpage53]

I do suspect that care has to be taken to ensure all the cells are in a similar state of charge when paralleled, otherwise some will draw more currents than others, possibly to a dangerous degree.
When I tried to repeatedly point this out last year, I got loudly shouted down. I don't see a 20A 10 cell parallel charger as a foolproof method without a lot of protective circuits that effectively turn it into 10 single cell chargers fed from a common source.
I am particularly concerned about a cell that was failing with a short circuit (extremely low resistance and low voltage) that would suck most of the 20A and blow.

The current limiting, voltage limiting 1 cell circuits are available for free from surplus battery packs.

Rick.

[RD Blakeslee]

Dave et al, sorry to get back here late. Some series-charging guy put a gremlin in my CPU, but The Force is with the parallel universe, so I'm back.

First, I'm no more savvy about things electronic than Dave. We're in dire need (at least, I am) of instruction, and thank you guys here for it.

The problem presented by low cell voltage(s) and/or extremely different voltages cell-to-cell when parallel charging hasn't caused trouble for me, simply because I don't discharge LiPo cells below 3 volts, to avoid cell damage, and I check current flow to each cell when I start the charge, in case something's wrong in one or more charging branches, so to speak. Also, I apply my high-tech FINGERONEM temperature probe from time to time during the charge.

But I can see where an ordinary CCVV would be dangerous for novices who don't have any electrical background (and resultant common sense) about charging batteries.

So, We'll have to have "trickle in parallel up front", I guess. How much extra complexity and cost wth that?

The single shorted cell Rick has mentioned becomes obvious as soon as it's connected in parallel with any other(s). Such a cell in a SERIES CHARGED pack is not self-evident and effectively converts a nominal XS pack into an (X-1)S pack.

- RD

[Marv829]

For parallel charging of multiple single cells I would think you should be able to take a design like this charger (http://cgi.ebay.com/ws/eBayISAPI.dll?ViewItem&category=2563&item=5943466178&rd=1&ssPageName=WDVW) and just rework the resistor values for a 4.2v output. You can then just add another kit(modual) to how many you want as long as your power supply has the current capacity to support/supply several moduals. I belive this kit uses the LM317 voltage regulator for both the voltage and current regulation.The LM317 is capable of being used from 1.2v - 37v and has a current capacity of 1.5A so that would be the limit for each modual as far as charging current goes. However the LM317 in its K package(TO-3) has rating of 2.2A with a max of 3.4A when it's input to output voltage ratio is < 15v, so that opens you up to some big single cell packs at a 1C charge rate. When the LM317 is used as a current regulator you coud use a POT for output adjustments but I would feel safer using a fixed values set via a jumper setting. Here is a link to it's spec sheet in PDF form - LM317 (http://www.national.com/ds/LM/LM317.pdf)

[Dave Hederich]

I think all of the discussion in this thread is on topic. As with other similar discussions, what we're all looking for is a simple, inexpensive solution for safely charging Li cells used in RC applications.

Parallel and single-cell Li charging have both been widely acknowledged as safer overall than series charging. But RC vendors continue to use series charging of Li cells as the industry standard. Some people prefer the concept of single-cell charging to parallel and vice versa. Each has its pros and cons.

I actually lean toward single-cell charging as the safest overall solution. There appear to be good efforts to bring single-cell charging systems to market, and these will be interesting to track. But I think parallel also deserves its fair hearing. No one should be loudly shouted down, because it's important to get good input from all those who have knowledge about various Li charging characteristics.

I don't think there will be a shortage of cheap, surplus Li cell phone chargers for many years. The cell phone industry has settled on single-cell Li packs as the standard for the industry. Different cell phone vendors use slightly different chargers for different models. So every time an old cell phone model is replaced with a new one, there's a good chance that the single-cell charger dedicated to that model will be replaced with a new design, and the old charger will be dumped on the surplus market.

The problem we always run into when discussing this topic is that there are such widely varying needs. At the low end, some small RC aircraft run on a single, tiny Li cell with capacities of <100mAh. At the high end, some of the larger, high-performance aircraft have XSXP packs where X can be a dozen or more, and cell capacities can run into thousands of mAh. A universal charging solution is elusive, to say the least.

This is further complicated by the fact that some people are only interested in factory-engineered solutions that can be purchased as plug and play from vendors, while others are interested in building their own. Taking all of this into consideration, it's no wonder that there's never universal agreement when we have these discussions.

DIY folks will continue to do whatever they want, as designing and/or building the projects is part of the entertainment for them. The commercial market will design, build and market the systems that they think make the best business sense. Otherwise, they would go out of business. The commercial market appears to be warming up to single-cell Li charging, but not parallel Li charging. By present trends, it appears that parallel Li charging will remain DIY.

Since RD Blakeslee has been the leading proponent of parallel charging, this subject is not going to disappear unless RD has a change of heart. This is good, because until one system is proven to be safer and more effective than others, we should keep all options open.

RD, please let us know if you ever decide that single-cell Li charging is ultimately safer for more people than parallel charging, and more likely to replace series charging. From what we've heard so far from many people, parallel charging appears to have an edge in lower cost while single-cell charging appears to have an edge in overall safety.

[Dan Baldwin]

Dave, just to clarify, I think parallel charging done right is definitely safer than the prevailing "no taps, no balancers, series charging" scheme.

I do suspect that care has to be taken to ensure all the cells are in a similar state of charge when paralleled, otherwise some will draw more currents than others, possibly to a dangerous degree. And it seems to me that using a dedicated CC/CV charger per cell is the safest solution of all, since it requires almost nothing from the user, beyond ensuring the polarity isn't reversed when the connectors are wired up.

I'll have to do some measuring on those surplus cellphone chargers to see what's going on. The open circuit voltage I measured from those surplus chargers was not 5V, but more like 4.4 - 4.5 V, IIRC. My thought is that it may only take a single resistor change to cut that to 4.2V - usually all you have to do to change the voltage from a regulator chip is vary the feedback amount by changing one resistor in a potential divider...

I do not know if current limiting is built into these switching regulators or not. I definitely need to crack one open to see whats in there, and make some measurements on what it does. If I learn anything useful, I will post the results.

While Kokam does not yet make 1-cell charge guards, I think Dan Baldwin's Lipo balancer circuit does exactly the same thing. The only catch is that you have to build it yourself, not hard if you are familiar with electronics construction, but daunting otherwise.

By the way, if you think this is not the thread for this topic, I will gladly start a different one. Just let me know.

-Flieslikeabeagle

I don't have an FMA charge guard, so I'm not positive about this, but I don't think they work like a balancer. As I understand it, the charge guard just ends the charge if the voltage on the pack gets over 4.2 volts/cell. A balancer will shunt away however much current it needs to to hold the cell to 4.2 volts.

One problem with using a charge guard with a constant current source as a charger is that you lose the constant voltage portion of the charge, so the pack would not be fully charged.

