Espritmodel.com Telemetry Radio
Reply
Thread Tools
Old Jan 05, 2010, 09:47 PM
ancora imparo
jj604's Avatar
Melbourne, Australia
Joined Jul 2005
6,067 Posts
Discussion
Using two power supplies for higher voltage/capacity chargers: safety issues

NOTE: Post 183 has my conclusions from the discussion in this thread.

In addition, we know a lot more about these supplies since this thread was started. For me it is now preferable to isolate the DC output ground on the second supply rather than isolate a ground on the AC side (which is what this thread is about). However this requires disassembling one of the power supplies and having accurate and correct information about what to do to modify it.

If you have the necessary skills and knowledge it is in my opinion the better way to go.

If you don't have that skill and knowledge you need to ask yourself whether you should be doing this. These supplies can kill you if you you don't know what you are doing.

I am recommending nothing - it is your decision if you want to tackle a dual series supply project or not.


With the growth of high capacity battery chargers there has been interest in beefier power supplies capable of managing two chargers simultaneously of up to 500Watts each. Commercial supplies of this capacity are expensive, particularly if you want 24V to get the maximum performance out of some of the new chargers like the Hyperions and others.

There have been a number of posts on modifying IBM, Compaq and HP server power supplies to provide 12V at 40-110A and this is pretty simple. A number of people such everydayflyer and feathermerchant have successfully combined two of these to provide 24V at high currents at moderate cost. These guys are experienced experts and know what they are doing but not everybody is, and there are some genuine risks in this approach.

I thought it would be helpful to start a thread where we could amass some reliable opinion to help the rest of us. If you are a lurking qualified electrician who really understands wiring practice please contribute!
NOTE: This is not the place for the “I made a $20 PC supply work by connecting the green wire….” stuff. There is plenty of info on that all over the net. This is specifically about connecting two high power server supplies in series to provide double the voltage.

To get started here is the problem, as I understand it – corrections and clarifications welcome.
The crude diagram attached shows how 12V two supplies are connected with their outputs in series to provide 24V.
The difficulty is that these supplies are all metal cased boxes with a common ground connection for the output and the 110 or 240V Mains ground wire. Under these conditions there is a connection ABCDEFG that is essentially a short across the output of Supply#2. The way round this is to break the loop at EF by disconnecting the ground pin on the #2 Mains plug. The following issues arise:

1.If the cases touch, then the short is re-established. Easy to fix by simple insulation or mechanical separation.

2.The manufacturer didn’t ground the case to protect the power supply, they did it to protect you. In the unlikely event a PS fails by a short from one of the mains inputs to the case then the resulting surge to ground will blow the fuse. If there is no ground, the surge goes through you if you are touching the case. This can be protected against by FULLY insulating Supply#2 but it’s not as easy as 1 above.

3.Even if Supply#2 is “untouchable”, it is connected to your charger by the –ve lead and if the charger case is connected to that, then the charger and possibly your LiPo can reach lethal mains voltage.

Welcome informed comments on how real the risk is and how it can be managed. For instance, is a fully insulated PS fed from an Earth Leakage Circuit Breaker acceptable practice?
Let the conversation begin.

SEE POST 183 for the final conclusions.

http://www.rcgroups.com/forums/showp...&postcount=183

THE DIAGRAM BELOW IS NOT HOW TO DO IT
.


Also, as noted above, we now know a lot more about these server supplies than when this thread was started 3 years ago largely thanks to xandrios and other experts on the 12V 100Amp supply thread.

Most people would now float the DC ground on the 12-24V supply, not the AC ground if there is information available on how to do that. It is supply specific and some are already built with isolated outputs.

Here's how one guy did this for the supplies I have used in this thread.

http://www.rcgroups.com/forums/showp...&postcount=678
jj604 is online now Find More Posts by jj604
Last edited by jj604; Feb 27, 2013 at 03:33 PM. Reason: Added reference to Post with the conclusions
Reply With Quote
Sign up now
to remove ads between posts
Old Jan 05, 2010, 10:25 PM
Use the 4S Luke
feathermerchant's Avatar
USA, TX, Euless
Joined Aug 2003
9,341 Posts
First off, thank you for starting this thread John.
My observations of the 3 points above -
1) If the cases touch, then one supply's 12V output is shorted and as these are high quality supplies, the shorted supply will just turn off its output. No big deal. MAybea spark and a burned spot on the case. I've seen it before...

