Thread Tools
Aug 21, 2009, 06:06 AM
7000mw of raw power!
rich smith's Avatar
Thread OP
Mini-HowTo

High current 24v DC supply for almost nothing


Here's a highly regulated high current 5v 12v 17v 24v DC supply that costs almost nothing and only takes minutes to build with no soldering. It is based on two junk PC supplies in series (see diagram) and can put out 20a-30a depending on units. Specially useful for battery charging or anywhere high power regulated DC is needed. Originally I was going to put this in the Battery/Charger area but a few felt it was beyond typical users there safety wise so here we are.

My motivation for building this came from need for 22v+ to simulate photovoltaic panels for testing Harbor Freight 45watt solar charger. Some old PC supplies were gathering dust because they lacked the 4 pin 12v connector most new Intel motherboards require. Why not put them to use? Many supplies can be had for free from PCs literally left by the roadside or almost free at holiday sales.

DC +12v (yel) from the bottom supply is wired to DC ground (blk) of the top one which effectively puts the supplies in series. The two cases must be insulated from each other because the top one ends up at +12v. I used several pieces of double side foam tape to separate them physically and electrically. Note that it is best to mount them with ventillation grills facing away from each other. A side benefit is the covers can be removed without undoing the foam tape. Care must be taken to avoid shorting the cases. This will generally cause the +12v to shut down which at best means power must be cycled to restore. Worst case a fuse will blow and need replacing. A strip of packing tape or duct tape around the gap can help prevent this.

The top case must now be isolated from the AC ground pin. A 2 prong adapter can be used to do this or the green AC wire can be disconnected internally. There were concerns about disconnecting earth ground from the top case but closer inspection shows this is not really a problem. There are no dangerous line voltages floating around or exposed AC if done properly.

Another possibility is isolate the 4 internal PCB grounds of the top unit which means no need for the 2 prong adapter. In this "case" both end up at earth ground like the original supplies. It involves a lot more work and soldering so I've chosen the path of least "resistance". (do I detect 2 puns in the space of 2 sentences here? )

As always take care when dealing with these kind of voltages. Use a power strip that has a breaker. Don't work on them while plugged in and check all wiring with a meter at each stage. Also a good idea to verify voltages on both supplies before starting so you won't waste time with a defective unit. If you are not familiar with these types of projects then have someone do it who is.
Last edited by rich smith; Jan 06, 2010 at 07:54 AM. Reason: 24v
Sign up now
to remove ads between posts
Aug 21, 2009, 07:08 AM
7000mw of raw power!
rich smith's Avatar
Thread OP
Quote:
Originally Posted by Kaz the Minota
Yeppers I will. I have some double sided sticky foam and I bought an adapter to disable the ground on 1 PS.

It doesn't matter which PS ground you disable does it?

What were you using for connectors on the ends of the wires? I'dlike to find some large crip on rings or something.
Again, hope you don't mind moving this to threads as PM keeps filling up. And others may have similar questions. (deleted comments about posters in the other thread for obvious reasons)

I suggest isolating the top supply otherwise the bottom case will be -12v compared to nearby earth grounds which can be problematic. Better to have the lower one at earth ground, less likely to contact nearby objects. I suggest following orginal setup as close a possible. At least initially.

Those connectors are original Molex ones from the supplies. I tucked the unused ones inside the cases. Didn't want to do any unnecessary cutting or soldering. I swear the whole thing took me less than 10 minutes. I was in a rush to check that solar charger.

They actually plug into another assembly consisting of a UBEC modified for variable voltage and two meters similar to the ones in the $3 wattmeter thread:

https://www.rcgroups.com/forums/showthread.php?t=949923
Aug 25, 2009, 07:18 AM
7000mw of raw power!
rich smith's Avatar
Thread OP
Quote:
Originally Posted by Kaz the Minota
If I combine 2 in series do you see any real benifit to bringing the 12V rail on the one PS up as high as I can? By adding resistors on the 5V rail?

I understand that it isn't necessary in this case though. I hooked up (5) 1157 light bulbs uo to the 5V rail and got my +12 rail up over 13V.

That would probably be too much though...as in to much stress on that PS.
Not much advantage because the main purpose of this series idea is so we don't have to resort to that resistive load stuff. I had very bad luck with that approach. The supply get's so hot it was necessary to turn on the air conditioner!

And yes, I think it does put extra stress on the supply. All the power burned up in a bulb/resistor is that much less available to do something useful like charging a battery.
And I don't think it helps the life of the supply either.
Aug 30, 2009, 09:59 AM
I will be an angry old man!
marc1's Avatar
Wouldn't this setup benefit from back-biased diodes to protect each PS? Otherwise turning one on even slightly before the other... any cause for concern about a common ground?
Aug 30, 2009, 02:48 PM
Registered User
Andrew0820's Avatar
Quote:
Originally Posted by marc1
Otherwise turning one on even slightly before the other... any cause for concern about a common ground?
I've been publishing these conversions for several years now. The only consistancies to be found are the variations in PSU design. Many times these supplies have been pressed into service in environments far outside the design guidelines that dictated their construction. When they fail to perform as we expect, albeit as originally designed, we are sometimes at a loss to explain why. These supplies are built to power, and protect, our PCs --- using them for other functions is simply a plus.