I'll see if I can find any of those chargers at a 99 cent store. It sounds like something fun to play with.

Dan

[Art Horne]

Greetings,

I've beem following this thread for a while for obvious reasons and I feel i need to bring one very important caution to the forefront. From the postings I've read here so far I get the impression (I may be wrong) that some of you are intending to have multiple 4.2v chargers, all powered by a single hi current power source, plugged into a series wired pack through the taps to only charge each cell to 4.2v. Now this will work ONLY if there is no common ground between each charger circuit. If you don't have some form of transformer isolation between each unit you will instantly short out one or more cells in the serial connected battery. If the cells being charged are completely disconnected from each other then you have no problem. The DIY simple chargers have no isolation and cannot be used to charge multiple series connected cells simultaneously unless each charger has it's own isolated power source. Multiple DIY chargers all powered from a single DC source would all have thier negative terminals tied common to the negative side of the supply and therefore would cause shorts when plugged into the battery. I hope that this might save someone from an expensive and dangerous mistake.

Good luck with it.
Art

[flieslikeabeagle]

Art, thanks for the clarification. I was the one who mentioned the possibility of perhaps using some cellphone chargers in series to charge packs with taps - and I did mention that the chargers needed to be a "floating" design - but thank you for spelling out exactly what that means to anyone who is unsure.

-Flieslikeabeagle

[flieslikeabeagle]

Dan, if I understand you, you're saying the FMA guards are series devices rather than shunt ones? In other words, they go to high resistance when the voltage across the cell exceeds the safety threshold?

One problem with using a charge guard with a constant current source as a charger is that you lose the constant voltage portion of the charge, so the pack would not be fully charged.

I'm having difficulty seeing this. Since the charge balancers simply limit the voltage across the cell, and have no "foldback" or "crowbar" action, it seems to me the cells end up with the full 4.2V being held across them indefinitely...and that means they will charge fully, given enough time. In fact, it seems to me that once the charge balancers kick in, they actually turn the constant current source into a constant voltage source - think of the Thevenin equivalent circuit, a current source dumping into a device tht clamps the load terminal voltage to a constant value.

I'll see if I can find any of those chargers at a 99 cent store. It sounds like something fun to play with.

I was hoping you'd say something like that :) . I have no experience with switching regulators myself. Here's a picture to assist you in the hunt for the 99-cent chargers. The one on the left, the Qualcomm Rapid Charger, looks the more promising as it puts out more current.

(To anyone just jumping into this thread: the chargers pictured below are NOT SAFE to use as-is. Dan and I are considering the possibility of modifying them to be safe.
-Flieslikeabeagle

[jeffs555]

I have purchased quite a few cheap car cell chargers of various brands at dollar stores. A few were just a couple of dropping resistors, but most used a MC34063 switching regulator or equivalent, and were all set for around 1 amp current but also were all set for more than 4.2 volts, so the voltage regulation must be done in the phone. Post #22 in this thread http://www.rcgroups.com/forums/showthread.php?t=242053&page=2&highlight=charger+cell has a circuit by Quax, that uses the MC34063, and is very close to the circuit in most of the chargers I have. It should be easy to change the voltage and current resistors to get the voltage and current you need.
Here is the data sheet for the MC34063 http://www.onsemi.com/pub/Collateral/MC34063A-D.PDF

[RD Blakeslee]

... RD, please let us know if you ever decide that single-cell Li charging is ultimately safer for more people than parallel charging, and more likely to replace series charging. From what we've heard so far from many people, parallel charging appears to have an edge in lower cost while single-cell charging appears to have an edge in overall safety.

Dave,

I do think single-cell charging is a little safer:

http://www.rcgroups.com/forums/showpost.php?p=2288908&postcount=12

And, parallel charging is cheaper, as you point out.

I'm mostly concerned that neither method is likely to replace series charging; I would gladly settle for either one.

- RD

[Dave Hederich]

RD, thanks for the clarification. It appears that your name should be more closely linked with Li charging safety than with Li charging in parallel. Your receptiveness to both single-cell and parallel charging opens things up.

Perhaps it's best to refocus this thread on the original question, and carry the single vs. parallel charging discussion back to the Batteries and Chargers forum where similar discussions have been held.

As the starter of this thread, I have to say that my interest has been shifted back from parallel to single-cell Li charging. Anyone who's interested in a parallel Li charger is welcome to carry on here.

[Comatose]

How about a cheap hybrid? Charge in parallel, but sense, error out and disconnect bad cells individually? Has most of the cost benefits of parallel charging, and most of the safety benefits of individual charging.

[flieslikeabeagle]

rpage53 wrote:

When I tried to repeatedly point this out last year, I got loudly shouted down. I don't see a 20A 10 cell parallel charger as a foolproof method without a lot of protective circuits that effectively turn it into 10 single cell chargers fed from a common source.

Rpage, sometimes it takes several repetitions before people hear you. Don't feel bad, I'm sure your earlier attempts to get this considered have had some effect.

The current limiting, voltage limiting 1 cell circuits are available for free from surplus battery packs.

Rpage, can you give us any more information on this approach? Suitable surplus packs, sources, any mods required, anything? I have some older 1400 mAh prismatic Li-ion cells that may be able to donate their protection circuits.

If I can get protective circuits cheaply for my 99-cent chargers, that might be a very cost-effective approach to my lipo charging needs. :)

By the way, my wife and I visited Victoria last spring. It is one of the most beautiful places I have ever been. I envy you, living in the midst of so much natural beauty!

-Flieslikeabeagle

[Dan Baldwin]

Dan, if I understand you, you're saying the FMA guards are series devices rather than shunt ones? In other words, they go to high resistance when the voltage across the cell exceeds the safety threshold?

I'm having difficulty seeing this. Since the charge balancers simply limit the voltage across the cell, and have no "foldback" or "crowbar" action, it seems to me the cells end up with the full 4.2V being held across them indefinitely...and that means they will charge fully, given enough time. In fact, it seems to me that once the charge balancers kick in, they actually turn the constant current source into a constant voltage source - think of the Thevenin equivalent circuit, a current source dumping into a device tht clamps the load terminal voltage to a constant value.

I was hoping you'd say something like that :) . I have no experience with switching regulators myself. Here's a picture to assist you in the hunt for the 99-cent chargers. The one on the left, the Qualcomm Rapid Charger, looks the more promising as it puts out more current.

(To anyone just jumping into this thread: the chargers pictured below are NOT SAFE to use as-is. Dan and I are considering the possibility of modifying them to be safe.
-Flieslikeabeagle
I did manage to confirm that the FMA safety guard is not the same as a balancer. The following is a quote from an application note on the FMA site.

Safety Guard’s primary application is to prevent overvoltage when charging LiPo packs with a current-regulated LiPo charger. Connected between charger and pack, Safety Guard limits charge voltage to 4.2 volts per cell. When Safety Guard’s output voltage exceeds this amount, Safety Guard disconnects the pack from the charger—minimizing the chance of cell damage and dangerous conditions.