2)Even though one supply is ungrounded, by virtue of the fact that its ground contact is connected to the other power supply's +12V output, the ungrounded shpply's case is at 12V above ground. Not a big deal under normal conditions. Forthe case to go above 12V, the grounded supply must somehow have a mains fault to it's 12V or I suppose the ungrounded supply could have a mains to case fault. Maybe we (I) should construcht a test to see what would happen.
I did plug in the ungrounded supply and measure the case potential before connecting its outouts to anything and the case stayed within a few tenths of a volt of 0 anyway.

3)Since the power supply output is grounded on one side, I doubt there's any danger to anything connected to it.

The greatest risk I see is that because I cut off the ground prong on one of the power cords, it could find its way to another application where a ground should be used. I have marked it on both ends with red tape to call attention to its unusual configuration.

I have observed that if I power up the supply first and plug chargers in second, thereis a loud snapas I connect the first charger and one power supply shuts down. I must first connect the chargers then plug in the AC power.

I have been using my unit at our flying field for several weeks now with no problems. All our outlets are run from ground fault circuit breakers. The breakers trip if current traveling out does not match current returning. ie current is 'leaking' to another return path other than the neutral wire in the circuit.

I have posted pics and details on my blog. I can post them here too if helpful.
feathermerchant is offline Find More Posts by feathermerchant
Reply With Quote
Old Jan 05, 2010, 10:36 PM
Registered User
Himalaya's Avatar
Joined May 2009
503 Posts
That's what I did, with a minor mod for conveniency(one plug) and EMI consideration(shorter AC grounding path)



Note the inserted capacitor is just to make the grounding wire DC floating, all protection connections are still connected to ground like before, e.g. the two blue Y-capactors, the shielding case..etc.

This is what it looked like when finished. I got a high quality PS that's quiet(fanless), powerful(220x2=440W), and CHEAP.
Himalaya is offline Find More Posts by Himalaya
Last edited by Himalaya; Jan 05, 2010 at 10:53 PM.
Reply With Quote
Old Jan 05, 2010, 10:42 PM
Registered User
So. Cal.
Joined Oct 2004
7,832 Posts
I'm with Feathermerchant here. Before posting this last year I intentionally shorted the output of one of my 235's dozens of times to ensure that the protection circuitry was doing its job and adequately protecting the supply.

Obviously, running power supplies in series should only be attempted by those who understand and are willing to accept the risks involved. I emphatically agree that more discussion regarding mitigation of said risks is warranted given the number of high power chargers that are becoming commonplace and the resultant desire for affordable high current 24 volt DC. Thanks for starting this thread.

Mark
mrforsyth is offline Find More Posts by mrforsyth
RCG Plus Member
Reply With Quote
Old Jan 05, 2010, 11:39 PM
ancora imparo
jj604's Avatar
Melbourne, Australia
Joined Jul 2005
6,067 Posts
Thanks, for the quick responses everyone. Looks like we might get some good info all in the one place.

re:
Quote:
Originally Posted by feathermerchant View Post
Forthe case to go above 12V, the grounded supply must somehow have a mains fault to it's 12V or I suppose the ungrounded supply could have a mains to case fault.
That is exactly the case that concerns me. That is where the UNGROUNDED PS has a mains to case fault. Would be relieved if someone could authoritatively show it would never be more than a 12V potential but I'm not convinced.

John
jj604 is online now Find More Posts by jj604
Reply With Quote
Old Jan 06, 2010, 03:31 AM
Registered User
Melbourne, Australia
Joined May 2006
6,407 Posts
BTW - the risk of mains voltage appearing on the case and/or the output terminals is very real. I had this happen to me on a switched-mode power supply about a year ago. This was an unmodified commercial supply which simply had an internal component failure. Luckily the insulation breakdown on the coil was only partial so there was still considerable impedance and I received only a minor shock rather than a fatal one. It did trip the earth-leakage breaker as well. Whether that would have been fast enough to save me had the breakdown been worse is hard to say. I am glad it was not put to the test!

As you might expect, I did not replace the PS with another one from that manufacturer.

