SMALL - espritmodel.com SMALL - Telemetry SMALL - Radio
Reply
Thread Tools
Old Sep 08, 2009, 09:57 AM
Registered User
United States, AL
Joined Apr 2007
832 Posts
Parkjeff: that's my hope. Even with just a couple extra people posting quality discharge graphs it will be a huge help in sorting out battery performance!

One more question for smart people:
Is power dissipated by the mosfet dependent on the current and voltage? In other words, is 100w/50a/2v the same as 100w/50v/2a as far as the mosfet concerned? Or will the power dissipation capacity change with changing voltage/current even if power stays the same?
biskit is offline Find More Posts by biskit
Reply With Quote
Sign up now
to remove ads between posts
Old Sep 09, 2009, 03:59 AM
CamLight Systems
New York City, USA
Joined Oct 2003
1,172 Posts
Assuming that you're not exceeding any other specs, you're correct, it's the same for 50V/2A as it is for 50A/2V. Think of the MOSFET as just being a resistor for your calculations. Actually, in this, it is....a voltage controlled resistor.
JohnMuchow is offline Find More Posts by JohnMuchow
Reply With Quote
Old Sep 09, 2009, 04:16 AM
CamLight Systems
New York City, USA
Joined Oct 2003
1,172 Posts
Quote:
Originally Posted by biskit
Parkjeff: mosfet is "IPP048N06L G" from mouser. The shunt is just a 1 mohm resistor, but it has to be able to dissipate 2.5w at 50a. I used a "OARSXPR001FLF" from mouser because it was the cheapest I could find. Once I finalize everything I'll make a project list on mouser and you can just click "buy" and get all the parts you need.

John: that sounds like charlie brown's teacher to me! I found a little chip that monitors a thermistor and switches an output. Only two resistors to set the hi/low temperatures to switch at it seems like. Not worried about getting the temperature exact, just curious if it'd be simple to make it to where it wouldn't self-destruct if you used it without water running. I'll save the extra fancy stuff for after it's done I think, otherwise I'd never finish!

So who can make me a nice little aluminum block with a few holes in it?

I also noticed it'd be easy to make a simple 1-sided PCB without needing jumpers or anything. How much does it cost to have some made?
Good resistor choice. I've used them on some reflow-soldered (hot-air) boards.

Only one problem I've ever noticed with the OARS XP resistors. When hand soldering them, the size of the solder fillet can change the resistance value. If it's a high fillet, it's shorting solder around the resistive element. Solder shouldn't come above the flat bottom portion. You might have to test this. I had some problems using them for my electronic loads but I was trying to keep the tolerances below +/-1%.

Oops, sorry about the technobabble biskit.
The chip you have sounds just fine. Be sure to mount that thermistor carefully and protected from the water. Even epoxying it against the MOSFET's tab is good as you can bury it in epoxy to thermally insulate from the water's cooling effect.

I agree, temperature-reading accuracy isn't that important. Coming to within 5C is going to be just fine.

I don't know who has inexpensive single-sided boards but the double-sided ones are almost the same price in low quantities and you can use the second layer for a ground plane. Your op-amps will thank you.
JohnMuchow is offline Find More Posts by JohnMuchow
Reply With Quote
Old Sep 09, 2009, 08:36 AM
Registered User
United States, AL
Joined Apr 2007
832 Posts
Quote:
Originally Posted by JohnMuchow
.....it's the same ......
Four mosfets it is then. At 150w each that's 50a from 1.25v up to 12v. It's tempting to add more to get 50a at higher cell counts, but that's easy for people to do themselves if they need it. I wanted to use bigger ballasts, but even 0.1ohm is 15w at 50a - the 25w aluminum resistors are $2 each, but the 50w ones were $8 or something which blows the budget big time.
biskit is offline Find More Posts by biskit
Reply With Quote
Old Sep 09, 2009, 10:31 AM
Registered User
Rugby, UK
Joined Feb 2007
898 Posts
Quote:
Originally Posted by biskit
Four mosfets it is then. At 150w each that's 50a from 1.25v up to 12v. It's tempting to add more to get 50a at higher cell counts, but that's easy for people to do themselves if they need it. I wanted to use bigger ballasts, but even 0.1ohm is 15w at 50a - the 25w aluminum resistors are $2 each, but the 50w ones were $8 or something which blows the budget big time.
(OK John, I give in!)
Biskit,

No logic is there; it is cheaper to dissipate power in Fets than resisitors! So why not use more Fets and smaller resistors, say 6 fets + 6 x 0.1ohm 7Wresistors). The water cooling is a good idea to dump a lot of heat easily. You could use an aluminium block but water is cheaper and tends to flow a lot better! Calculation suggests the energy in a 3S2200pack would raise a 1Kg lump of aluminium almost 100deg.cent.
If you do use the Aluminium cased resistors, make sure they are properly heat sunk and cooled. If very over -run the epoxy block in the end comes out with the power and sound of a bullet, I've seen one clear a PCB of components.