That means that the FMA safety guards would not make a good charger along with a current source because you would be missing out on the constant voltage part of the charge.

Although it would be possible to use shunt balancers with a current source to make a charger, if the balancers are plugged into the pack separate from the charger, I don't consider it to be safe. It would be too easy to have a little too much resistance in the leads, or a bad connector, which could lead to dramatic cell failure. They would make a good backup to make sure that the charger doesn't overcharge in addition to balancing the pack.

I did go to the 99 cent store, but I couldn't find any of the 110 VAC travel chargers. I did find some of the car charger, though. They had a bunch of the Belkin F8V701. It uses the MC34063 chip, so I added a pot and cut out a zener, and converted it to a 2 cell charger. It would be easier to convert it to a 1 cell charger. You wouldn't have to cut out the zener, all that would have to be done is to add a pot to decrease the value of 2 resistors that are in parallel. The current limiting is on the primary (12 volt) side, so the output current will depend on the output voltage. With a short circuit, it puts out about 1 amp. When set for 8.4 volts, it only puts out about 500 ma. The output of this charger is not a floating type, so it could not be used along with others from a common power source to charge individual cells in a series pack.

Dan

[Dave Hederich]

That means that the FMA safety guards would not make a good charger along with a current source because you would be missing out on the constant voltage part of the charge.

Kokam Safety Guards can be used with certain types of power supplies as part of a Li charging system. They will limit total voltage to a series pack, but not individual voltage to individual cells. Is that what you meant by the statement above?

[flieslikeabeagle]

Dan, you wouldn't happen to have a pic of the Belkin F8V701? And/or the mods you made? I have a "99 cents only" store nearby and plan to go take a look at the chargers you found.

Meanwhile I just cracked open one of the "Qualcomm Rapid Chargers" I bought for 99 cents, and found a small circuit board with a single chip marked "PJ34063". Same chip by a different manufacturer, presumably. So I should be able to mod these the same way you did yours.

From a look at the MC34063 data sheet, (thanks for the link, Jeffs555!) it appears that all I have to do is ensure that there is 1.25V at pin 5 of the IC when the output voltage reaches 4.2 V. Correct? Also, what is the value of Rsc in your chargers (the current limit set resistor)?

Single-cell, 1A chargers for 99 cents + the cost of a pot or precision resistor - I love it! This is just about the perfect rating for my $7 Etec 1200 cells, and if I double the current limiting resistor, should work with the two-for-five-bucks 650 mAh palm-pilot Li-ion cells from All Electronics.

Nothing like Lithium-ion power on the extreme cheap! :D

-Flieslikeabeagle

[flieslikeabeagle]

Okay, with my interest piqued, I traced the schematic of the Qualcomm charger I cracked open earlier, by examining the traces on the PCB.

The schematic is very similar to the one that Jeffs555 linked to, except for the omission of the red LED and slightly different component values. My current limit resistor is 2.2 ohms, and the voltage set resistors appear to be 1100 ohms and 5190 ohms(??) respectively, if I'm reading the color code right (5 bands, 1% resistors). It looks as though all I have to do is put a trimpot across the 5190 ohm resistor and tweak it till I get 4.15 or 4.2 V at the output. Oh, and replace the 2.2 ohm resistor with a 0.47 ohm or two 1 ohm resistors in parallel.

Dan, anyone, are these switching regulators happy when unloaded (zero current draw), or do they need a minimum load current to work properly? I thought switchers needed some current flow through the inductor to work...

-Flieslikeabeagle

[RD Blakeslee]

Referring to post #40:

Dave, With respect to charging LiPos, my interest in safety is more specific: Charge at 4.2 volts only – no overvoltage charging is possible. That, alone, is an important safety enhancement for LiPo chargers, whether in the hands of a tyro who doesn’t understand what he’s doing, or the experienced series charger user who sets the wrong cell count, or the “automatic” charger which does his miscounting for him, or while attempting to charge unbalanced packs. Both individual and parallel charging can be done at 4.2 volts; Series charging cannot.

The difference in the charger for individual charging compared to one for parallel charging is really a matter of degree, namely, how much current can the charger supply? The chargers are identical in operation – a single-cell charger should put out a nominal C, whereas the parallel charger should output a nominal CX, where X is the number of cells in parallel.

In either case, we need a 4.2 volt charger with amperage output variable to accommodate various Cs and CXs.

So I hope we can continue to have the folks who can do that for us continue to work on the problem here.

Incidentally, re too much current into an overdischarged cell: I would settle for a simple low-voltage lockout, similar in operation to the Apache chargers’s.

- RD

[RD Blakeslee]

Maybe I should add that knowldgeable users already have all they need for 4.2 volt charging at any amperage up to 20 (including milliamp trickling, if necessary) with a Mastech 3020, or the like. However, it's possible to go overvoltage with it, since it's voltage is variable.

Perhaps a number of the considerably less expensive Mastech models which output lower amperages would be practical to individually charge cells, or parallel sets of cells, in a pack.

Again, for what you might call altruistic reasons, I would like to see a cheap "Mastech equivalent" with voltage clamped @ 4.2

- RD

[Dan Baldwin]

Okay, with my interest piqued, I traced the schematic of the Qualcomm charger I cracked open earlier, by examining the traces on the PCB.

The schematic is very similar to the one that Jeffs555 linked to, except for the omission of the red LED and slightly different component values. My current limit resistor is 2.2 ohms, and the voltage set resistors appear to be 1100 ohms and 5190 ohms(??) respectively, if I'm reading the color code right (5 bands, 1% resistors). It looks as though all I have to do is put a trimpot across the 5190 ohm resistor and tweak it till I get 4.15 or 4.2 V at the output. Oh, and replace the 2.2 ohm resistor with a 0.47 ohm or two 1 ohm resistors in parallel.

Dan, anyone, are these switching regulators happy when unloaded (zero current draw), or do they need a minimum load current to work properly? I thought switchers needed some current flow through the inductor to work...

-Flieslikeabeagle

Yes, these switching regulators are happy with no load. The plan you outlined above should give you the charger you are looking for, but one word of caution. If you replace the 2.2 ohm resistor with a .47 ohm resistor, you may push the current to a level that the coil can't handle. Also keep in mind that the current set resistor sets the current on the primary side of the SMPS, so it will not charge your 4.2 volt battery at 640 ma if you use a .47 ohm resistor. It will probably be double that. (12V X .64A = 7.68 W, 7.68W/4.2V=1.83A less losses)You should probably do that voltage mods first, see where the current is, and use that value to calc the new resistor. I'm not sure what value the current sense resistors are in my charger. One says 1R5, so probably a 1.5 ohm, but it has another resistor in parallel that says 5000. I'm not sure what value that one is. The combined value is too small to easily measure with my meter.