That being said, I have run supplies in series in order to realize higher voltages --- I have not included that conversion in the website simply because it adds another variation that falls even further outside the design parameters. I don't recommend using supplies that require external isolation. Workbenches are many times cluttered and the opportunity of shorting the cases together is a distinct possibility. If it does happen, the most likely scenario is that the overload circuitry of one or both supplies will initiate a latch-off. Worst case will be an exciting pop and possible thrown breaker.

Supplies that I have used in series have the DC ground isolated from AC and earth ground. Some designs isolate DC ground and some do not. If DC ground is isolated, it is not necessary to defeat case grounding or insulate the cases from each other. It also alleviates MARC1's concern for a common ground. Secondly, with the PSU's isolated and running as standalone units, the order of powering up becomes immaterial. Wiring becomes basically daisy chaining positive to negative to positive to negative, just as if you were wiring batteries in series.

Lastly, I've read many comments about the need or lack of need for loading the 5v rail to get the PSU to latch. Older supplies required a load for latching; newer supplies will usually latch without a load. However, I generally advise folks to add a load simply to improve voltage stability on the 12v when under load. Start with 10 ohms and drop as needed. I have not seen any advantage in dropping resistance below 2 ohms. There is seldom any improvement in output voltage and the by-product is additional heat and current consumption.

Because of the great variation in designs, your milage may vary.

andrew
Aug 30, 2009, 04:37 PM
7000mw of raw power!
rich smith's Avatar
Thread OP
Quote:
Originally Posted by marc1
Wouldn't this setup benefit from back-biased diodes to protect each PS? Otherwise turning one on even slightly before the other... any cause for concern about a common ground?
Somebody brought up the diode issue in the thread that started all this. First, 30a diodes ain't cheap and probably cost more than the 2 supplies together. Second, I don't think it buys us anything.

I've been turning both supplies on at the same time but just for kicks just tried one first and then the other first. As expected nothing bad happened.
Aug 30, 2009, 04:48 PM
7000mw of raw power!
rich smith's Avatar
Thread OP
Quote:
Originally Posted by Andrew0820
I've been publishing these conversions for several years now. The only consistancies to be found are the variations in PSU design. Many times these supplies have been pressed into service in environments far outside the design guidelines that dictated their construction. When they fail to perform as we expect, albeit as originally designed, we are sometimes at a loss to explain why. These supplies are built to power, and protect, our PCs --- using them for other functions is simply a plus.

That being said, I have run supplies in series in order to realize higher voltages --- I have not included that conversion in the website simply because it adds another variation that falls even further outside the design parameters. I don't recommend using supplies that require external isolation. Workbenches are many times cluttered and the opportunity of shorting the cases together is a distinct possibility. If it does happen, the most likely scenario is that the overload circuitry of one or both supplies will initiate a latch-off. Worst case will be an exciting pop and possible thrown breaker.

Supplies that I have used in series have the DC ground isolated from AC and earth ground. Some designs isolate DC ground and some do not. If DC ground is isolated, it is not necessary to defeat case grounding or insulate the cases from each other. It also alleviates MARC1's concern for a common ground. Secondly, with the PSU's isolated and running as standalone units, the order of powering up becomes immaterial. Wiring becomes basically daisy chaining positive to negative to positive to negative, just as if you were wiring batteries in series.

Lastly, I've read many comments about the need or lack of need for loading the 5v rail to get the PSU to latch. Older supplies required a load for latching; newer supplies will usually latch without a load. However, I generally advise folks to add a load simply to improve voltage stability on the 12v when under load. Start with 10 ohms and drop as needed. I have not seen any advantage in dropping resistance below 2 ohms. There is seldom any improvement in output voltage and the by-product is additional heat and current consumption.

Because of the great variation in designs, your milage may vary.

andrew
I agree with pretty much everything you said. However this trick is for those like myself who had a couple unused supplies kicking around and can be put to good use. Personally I would not go out and pay full price for two new PC supplies for this purpose. But then I only needed this to test a few solar panels.

In my case benefits outweighed drawbacks and it was easy to do. Easier than soldering up bulbs or resistors. And less power wasted and more available for use.

As you say, the mileage of others may vary.
Aug 30, 2009, 05:24 PM
I will be an angry old man!
marc1's Avatar
Quote:
I've been turning both supplies on at the same time but just for kicks just tried one first and then the other first. As expected nothing bad happened.

The failure would not be instantaneous (like temporarily overloading an ESC or Lipo) ... but the stress would accumulate on the components over time. You can't just say "hey look: here 12V, there's another 5V, let's put them together and we'll have 17V" without taking a serious look the the circuitry inside.