Wayne
Wayne Giles is online now Find More Posts by Wayne Giles
Reply With Quote
Old Sep 09, 2009, 12:22 PM
Registered User
United States, AL
Joined Apr 2007
832 Posts
Quote:
Originally Posted by Wayne Giles
...why not use more Fets and smaller resistors, say 6 fets + 6 x 0.1ohm 7Wresistors). ........ You could use an aluminium block but water is cheaper and tends to flow a lot better! .....If you do use the Aluminium cased resistors, make sure they are properly heat sunk and cooled.
Guess I wasn't clear, I'm going to mount the mosfets and resistors on a small piece of aluminum or copper with water flowing through passages inside the metal. A heat sink internally cooled with water if you will. I'm not going to sit and hold the thing under the faucet every time I use it!

The mosfets are $2.8 each, the 25w resistors are $3.3 - $24.4 for four of each. six each with $0.55 resistors is $20.4. The aluminum ones can be easily mounted on the metal while the big ceramic ones can't. It's only $4 more and I think it's worth it for me to have fewer parts and make it overall smaller. I want it as compact as possible as the whole thing will be mounted in a small box about the size of a coke can.
biskit is offline Find More Posts by biskit
Reply With Quote
Old Sep 09, 2009, 03:09 PM
Registered User
Rugby, UK
Joined Feb 2007
898 Posts
Quote:
Originally Posted by biskit
Guess I wasn't clear, I'm going to mount the mosfets and resistors on a small piece of aluminum or copper with water flowing through passages inside the metal. A heat sink internally cooled with water if you will. I'm not going to sit and hold the thing under the faucet every time I use it!

.
Yes sorry, I should have understood - its obvious now I read it again.
Running water is a great way of removing heat. I might consider it on a larger version of my discharger. Although 1600W sounds a lot it was at its limit when I tested a 5S4300 20C pack at its max rating, so I am aware that a 6S 5000 30C pack would be way beyond its capability.
Can I suggest an automatic cutoff at, say, 3.0V/cell so that when you are running comparisons of various packs they will all terminate at the same voltage. You can easily do it using an LM393 comparing part of the pack voltage with a reference voltage. Because the pack voltage recovers immediately, you need to put positive feedback on the comparator to latch it. An easy way is a signal diode from output to non-inverting input.

Wayne
Wayne Giles is online now Find More Posts by Wayne Giles
Reply With Quote
Old Sep 09, 2009, 05:13 PM
Registered User
United States, AL
Joined Apr 2007
832 Posts
Inverting latch doohickies is far beyond my comprehension. I do want to make it auto-shutoff and temperature protected(so it doesn't self-destruct when you forget to turn the water on), but I'm nearly clueless on how to do either. Both will be an add-on after I get the basic discharger finished. I wouldn't have gotten this far without John babying me along the path to understanding!

Another question: I've seen devices where you set a value(voltage, current, whatever) with switches or decade-style knobs... how is this done? Is there some logic so that when you dial in, say, 10a, it goes to 10a? Or does it just switch in the resistance that should correspond to 10a? Does that make sense? Rather than a knob I'd like a rotary switch to select a few pre-determined current amounts. Or maybe I should splurge on a nice multi-turn potentiometer? I guess the question is is there any logic or does it depend on having the correct value components?

I'm stuck now... I'm traveling for a few months for work and don't have any way to make a suitable water block device. Hoping a kind soul will offer to make one for a reasonable price! I'm going to hunt around this weekend for a drill press and some taps I can use, we'll see how it goes. Also might try to cobble one up from a small flat piece of copper plate and solder copper tubing on the backside for the water. That'd be easy if I can find a flat piece of copper plate(not so easy).
biskit is offline Find More Posts by biskit
Reply With Quote
Old Sep 09, 2009, 06:46 PM
Dieselized User
gkamysz's Avatar
Chicagoland
Joined Feb 2000
7,415 Posts
Quote:
Originally Posted by biskit

Another question: I've seen devices where you set a value(voltage, current, whatever) with switches or decade-style knobs... how is this done? Is there some logic so that when you dial in, say, 10a, it goes to 10a? Or does it just switch in the resistance that should correspond to 10a? Does that make sense? Rather than a knob I'd like a rotary switch to select a few pre-determined current amounts. Or maybe I should splurge on a nice multi-turn potentiometer? I guess the question is is there any logic or does it depend on having the correct value components?
This is done with fixed resistors rather than a potentiometer. There are a few ways to do it. If you need to be very accurate with the current you will want to put a trimmer on each setting I would think.