The nifty curly cord should probably be replaced with a larger gauge, shorter wire because the resistance of the tiny wire will significantly slow the charge down.

I can post a picture of the charger I bought and the mods that I made, but I'm not sure when it will be, because I'll be busy most of the day. Gotta fly you know.

Dan

[Dan Baldwin]

Kokam Safety Guards can be used with certain types of power supplies as part of a Li charging system. They will limit total voltage to a series pack, but not individual voltage to individual cells. Is that what you meant by the statement above?

Yes, Kokam Safety guards can be used as a charger, but they don't gradually limit the charge current when the batteries reach 4.2 volts/cell, they turn off completely. I don't know if they will cycle on and off. If they do that, they would end up charging the battery up competely, similar to the Astro 109. If they just turn off and stay off, they would only give you the constant current part of the charge. Perhaps Electroman can shed some light on this issue.


Dan

[Dave Hederich]

Dan, now I get what you mean. The following text from the Safety Guard White Paper on the FMA Direct site may explain it:

"You can charge 2s LiPo packs with Safety Guard and a NiCd/NiMH fast charger rated for 7 cells or less. These fast chargers output constant current, but LiPo packs must be charged with current limiting followed by voltage regulation. The result is that a fast charger will apply too much voltage near the beginning of the charge cycle. Charging this way shouldn’t damage the LiPo pack, but the pack won’t reach the same charge level as it would with a dedicated LiPo charger. For best results with this arrangement, use a lower current setting. The lower the current, the more charge can be applied to the pack. Safety Guard terminates charging when its output voltage reaches 4.2V per cell."

https://www.fmadirect.com/support_docs/item_1143.pdf

As I understand CCCV charging Li cells, when the current rate is lowered during the CC stage, the CC stage will require longer to complete but will charge the cell(s) closer to capacity. This results in the CV stage requiring less time to complete, which helps explain why doubling the charge rate from, for instance, 0.5C to 1.0C does not cut charge time in half.

From the description in the SG White Paper, it sounds as if the SG terminates charging at the conclusion of the CC stage, which is when the total pack voltage hits 4.2V times the number of cells in the pack. The suggestion in the White Paper to use a lower current setting as "the lower the current, the more charge can be applied to the pack" fits with the concept of charging at lower current during the CC stage to take the cell(s) closer to full capacity before entering the CV stage.

If this deduction is accurate, then it makes the SG less desirable as an element in a system where charging Li cells to full capacity is desired. On the other hand, for those who prefer only using the CC stage and eliminating the CV stage to save total charge time and perhaps extend cell life due to not charging to full capacity, Safety Guards may still have merit.

[flieslikeabeagle]

Dan wrote:

You should probably do that voltage mods first, see where the current is, and use that value to calc the new resistor.

Good idea! I have an 8V, 6.5A dc power supply bought from All Electronics. It should be able to feed at least four and maybe upto eight of the 99 cent chargers simultaneously, depending on their current limit and efficiency. If 8V is too low for them, I have an adjustable 0-15V, 25 A power supply that can feed as many of the 99c chargers as I can find :) . One per Lithium cell I own would be nice!


One says 1R5, so probably a 1.5 ohm

Yeah, that's pretty standard notation outside the USA. Putting the "R" or "K" or "M" in the location of the decimal point reduces typos that turn 4.7 ohms into 47 ohms, for instance. The idea being that it is easier to misread a decimal point (or be confused by a speck of dirt) than an alphabet.

The schematic by Quax that Jeffs555 linked to uses two 1-ohm resistors in parallel, that's why I thought I'd try a .47 ohm. But I like your suggestion of measuring what I have first, just to be on the safe side.

I'll be busy most of the day. Gotta fly you know.

Ah, flying most of the day - it's a dirty job, but someone's gotta do it! :D

-Flieslikeabeagle

[jeffs555]

It should work with 8v input to charge a single cell. The chip is spec'd for 3v-40v input. Like Dan said earlier, with the current sense on the input like it is, the actual current limit will depend on the ratio between the input voltage and output voltage, so if you decrease the input voltage and keep the output voltage and current sense resistor the same, the output current will drop. Best to do like Dan said, set the output voltage divider, then measure the output current with whatever power supply you are going to use, and change the current sense resistor accordingly.

[flieslikeabeagle]

Jeff, sounds good. I will do as you and Dan suggest.

I have to remind myself these things (switching regulators) don't work the same way as an analog series voltage regulator, and that the efficiency doesn't drop when you feed it a higher input voltage. In addition to the 8V, 6.5 A power supply, I also have a 20V, 3A power supply. If I use this one I may not need to change the current limit set resistor. I've confirmed that the electrolytic caps on the board are rated at 25V, so 20 V input should be okay.

Now if I can find where I put those 10-turn, 1K trimpots I know I have somewhere in the house...

-Flieslikeabeagle

[RD Blakeslee]

The thread seems to have morphed - where's the variable amperage capability the thread was predicated upon?

Oh well. Maybe some other time, some other place...

[Dave Hederich]

RD, what I've picked up from this thread is that there appears to be no great need in the world for a high-current variable current controller. In addition to not scoring high on the recognition charts in the DIY Electronics forum, I couldn't find anything on the internet with a Google search.

From my limited electronics knowledge, I'm thinking the way to go might be to get a schematic from a variable voltage/current power supply, or take one apart and try to eyeball the variable current circuit, and see what's there. What we're really talking about here is taking the variable current circuitry out of a variable voltage/current power supply and putting it on a a 5V power supply along with another device to reduce voltage to a fixed 4.2V.

I'm afraid that this project lost a little interest for me with the understanding that the Kokam Safety Guards are set to cut off charging when pack voltage reaches 4.2V. This effectively eliminates the CV stage of the CCCV charge. So even if a 4.2V SG were made available, it would not be a silver bullet to limit the 5V power supply to 4.2V for purposes of a CCCV charge. That means a more DIY approach would be needed to turn a 5V power supply into a 4.2V power supply. Some of the folks on this forum would probably welcome the challenge. But for most people it would be a problem if it's not a simple "bolt-on" solution.

The barriers appear to be pretty high to creating the elusive safe, reasonably priced, easy to procure 4.2V/variable current power supply to use as a charger for parallel Li cells that may require fairly high current. The barriers appear to be lower as current requirements get lower, and lower current requirements point us at single-cell charging.

In selecting the best 4.2V charging option, it's looking more and more as if the single-cell solution is making more sense than the parallel solution. As we discussed earlier, the single-cell solution is a little more expensive to implement, but it's also safer. That's not a bad trade-off.

[flieslikeabeagle]

Dave wrote:

...it's looking more and more as if the single-cell solution is making more sense than the parallel solution. As we discussed earlier, the single-cell solution is a little more expensive to implement, but it's also safer.