Most of us are very complacent about electrical products primarily because of the high level of designed-in safety that is provided. My own view is that encouraging folk to bypass the safety features in order to save a few bucks is unwise and I recommend against both giving such advice and taking it. No matter how many caveats are provided, there will be folk who do not understand what they are doing that will try it out. The risk of bad outcomes is real.
kgfly is online now Find More Posts by kgfly
Reply With Quote
Old Jan 06, 2010, 05:05 AM
Use the 4S Luke
feathermerchant's Avatar
USA, TX, Euless
Joined Aug 2003
9,341 Posts
kgfly - If the case was grounded, how did it have any voltage on it at all? Even if there in an internal failure, the case ground should prevent it from being dangerous.
feathermerchant is offline Find More Posts by feathermerchant
Reply With Quote
Old Jan 06, 2010, 06:23 AM
Southern Pride
everydayflyer's Avatar
Haralson County GA. USA
Joined Oct 2004
34,510 Posts
How to, why we should not or just this is what I think?

I have a 40 year old Black & Decker Drill Motor ( hand drill)with a metal housing and I cut the earth ground off of it approx 30 years ago . Why not I never connected the earth ground adapter to ground anyway.

When I was young 120 AC plugs were not polarized,they could be plugged in either way. They were two prong only. Appliances has metal housings. If a toaster and a coffee pot were both turned on and connected the right(wrong) way and you touched both you got a tingle. Unplug one of them and flip the plug ,problem solved.

Now we have polarized plugs, 3 prong plugs, plastic double insulated cases and we still need(?) ground fault circuits to protect us.

A 120VAC circuit has a hot (line) and has two grounds. One ground is called a neutral and the other is called a ground or an earth ground, both grounds connect to the same place that being Earth Ground.

Some bumass DIY type and even some so called electricians have been known to reverse the hot and netural connections at the wall socket. Now if you plug a 3 prong in to such a socket the earth ground (round pin) does a great service.


Received report that latest find are not the same as others. Whomever figures out power up be sure to post.

FYI I have ran two IBM 235s in parallel at 12+ volts up to just past 60amps. I have ran two of them in series at just over 24 volts and 800 watts output. At this point they were causing the overhead room light in my room to flicker a bit.

My series pair has served me well and is used daily.

http://www.rcgroups.com/forums/attac...7&d=1262697082

Charles
everydayflyer is offline Find More Posts by everydayflyer
Last edited by everydayflyer; Jan 06, 2010 at 10:52 AM.
Reply With Quote
Old Jan 06, 2010, 06:30 AM
Registered User
Melbourne, Australia
Joined May 2006
6,407 Posts
Quote:
...the risk of mains voltage appearing on the case and/or the output terminals is very real..
The 240V was present on the output terminal, not the case. You are correct that if the internal fault had been to the case there would have been no risk since the breaker would blow whenever the unit was plugged in or turned on depending upon where the failure lay. My point was simply that these kinds of failures are not unheard of.
kgfly is online now Find More Posts by kgfly
Reply With Quote
Old Jan 06, 2010, 06:49 AM
7000mw of raw power!
rich smith's Avatar
New Hampshire (not the old one)
Joined Dec 2006
5,987 Posts
The same piggyback supply issues were brought up some time ago in this thread:

http://www.rcgroups.com/forums/showthread.php?t=1098857

Risk is real but in my case worth it and solved several power supply problems at almost no cost. Risk of getting hit by meteor or "green ice" is also real and everyone must decide for themselves.
rich smith is offline Find More Posts by rich smith
Reply With Quote
Old Jan 06, 2010, 01:34 PM
G=667x10^-8 cm^-3 gm^-1 sec^-2
dissymmetry's Avatar
Joined Nov 2009
83 Posts
Personally, I keep a car battery on a car battery charger, all the time, I run everything dc off of that, and I have another battery and charger I can bridge to 24v, if I need that much, I have before. The batteries can be individually charged with their own 12v chargers, even while bridged in a series to 24v, without any issues. I know that my setup uses just a 2-10 amp charger, but the battery will make over 600 amps, so since I always keep the charger on it, I can get away with pulling 100 amps for at least an hour or two, without discharging the battery completely.
It works, and doesnt complain if the plug gets pulled, the battery will continue on, for a while, depending on the load.
dissymmetry is offline Find More Posts by dissymmetry
Reply With Quote
Old Jan 06, 2010, 04:32 PM
Southern Pride
everydayflyer's Avatar
Haralson County GA. USA
Joined Oct 2004
34,510 Posts
Why can a Thread never stay on subject here on RC Groups?