Greg
gkamysz is offline Find More Posts by gkamysz
Reply With Quote
Old Sep 10, 2009, 06:10 AM
CamLight Systems
New York City, USA
Joined Oct 2003
1,172 Posts
I do it with rotary decade switches, no resistors. But, the switches go to a PIC microprocessor that drives a digital-to-analog converter (DAC) to create the correct voltage to the op-amps (which results in the correct current level). Certainly more complex than just a pot (which you'd have to calibrate the positions for) or fixed resistors, but it can result is a very accurate current setting. And with only three decade-switches, I get 150 current settings with my loads. That's not practical when using fixed resistors.

I recommend building one unit with a pot and measure the resistances that give you a range of specific current settings. Then use 1% tolerance fixed resistors and a multi-position switch to select the current values in the final design. You'll probably be within 2% of the desired current value this way.
JohnMuchow is offline Find More Posts by JohnMuchow
Reply With Quote
Old Sep 10, 2009, 06:11 AM
CamLight Systems
New York City, USA
Joined Oct 2003
1,172 Posts
Quote:
Originally Posted by Wayne Giles
(OK John, I give in!)

Wayne
Huh?
JohnMuchow is offline Find More Posts by JohnMuchow
Reply With Quote
Old Sep 10, 2009, 09:06 AM
Registered User
Rugby, UK
Joined Feb 2007
898 Posts
Quote:
Originally Posted by biskit
Inverting latch doohickies is far beyond my comprehension. I do want to make it auto-shutoff and temperature protected(so it doesn't self-destruct when you forget to turn the water on), but I'm nearly clueless on how to do either. Both will be an add-on after I get the basic discharger finished. I wouldn't have gotten this far without John babying me along the path to understanding!

Another question: I've seen devices where you set a value(voltage, current, whatever) with switches or decade-style knobs... how is this done?).
Unfortunately my scanner is kaput so I can't post a cct. at present but will try to do so in next day or so. It is simple and cheap to make a latch with a comparator or op-amp.

I had the same problem in setting the current up to 100A in 1A steps and came up with a cheap answer which is to use two 10 way rotary switches, one to set 10A/step and one for 1A/step. I mounted 470ohm resistors x 10 around the tags on the 10A one and 47ohm around the 1A switch. Resistors just go from one tag to next. This means that with the switch set on "0" there is no resistance between "0" position pin and the centre pole, but as you rotate the switch you will get one resistor put in series with each step. If you connect these two assemblies in series and feed a constant current down the chain of resistors the voltage across the two assemblies will be an exact analogue voltage of the total resistance in cct. I used 1% resistors and 1.596mA to give me 75mV/amp load but obviously you can scale it to fit the drive requirements you have. The constant current was from an LM317L which only needs a single restor and pot to set the current to the exact value you need. That all sounds compicated but as a cct. it is simple and cheap (<$3 total) and works perfectly.
I see you are only using a 1 milliohm shunt whereas I used all the ballast resistors and averaged all their voltages to a common point with 100ohm resistors which gave me a much larger voltage to work with. Problem with such small voltages is that it is easy to get an error which multiplies up. If you do use the above system, remember that you cannot take any current from the voltage generated. If you wish to divide the voltage down you must use an op-amp.
Hopefully I can post a cct. soon so that you will see that it is much simpler than it sounds from the above.

Wayne

John - Just meant I did not intend to post but couldn't resist it! As Oscar Wilde and Woody Allen have said "the best way to overcome temptation is to give way to it"
Wayne Giles is online now Find More Posts by Wayne Giles
Reply With Quote
Old Sep 10, 2009, 12:44 PM
Registered User
United States, AL
Joined Apr 2007
832 Posts
Those are all too complicated and will break our budget. I'm thinking a switch that can switch to three ranges by changing resistors, like 0-25a, 25-50a, and 0-10a or something like that. All great suggestions, but this has to stay as cheap and simple as possible otherwise it's just another thing that most people can't justify owning and that defeats the whole purpose.