I think you summarised it nicely. I'd like to clarify something already pointed out earlier in this thread by someone else (Rpage??). In order to make a high-current parallel charger truly safe, it would be necessary to add circuitry to monitor each of the individual cell currents, and take appropriate action if overcurrent was detected in any single cell. These monitoring circuits would have to have electronics to measure the voltage across the cell or the current through it, and some kind of series device that could be used to shut off any cell drawing too much current. Once you have individual circuits capable of doing all this, you might as well make them also do the CC/CV thing, as there will probably be negligible (if any) cost difference. At this point you are back to a single power supply feeding a large number of individual, single-cell CC/CV chargers...

-Flieslikeabeagle

[RD Blakeslee]

Dave, I agree with everything you say.

I would make the additional point that, To be able to charge the range of most commonly used LiPo cells individually, we would need cc/cv chargers with current-varying capability (collectively?) between .5 and 5 amps. If we were going to include capability to charge parallel sets in an XP pack, it might be useful to have .5-10 or so.

- RD

[RD Blakeslee]

... The LM317 is capable of being used from 1.2v - 37v and has a current capacity of 1.5A so that would be the limit for each modual as far as charging current goes. However the LM317 in its K package(TO-3) has rating of 2.2A with a max of 3.4A when it's input to output voltage ratio is < 15v, so that opens you up to some big single cell packs at a 1C charge rate. When the LM317 is used as a current regulator you coud use a POT for output adjustments but I would feel safer using a fixed values set via a jumper setting. Here is a link to it's spec sheet in PDF form - LM317 (http://www.national.com/ds/LM/LM317.pdf)

Marv,

This seems pretty good to me, unaccustomed as I am to things electronic. Jumpers are OK. I guess each cell's charger would have to have it's jumper set. Is there any way to do this collectively, i.e., a simple circuit to collectively set all the charger amperages through a single jumper?

- RD

[RD Blakeslee]

We can bear in mind that the amperage setting is not nearly as critical as the permanent 4.2 volt setting, but I guess we would need a low-voltage lockout, for safety when used by novices.

[Dave Hederich]

In order to make a high-current parallel charger truly safe, it would be necessary to add circuitry to monitor each of the individual cell currents, and take appropriate action if overcurrent was detected in any single cell. These monitoring circuits would have to have electronics to measure the voltage across the cell or the current through it, and some kind of series device that could be used to shut off any cell drawing too much current. Once you have individual circuits capable of doing all this, you might as well make them also do the CC/CV thing, as there will probably be negligible (if any) cost difference.

This does tend to negate the price difference and point again to single-cell charging as the safest and best overall solution.

I hope everyone else is finding this discussion as productive as I am. It's amazing what can be accomplished when a lot of people contribute their thoughts on a subject. I started this thread leaning toward parallel Li charging, and now see the wisdom of focusing on single-cell charging thanks to the comments of many contributors.

Solutions for those who use simple 2S and 3S packs are relatively simple. Solutions for those who use complex packs such as 4S5P are much more complex. I don't see anything in my near future beyond simple 3S packs, so if I just focus on my own needs, life is simple. When I try to think of solutions that will handle everyone's needs, I start getting a big headache!

My variable voltage/current power supply will see me through until various single-cell charging options become available. I'll continue to visit DIY Electronics to see what kind of DIY projects the regulars on this forum come up with as options to the upcoming commercial units.

[Marv829]

Marv,

This seems pretty good to me, unaccustomed as I am to things electronic. Jumpers are OK. I guess each cell's charger would have to have it's jumper set. Is there any way to do this collectively, i.e., a simple circuit to collectively set all the charger amperages through a single jumper?

- RD

I can't think of a way to set the output current to multiple chargers with a single jumper. However it could be done with a rotary switch but you would be limited by the switches poles & positions for how many chargers and current choices you would have. As Art H stated each cell with this type of setup must be charged individually- ie: NOT charged individually while still being wired in a series wired pack.

[Dan Baldwin]

I think you summarised it nicely. I'd like to clarify something already pointed out earlier in this thread by someone else (Rpage??). In order to make a high-current parallel charger truly safe, it would be necessary to add circuitry to monitor each of the individual cell currents, and take appropriate action if overcurrent was detected in any single cell. These monitoring circuits would have to have electronics to measure the voltage across the cell or the current through it, and some kind of series device that could be used to shut off any cell drawing too much current. Once you have individual circuits capable of doing all this, you might as well make them also do the CC/CV thing, as there will probably be negligible (if any) cost difference. At this point you are back to a single power supply feeding a large number of individual, single-cell CC/CV chargers...

-Flieslikeabeagle

As long as you only try to charge a single pack in parallel, all the cells should be very close to balance, so there shouldn't be any problem with cells out of balance, and taking a great deal more current than the others. It would be necessary to check the voltage to confirm that all cells are very close in voltage, and no cell is shorted. After that is done, about the only thing that could cause a major problem would be if a single cell were to short during the charge. A fuse could be put into the lead for each cell to prevent that eventuality from taking out the other cells. As RD can tell you, parallel charging does work, and can be done safely.

Dave
If a circuit was added to the output of a 5 volt supply that would limit current, it wouldn't take much more to make it limit voltage too, so the Kokam charge guard would not be needed at all. The main problem with that is that you only have .8 volts (5 volts-4.2 volts) of voltage drop to work with. It can be done, but it's a bit more difficult than building a charger using a higher voltage. I have been playing with a simple linear circuit to do it, and when I get a chance I'll do some more playing. I'll let you know what I come up with. A switching charger would be better, but the inductor needed for 20 amps would be pretty big and expensive.

Dan

[flieslikeabeagle]

As RD can tell you, parallel charging does work, and can be done safely.

I agree, Dan, however as both you and RD mention, some manual care has to be exercised, i.e., verifying that cells are closely balanced before starting the charge, checking that no cells are shorted, etc. If these things are done, evidently parallel charging is safe and effective, and has served RD and some other folks well.

I was thinking about ways to make the charging almost idiot-proof, requiring no manual measurements or safety checks. To do that, I think it is necessary to have electronics to individually monitor each cell, so the user doesn't have to. And that path more or less logically leads to the "single power supply, individual CC/CV chargers per cell" approach.

To clarify: I talk about "idiot-proofing" not because idiots should be using Lipos, but because everyone, idiot or not, makes mistakes sometimes. "Idiot-proofing" is a way to reduce the number of mistakes we make. When it comes to potentially hazardous things like charging Lipos, I'll take all the "idiot-proofing" I can get, to save me from my own mistakes. It is a badly named term, it would be better termed something like "mistake-proofing", which better describes its real purpose.

The main problem with that is that you only have .8 volts (5 volts-4.2 volts) of voltage drop to work with. It can be done, but it's a bit more difficult than building a charger using a higher voltage. I have been playing with a simple linear circuit to do it

P-channel logic-level power MOSFET with its source tied to the +5V line? I think some of those have low enough Rds to handle 20 A at less than .8 V drop. And a CMOS op-amp with rail-to-rail output swing to drive the MOSFET gate, and the usual voltage reference and feedback.