... Then publishing this as a practical solution on a forum where many people look for advice when they may not have the electronics experience to evaluate whether this is a good or bad design idea.
Aug 30, 2009, 05:42 PM
7000mw of raw power!
rich smith's Avatar
Thread OP
Quote:
Originally Posted by marc1
You can't just say "hey look: here 12V, there's another 5V, let's put them together and we'll have 17V"
Yes I can. Watch... I'll do it again:

"hey look: here 12V, there's another 5V, let's put them together and we'll have 17V"

Here it is twice in a row:

"hey look: here 12V, there's another 5V, let's put them together and we'll have 17V"

"hey look: here 12V, there's another 5V, let's put them together and we'll have 17V"
Aug 30, 2009, 07:38 PM
Registered User
Andrew0820's Avatar
Quote:
Originally Posted by rich smith

In my case benefits outweighed drawbacks and it was easy to do.

As you say, the mileage of others may vary.
It was not my intent to be critical of your setup, only that I would probably not recommend it to the casual DIYer. Your experience makes the difference.

On occasion, I get asked if some circuit or lashup is possible. Sometimes, my response is, "Yes, but I don't think you should try it." In some cases, if someone has to ask why, then they probably need a little more OJT before jumping in.

andrew
Aug 30, 2009, 07:44 PM
7000mw of raw power!
rich smith's Avatar
Thread OP
Quote:
Originally Posted by Andrew0820
It was not my intent to be critical of your setup, only that I would probably not recommend it to the casual DIYer. Your experience makes the difference.

On occasion, I get asked if some circuit or lashup is possible. Sometimes, my response is, "Yes, but I don't think you should try it." In some cases, if someone has to ask why, then they probably need a little more OJT before jumping in.

andrew
Again, you appear to be the voice of reason. And I also recommend against this for those who are not familiar with soldering or electricity.
Aug 30, 2009, 07:57 PM
Dave the Rave
dmccormick001's Avatar
By Jove, he did it! Twice!
Jan 06, 2010, 03:01 PM
7000mw of raw power!
rich smith's Avatar
Thread OP
Quote:
Originally Posted by rampman
I tried to put two 17A 12V supplies in parallel as 17A was not enough for my FB's Duo charger at max amps with two packs, one on each circuit.
It still did not carry the charger with what I thought would have been 34 amps avialable. When the second pack went into CC phase it shut down momentarily due to low volts into the charger.
Would a shared ground from wall outlet be the reason? Any ideas on this electron guru?

Thx
Hi Rampman,

Hope you don't mind moving this here because there's a space problem in my email area. Also others may benefit.

There are a couple reasons it may not work and probably not related to the AC outlet at all. There are two other more likely suspects:

First of all if these are typical PC supplies rated for 17a then they are probably not even close to that in reality so you won't get the expected amps in most cases. I derate by 50% for cheapies and 75% for expensive ones.

Secondly when you put two high current switching supplies in parallel it does not always result in twice the current like it does with batteries and BECs. The feedback circuitry is more sophisticated and they may oscillate or "fight" each other. Not common but I've seen it happen.

Moral of the story: better to put these type supplies in series than parallel. There's another thread similar to this one but for server supplies that started today in Battery Charger area. I'm not sure if any new info will crop up there.

Also if you have a "good" charger it's better to put double the voltage in than double the current. This results in more effiicient and cooler operation. Unfortunately you can't do this with the cheapies.
Jan 06, 2010, 07:27 PM
Registered User
Interesting case of 2 switchers in parallel. My 2 cents is try putting a series
diode on each one, tie the cathodes and take the combined output there.

2 resistors instead, of .3-1 ohm might help too. Then add a nice big cap on
the final output to ground to smooth things out.
Jan 07, 2010, 04:24 AM
7000mw of raw power!
rich smith's Avatar
Thread OP
For those concerned about mains isolaton etc. here's an idea from the copycat thread:

Quote:
Originally Posted by BobOD
By the way, my suggestion, assumming no liability of course, is to get yourself a used PC case. Mount the 0-12v one grounded and the second (or third...) on insulated standoffs. Then everything is inside a grounded case.
I don't have the space or motivation but it certainly removes most of the risk for those that do so worth mentioning.


Quick Reply
Message:

Thread Tools

Similar Threads
Category Thread Thread Starter Forum Replies Last Post
Question High Current Variable Power Supply bos10 Dock Talk 7 Nov 16, 2008 04:39 PM
For Sale 1000w Mastech High Current Variable Regulated Dc Power Supply 50v 20a Hy5020e Mastech Aircraft - Electric - Batteries & Chargers (FS/W) 0 Nov 15, 2008 02:15 PM
AC power supply for Hitec DC charger? John Webb Power Systems 4 Nov 14, 2001 01:25 AM
Sources for high current wire? MtnGoat Power Systems 2 Oct 16, 2001 10:53 PM
Vendors for high current/capacity single cells? Paul Electric Plane Talk 2 Jun 06, 2001 11:02 PM