So I know you can get PCBs made but that's something beyond my experience and available time. How much would a small 5x5cm board cost? I'm thinking about assembling all the parts into a few kits to sell at cost. Mouser shipping was $9 for me, but all the parts could be put in a padded envelope and mailed for $2-3. So 10 people could save $5 each just on shipping. Plus another $ or two saved by buying the parts in larger quantities... If I can find a simple solution to the heat sink maybe this thing could come in under $40 including shipping. Surely that'd get a few people to build one!
biskit is offline Find More Posts by biskit
Reply With Quote
Old Sep 10, 2009, 05:36 PM
CamLight Systems
New York City, USA
Joined Oct 2003
1,172 Posts
Quote:
Originally Posted by Wayne Giles
John - Just meant I did not intend to post but couldn't resist it! As Oscar Wilde and Woody Allen have said "the best way to overcome temptation is to give way to it"
Wayne, I'm glad you did! You've built some great loads and I think your contributions to this thread will be invaluable.

biskit,
One thing to consider is the consistency of using your 3-setting-plus-pot setup. It might cost a dollar or two less than Wayne's idea but his idea offers accuracy and repeatability for the extra money. Depending on the pot and available current ranges it might be hard to dial in closer than +/-0.5A of the desired current. This will make it harder for users to compare their results. It's not a huge difference in current repeatability, but it might matter to a lot of potential users of your load. Just a thought....

Also, pots will get "noisy" as dirt will inevitably get inside, the wiper or resistive element wears out and the connection becomes intermittent. Especially if the users bounces back and forth between a couple of common settings. You'll need a way to keep the unit from jumping to wildly different current levels when the pot's voltages rapidly bounce all over the place when they're adjusted.

Also, have you run the numbers for the 3-setting-plus-pot setup? I'm worried that as you switch to different current ranges that the resistance range of the pot will shrink too much. That is, the pot might be able to change the current by 25A at one setting but only an amp or two at another setting.

I'm really looking forward to seeing this unit come to life!!
JohnMuchow is offline Find More Posts by JohnMuchow
Reply With Quote
Old Sep 10, 2009, 07:54 PM
Registered User
United States, AL
Joined Apr 2007
832 Posts
Priority one is a cooling device. A potentiometer is good enough for now and features can wait.

I swapped the resistor on the potentiometer for the 50a value and re-tested. No pictures, but I got 250w from the single mosfet with the same dinky little aluminum plate. It required much more water flow and the plate was still warm. The poor little plate couldn't take much more, but a decent block of Al/Cu could easily.

For $40 in parts this could easily be a 1kw discharger if I can find a cheap way to make the heat sink!

I want the aluminum resistors because they can be heat-sunk like the mosfets and it can all be made compact and won't glow red hot while running, but it would be cheaper to use smaller resistors and a couple extra mosfets. Which leads to the question about the ballast resistance - what is the lowest resistance that's practical to use? (ie. smaller resistance means less power in the resistors which means smaller/cheaper resistors). I assume the ratio of ballast/mosfet resistance effects the required ballast value? A quick computation at 100-250w per mosfet shows the mosfet resistance varies from 0.65-1.9 ohms depending on how many mosfets you have in parallel... Obviously 0.1 ohms for ballast has a varying effect. If I can get the ballast resistance down safely then I can use cheaper resistors and fewer parallel "channels" and the cost will drop through the floor. So what do you smart guys think about the ballast value?

One more question - what about a different form factor mosfet? I looked, but this "to220" is rated at 300w! There were mosfets up to 600w, but the cost was going up exponentially(double the power, but 7x the price). Any other options I'm not seeing for a single super-power mosfet ?
biskit is offline Find More Posts by biskit
Reply With Quote
Reply


Thread Tools

Similar Threads
Category Thread Thread Starter Forum Replies Last Post
Discussion PIC PWM and FETs as constant-current discharger? meteor DIY Electronics 21 Jan 09, 2010 11:17 AM
Mini-HowTo How to build a constant current battery discharger biskit Batteries and Chargers 25 Sep 06, 2009 08:53 PM
how to build a vaccum forming box? rayc Scale Kit/Scratch Built 14 Feb 28, 2008 11:46 AM
Discussion DIY analog constant-current discharger? meteor DIY Electronics 7 Jan 06, 2006 11:09 AM
Anbody know how to build a prop? Pork Tornado Parkflyers 4 Aug 03, 2001 05:10 PM