A suggestion as to the inductor: All Electronics sells a variety of inexpensive ferrite cores and toroids, up to some fairly large sizes.

-Flieslikeabeagle

[Art Horne]

Flieslikeabeagle,
Greetings,
Have you seen this link: http://www.rcgroups.com/forums/showthread.php?t=269638&page=1&pp=15 ???

It essentially does what you want, I think. All cells are monitored and charged individually.

Regards,

Art

[Dan Baldwin]

P-channel logic-level power MOSFET with its source tied to the +5V line? I think some of those have low enough Rds to handle 20 A at less than .8 V drop. And a CMOS op-amp with rail-to-rail output swing to drive the MOSFET gate, and the usual voltage reference and feedback.

A suggestion as to the inductor: All Electronics sells a variety of inexpensive ferrite cores and toroids, up to some fairly large sizes.

-Flieslikeabeagle

The experimentation that I've done so far is using a P channel fet (STP80PF55) a pair of op amps; one for voltage control, and one for current control and a small shunt for sensing current. It works, but so far I haven't been able to get 20 amps out of my PC power supply and still have it hold 5 volts. Part of the problem is that I'm only using one of the original wires coming out of the power supply, and it has too much voltage drop at high currents. Another problem I have is coming up with a load for 20 amps. That's a lot of tail light bulbs.

Dan

[RD Blakeslee]

When will your Tanic charger be built for sale in substantial numbers, Art?

- RD

[RD Blakeslee]

Dan, for charging LiPos, does the PS have to hold 5 volts @ 20 amp output?

As the LiPo approaches the transition from limit amperage to limit voltage, won't the current trend downwardly even if the voltage is below the limit? Won't the voltage then rise to the limit as the current drops? In other words, the "knee" from limit current to limit voltage would be rounder - not so sharp?

[flieslikeabeagle]

Art, your charger looks very interesting! Very complex, too!

I thought about this approach briefly at one time (charging via taps and using Kirchoffs current laws and the current through each tap to monitor the current through each cell in the series stack). Somewhere on RC Groups is a little schematic I posted showing how the current through each cell could be measured by sensing the current at each tap as well as the endpoints of the pack.

It became obvious that this method was going to be complex, and would need to use a microprocessor of some sort. Beyond my electronics skills...so I dropped the idea.

I wish you well with your charger, I think it has the potential to be the "Rolls Royce" of chargers when its done. (Or should that be "Volksvagen" of chargers, given what happened to Rolls?). I expect it will be beyond my budget, though.

I have been thinking about a simpler idea of late, and that is an analog circuit that puts out regulated voltages at 4.2 V, 8.4V, 12.6V, and so on. The highest output voltage needs to only source current, while the intermediate voltages need to be able to both source and sink current. Each output will be current-limited. This would not be hard to achieve, the highest voltage output would be a standard CC/CV regulator, the in-between voltages would use a complimentary push/pull output stage as used in most audio power amps, complete with current limiting circuitry. If fact, it may be possible to use an inexpensive audio power amp IC for these outputs.

A circuit like the one I am describing above would do essentially the same thing - charge series packs, using the taps to maintain charge balance during the charging process. An analog circuit of this sort cannot easily monitor all the currents, so safety will be ensured by setting all current limits to 1/2 C (in the worst case of extreme cell imbalance, with 1/2 C flowing down through the cell above, and 1/2 C flowing through a tap, any single cell could recieve up to 1 C current).

Anyone think this is worth following up on? I have little time to spare, but if anyone else wants to run with it, I'd be happy to start the ball rolling. I think a couple of cheap integrated circuit audio amp IC's and a voltage regulator chip might do the trick, if we can find some audio amp circuits with the right current limit values built in. If not, discrete power transistors can be used, but the parts count will rise a lot.

RD, I think you make a good point. During the CC portion of the charge, current draw is at a maximum, but less than 4.2V/cell is being used, so some voltage droop from the supply is acceptable. During the CV portion of the charge, 4.2V/cell is necessary, but less currentt is being drawn, so the power supply will droop less. I think you're onto something here!

-Flieslikeabeagle

[Dave Hederich]

Dave
If a circuit was added to the output of a 5 volt supply that would limit current, it wouldn't take much more to make it limit voltage too, so the Kokam charge guard would not be needed at all. The main problem with that is that you only have .8 volts (5 volts-4.2 volts) of voltage drop to work with. It can be done, but it's a bit more difficult than building a charger using a higher voltage. I have been playing with a simple linear circuit to do it, and when I get a chance I'll do some more playing. I'll let you know what I come up with. A switching charger would be better, but the inductor needed for 20 amps would be pretty big and expensive.

Dan, I think I've pretty much written off the 4.2V Safety Guard, which doesn't exist, anyway. ;) So, yes, if voltage limiting could be added to the variable current circuit without too much trouble, that would be the way to go. It's important to have something limiting the current to precisely 4.2V, because as RD says, that's the key to having a safe parallel charger due to the sensitivity of Li cells to overvoltage.

I have to go back to the bogey of the Matech 0-20A/0-30V power supply for $250. I was hoping it would be possible to start with a 20A/5V power supply and turn it into a 0-20A/4.2V power supply with minimal effort and considerable cost savings over the $250 Mastech. Other than, hopefully, a lower price, the only advantage it would have over the Mastech would be the safey of being fixed at 4.2V.

Also, my 20A limitation was based on the fact that the existing Kokam SGs are only rated for up to 20A. With the SGs out of the equation, there's no reason to limit the power supply to 20A. For example, I ran a quick internet search the other day and found several different 60A+/5V switching power supplies for <$50 (check the 5th one from the top on this site: http://www.alltronics.com/power_sources.htm ). So, depending on the cost and complexity of the add-on variable current/4.2V circuit, it seems that a high-current Li parallel charge could be produced at fairly reasonable cost. But you would have a much better idea than I do of the approximate cost of such a circuit.

[RD Blakeslee]

Dave, that's an interesting switcher you found!

Someone who knows what they're doing could use that supply and a five or ten amp Mastech, consecutively, to charge 40 AHs worth of LiPos in parallel @ 1C in a couple of hours: Charge on the 60/5, monitoring until 4.2 volts is reached, then disconnect and conect to the Mastech for the final taper.

- RD

[Dan Baldwin]

RD
Yes, if the power supply sagged a bit below 5 volts under full load, the only effect should be to soften the knee, but my control circuitry runs off the same 5 volts, and I'm afraid it was causing problems there.

Dan, I think I've pretty much written off the 4.2V Safety Guard, which doesn't exist, anyway. ;) So, yes, if voltage limiting could be added to the variable current circuit without too much trouble, that would be the way to go. It's important to have something limiting the current to precisely 4.2V, because as RD says, that's the key to having a safe parallel charger due to the sensitivity of Li cells to overvoltage.