Charles
everydayflyer is offline Find More Posts by everydayflyer
Reply With Quote
Old Jan 06, 2010, 04:43 PM
Registered User
USA
Joined Jan 2002
4,520 Posts
Quote:
Originally Posted by everydayflyer View Post
Why can a Thread never stay on subject here on RC Groups?

Charles
Perhaps you can tell us....

I'm still wondering what, if anything, the cost to ship a ten pound bag of sugar or potatoes to Australia has to do with the "HYPERION EOS NET, DUO, and DUO2 Chargers" thread.

http://www.rcgroups.com/forums/showp...postcount=1931
Ohmic is offline Find More Posts by Ohmic
Reply With Quote
Old Jan 06, 2010, 05:17 PM
Registered User
earlwb's Avatar
USA, TX, Grapevine
Joined Dec 2008
12,329 Posts
My question si why would someone want to ship a ten pound bag of potatoes or sugar to Austrailia. it can't be that expensive there. Maybe in Tokyo japan.


Anyway I think jj606 has the right approach to it. the problem is trying to keep the power supply cases apart at least the metal from touching. Plus covering the power supplies somehow to accidentally prevent yourself from touching them too. it doesn't happen often, but every now and then a failure can cause mains voltage to be on the case. usually the powersupply will shut down when that happens as it is grounded, but on the floating ungrounded power supply it needs to get protected somehow.

I like the other guy's idea on the capacitor across the cut ground connection too. it is a good thought.
earlwb is offline Find More Posts by earlwb
RCG Plus Member
Latest blog entry: My Low Wing plane
Reply With Quote
Old Jan 06, 2010, 05:43 PM
ancora imparo
jj604's Avatar
Melbourne, Australia
Joined Jul 2005
6,067 Posts
Why I started this thread

I think I need to clarify something here since this is the kind of thinking I was hoping to avoid.

1) I NEVER "advocated" this approach. That's your decision. What I was trying to do was get some discussion from people who actually knew something about the subject. There is no shortage of opinion on the forums. It isn't always connected to fact. I happen to think, until someone authoritatively convinces me otherwise, that I won't take the risk. Other equally intelligent people will decide otherwise because in their informed opinion the risk is negligable. That's perfectly OK. What is important is that decision is based on facts.

2) "usually the powersupply will shut down when that happens as it is grounded, but on the floating ungrounded power supply it needs to get protected somehow" is EXACTLY what I don't care about. I am aware these are well designed machines that will normally look after themselves if they fail. There is a whole room of them in the floor below my office and they do that very well. But they are in professional racks in a professionally designed data centre. If we use them as Charger power supplies on the bench and disconnect a designed safety feature to do so then it is irrelevant if the PS can protect itself - it's protecting me that this thread was supposed to be about.

John
Quote:
Originally Posted by earlwb View Post
Anyway I think jj606 has the right approach to it. the problem is trying to keep the power supply cases apart at least the metal from touching. Plus covering the power supplies somehow to accidentally prevent yourself from touching them too. it doesn't happen often, but every now and then a failure can cause mains voltage to be on the case. usually the powersupply will shut down when that happens as it is grounded, but on the floating ungrounded power supply it needs to get protected somehow.

I like the other guy's idea on the capacitor across the cut ground connection too. it is a good thought.
jj604 is online now Find More Posts by jj604
Reply With Quote
Reply


Thread Tools

Similar Threads
Category Thread Thread Starter Forum Replies Last Post
For Sale Used PC Power supplies converted for 12V DC chargers ddierking Aircraft - Electric - Batteries & Chargers (FS/W) 3 Jan 13, 2010 08:14 AM
Question Power supplies for lipo chargers? steelfinger Batteries and Chargers 4 Sep 04, 2009 04:28 AM
For Sale 12V Power supplies 30A $40 and 42A $50 for your chargers feathermerchant Aircraft - Electric - Batteries & Chargers (FS/W) 1 Sep 03, 2009 08:07 PM
Discussion Using PC power supplies for chargers? shurcooL Batteries and Chargers 9 Jul 25, 2008 01:44 PM