I have to go back to the bogey of the Matech 0-20A/0-30V power supply for $250. I was hoping it would be possible to start with a 20A/5V power supply and turn it into a 0-20A/4.2V power supply with minimal effort and considerable cost savings over the $250 Mastech. Other than, hopefully, a lower price, the only advantage it would have over the Mastech would be the safey of being fixed at 4.2V.

Also, my 20A limitation was based on the fact that the existing Kokam SGs are only rated for up to 20A. With the SGs out of the equation, there's no reason to limit the power supply to 20A. For example, I ran a quick internet search the other day and found several different 60A+/5V switching power supplies for <$50 (check the 5th one from the top on this site: http://www.alltronics.com/power_sources.htm ). So, depending on the cost and complexity of the add-on variable current/4.2V circuit, it seems that a high-current Li parallel charge could be produced at fairly reasonable cost. But you would have a much better idea than I do of the approximate cost of such a circuit.

Trying to test a 60 amp charger would take some SERIOUS tail light bulbs. Not sure that I could stay within my budget. If the charger were cheap enough, you could build 2 or 3 that would handle 20 amps each and run them off the same power supply. I'll see if I can make the 20 amp charger work first, then go from there.

Dan

[Dave Hederich]

Dan, I just want to clarify that I'm only looking at the really high current power supplies as an option for the relatively small number of Li users who have the really big packs that can include dozens of multiple series and parallel cells. I look at this as being a highly specialized application for a limited market.

I think that 20A is more than enough for the average Li user if they wanted to switch to parallel charging. For my own personal needs, I don't see a need for anything over 10A. Assuming the design gets more difficult with rising maximum current levels, then it probably makes more sense to pick a current level that would satisfy the mainstream of Li users.

Reading through all the Li messages in the Batteries and Chargers forum, I see the typical Li user mostly using 2S or 3S packs of 2500mAh cells or lower. Let's say 3S 3000mAh capability would take care of a large percentage of users. Assuming the 3S 3000mAh pack is wired for series discharge and parallel charge, a parallel charger would need to deliver ~9A @4.2V to charge at 1C.

So a 10A switching power supply should be a sufficient starting point for a basic design that could then be scaled up for the smaller part of the market dealing with higher capacity cells and/or more of them. If you go back to the Alltronics link in my previous message, you'll see that the 2nd item down is a 10A/5V switching power supply for just $14.95.

Assuming that this is an appropriate power supply to serve as the basis of a 0-10A/4.2V parallel Li charger, and that the cost of the variable current/fixed 4.2V circuit is reasonable, this could result in an inexpensive parallel Li charger that would fit the needs of a large percentage of Li users who would like to parallel charge their cells.

[Art Horne]

Hi Guys,

Well, I've got good news and bad news. I've been fighting a problem for the last two weeks that I've finally figured out what was happening. Switching noise in the ground circuit was resetting the microprocessor. As a result I've got to do some major board re-layout which means my release date is going to be pushed out a bit. The good news is I'm using this opportunity to make it a 6-channel insead of 5-channel. Now you will be able to charge two 3SxP packs at a time. Also, I'm making the channels modular so it can have one, two, three, four, five or six channels in the box. Of course the most bang for the buck will be the six channel as the interface part of the charger is the same for any number of channels.

RD,
I'm sorry I can't give any definite date for production yet. I had hoped torelease the prototype to Brin for his evaluation last week but no joy. As soon as he gets his prototype I'll let everybody know. It's not becasue I'm not trying. Only so many hours in a day.

Regards,

Art

[Dave Hederich]

Art, I'm really glad that you are working on this project, as it sounds much more complex than a simple parallel charger. Looking forward to seeing your finished product.

[flieslikeabeagle]

Hi all, just wanted to draw your attention to another (and in my opinion, better) way to balance a Lipo pack while charging it. This method requires taps on the pack. It seems to have all the safety advantages of the "multiple single-cell chargers" approach, while at the same time having the wiring simplicity of the series pack with low current inter-cell taps (rather than having two wires per cell brought out to a connector capable of handling the full discharge current).

I was thinking about the basic concept I outlined in post #69 above, and today I had a flash of inspiration, a way to make it much better. I decided to call it "active balancing", and started a thread on it. Then another RC Groups user (RMFISH) informed me that someone else (a Mr. Chuck Bernard) had this idea, a couple of years before me, and already had a very similar schematic up on his website.

No matter, whoever had the idea first, its still a good idea, IMHO. Mr Bernards deserves the credit for what I think is the best approach to Lipo balancing yet. Take a look at the thread I started, and decide for yourself:
http://www.rcgroups.com/forums/showthread.php?p=3070971#post3070971

-Flieslikeabeagle

[Dan Baldwin]

There may be another approach to making a high current single cell CCCV charger. I modified a 12V switching power supply with a simple 3 transistor circuit to add current shutdown, turning it into a 3 cell CCCV charger. It might be possible to do something similar with a 5 volt supply. The charger I built is fixed current (1.2 amps) but I think I could make one that the current could be variable.

Dan

[Dave Hederich]

Dan, I probably won't understand it all, but I would be interested in seeing what you accomplish. If I spend much more time in DIY, I'm going to end up visiting Radio Shack to buy one of those junior electronics learning labs. ;)

[Dan Baldwin]

You may notice that there are no big, high current devices on my board. I'm letting the power supply do all the heavy lifting. It already has a shutdown circuit to regulate the voltage to 12 volts, or in this case 12.6 volts. I just added another circuit that will also shut the power supply down based on current. I'll post a schematic tomorrow.

Dave, may not want to get into electronics unless you have time for another full time hobby. It can be as habit forming as RC. RC takes all my money, electronics takes all my time. Do you suppose that's why my wife left?

Dan

[Dan Baldwin]

These are the modifications I made to the 12 volt power supply to turn it into a 3 cell CCCV charger with fixed 1.2 amp charging rate. The power supply has a pot for setting voltage, and 12.6 volts was in the middle of the range. I replaced a coil in the output circuit with a .5 ohm resistor, and added a transistor circuit that shuts down the power supply when the voltage across the resistor reaches about .6 volts, which happens at about 1.2 amps. I also added an LED circuit that turns off the LED when the charge current gets down to about 1/20 C.

The power supply I used is a surplus supply that is available from a local electronics store for $2.50.

Dan

[flieslikeabeagle]

Dan, I am so sorry to hear that your wife left. Broken marriages seem to be the curse of our time. It seems that at least two-thirds of the people I know are dealing with something similar, or worse, are still stuck in an unhappy marriage. This always sounds trite, but it's true: give it time, and things will get better.

Back to the topic at hand, thank you for the schematics. It's nice to run into someone else who still thinks analog. And who knows that 5.1 - 5.6 V Zeners have about the lowest tempco of any zener voltage. :)

If I understand you correctly, you set the CV voltage to 12.6 V with the existing trimpot, and added the 0.5 Ohm resistor and PNP/NPN combo to limit the current to about 1.2 A. Correct? Does the 12.6V setpoint seem stable enough for Lipo charging needs, where an extra 50 mV per cell can be too much?

Also, can you post any details of the switching power supply you are using? Make, model, where did you buy it from?

It might not be a bad idea to mention that high voltage switching circuits can have dangerous, even lethal amounts of electric charge stored in capacitors, even after the plug is pulled...extreme care should be taken any time work is done on a circuit like this.

-Flieslikeabeagle

[Dan Baldwin]

Actually, my wife left about 20 years ago, so I'm just about over it.

I think in analog a lot, but I also think in digital quite a bit. I'm ambidextrous that way......or is it schizophrenic.

This power supply uses a TL431 to set the voltage, which has a max delta Vref of 2.7 mv/v over the entire temperature range(-40C to 105C). That would translate into a maximum of about 34 MV at 12.6 volts, which seems acceptable to me. In actuality the temperature deviation over the range of about 20C to 40C is a tiny fraction of that.

The power supply is a 12 volt 5 amp 60 watt surplus supply bought at J K electronics in Westminster for $2.50 The last time I checked a few months ago, they still had hundreds of them. They don't have a part number, you just ask for the surplus 12 volt 5 amps switching power supply. The phone number is (714) 890-4001.

Yes, any time you poke around with something that plugs into a 120 volt outlet, there is always the danger of lethal shock, and SMPS's have a capacitor that stays charged to about 170 volts for a time after power is disconnected, but these power supplies, like most, have bleed resistors, so the voltage on the high voltage capacitors bleeds down fairly quickly.

Dan

[Dave Hederich]

Nothing like talk of a failed marriage to kill a thread. ;)

(Been there. Done that. Got it right the second time.)

Dan, you said the following earlier in this thread:

Dave
If a circuit was added to the output of a 5 volt supply that would limit current, it wouldn't take much more to make it limit voltage too, so the Kokam charge guard would not be needed at all. The main problem with that is that you only have .8 volts (5 volts-4.2 volts) of voltage drop to work with. It can be done, but it's a bit more difficult than building a charger using a higher voltage.
The only reason that I was looking at 5V power supplies is because I thought it would be easier to limit 5V to 4.2V than if starting with a higher voltage. If I understand your statement above correctly that it would be easier to make a fixed 4.2V Li charger out of a power supply with >5V, I can do a little research. Just let me know what voltage I should be looking for.

[Dan Baldwin]

Nothing like talk of a failed marriage to kill a thread. ;)

(Been there. Done that. Got it right the second time.)

Dan, you said the following earlier in this thread:


The only reason that I was looking at 5V power supplies is because I thought it would be easier to limit 5V to 4.2V than if starting with a higher voltage. If I understand your statement above correctly that it would be easier to make a fixed 4.2V Li charger out of a power supply with >5V, I can do a little research. Just let me know what voltage I should be looking for.

For a simple linear circuit, the best power supply would probably be about 7 volts, but I'm afraid that there isn't anything like that available. There may be a few 6 volt supplies around, but probably very few. The most widely available supplies will be 5 volt, and 12 volt. 12 volts would be too high for a linear circuit because the circuit would get very hot and require a large heat sink. 5 volts is probably going to be the ticket, it will just take some careful designing.

I think the best prospect is to find a widely available 5 volt supply that can easily be modified to turn it into a single cell charger, as I did the 12 volt supply for 3 cells.

Another posibility would be to use a 12 volt supply and a switching voltage/current regulator, but I'm trying to keep this as simple as possible so I would prefer a linear solution if possible.

Dan

[Dave Hederich]

Dan, in the previous link I posted to Alltronics, the 6th power supply down is an 8V, 68A Astec switching power supply. At $49.95, it's not a low-buck solution. But it does give you voltage room to play with, and at 68A it would handle the needs of many of the high current users.

http://www.alltronics.com/power_sources.htm

[Dan Baldwin]

8 volts would certainly give plenty of voltage head room (about 4 volts), although the power dissipated would be about 40 watts per cell at 10 amps, so it would require a large heat sink with a fan to keep the pass transistors cool. I could probably do some modifications to the charger that I've already posted to make them work with an 8 volt supply. http://www.rcgroups.com/forums/showthread.php?t=304448 They are fixed current chargers, but different current sense resistors can be switched in for different currents.

I would still prefer to do something with a 5 volt supply, since they are VERY cheap, and power dissipation would be very small, but unfortunately I'm not going to have time to work on it for the next month. I may have time to post the mods to my existing chargers for single cell, 8 volt input, 10 amp max charge rate.

By the way; I went to JK electronics over the weekend, and they still have LOTS of the 12 volt 5 amp power supplies that I used in make a 3 cell charger shown in the post above.

Dan

[Dave Hederich]

Thanks, Dan. No big priority on this. There just doesn't appear to be a huge amount of interest in parallel charging Li cells. So a fixed voltage 4.2V power supply with variable current control is never going to be a big priority for the vast majority of Li users.

[Dan Baldwin]

Parallel charging wouldn't be my favorite method of balancing. I think the safest method would be to charge in series using shunt balancers, and use a circuit that shuts down the charger if the balancing current of any balancer gets too high, but I've never seen anything like that. I guess that will have to be my next project.

Like I needed another project:)

Dan

[nsp]

Thnx to Dave's topic!
I'm newbie and I know a little about electronic.
Now, I want to design a Variable Current Control DC Power Supply. I have read all article but cant find any clue!
I think that the switching PS is best choice.
My demand is certain circuit or schema to design a 0-15A DC PS. The maximum voltage is 18VDC. The current is regulated and controlled by user.

[Dan Baldwin]

Thnx to Dave's topic!
I'm newbie and I know a little about electronic.
Now, I want to design a Variable Current Control DC Power Supply. I have read all article but cant find any clue!
I think that the switching PS is best choice.
My demand is certain circuit or schema to design a 0-15A DC PS. The maximum voltage is 18VDC. The current is regulated and controlled by user.

Welcome to RC Groups.

A 0-18 volt regulated, 0-15 amp regulated switch mode power supply is not something that would normally be considered a DIY project unless you have a heavy electronics background. It's a very ambitious project.


Dan

[coro]

A 0-18 volt regulated, 0-15 amp regulated switch mode power supply is not something that would normally be considered a DIY project unless you have a heavy electronics background. It's a very ambitious project.


Dan
Dan, if one could make LINEAR psu of satisfactory parameters, it is much easier than making of switching psu.
Another part of project could be swithing "prescaler" part, taking 110 or 230V, or 12V, as an input, feeding linear part with appropriate voltage - this switching part greatly reduces heat at linear part. This way, linear part could improve dynamic parameters against switching psu, and help to reduce ripple...