Need help with server psu conversion

[Rocketzx1]

So I need some help here. I bought 2x HP 750w server power supplies off eBay. Before I started to do the conversion to wire them into series for 24v output, I did the 33pin and 36pin “trick” to make the power supplies turn on. I checked the out puts of both power supplies, a steady 12.3 volts each.

I label them as slave and master, as the slave is the one to float the ground to allow for 24v when wired in series together. I go through all that and the slave power supply now has a floated ground.

I now go to recheck the slave power supply voltage output and it’s all over the place. Starts off at 12.3 goes to 2.4 drops to zero jumps to 11.1 drops to 5.6 etc. Is that normal for the power supply with the floated ground? Or have I done something wrong?

Please help!

 

[Tex Koder]

So I need some help here. I bought 2x HP 750w server power supplies off eBay. Before I started to do the conversion to wire them into series for 24v output, I did the 33pin and 36pin “trick” to make the power supplies turn on. I checked the out puts of both power supplies, a steady 12.3 volts each.

I label them as slave and master, as the slave is the one to float the ground to allow for 24v when wired in series together. I go through all that and the slave power supply now has a floated ground.

I now go to recheck the slave power supply voltage output and it’s all over the place. Starts off at 12.3 goes to 2.4 drops to zero jumps to 11.1 drops to 5.6 etc. Is that normal for the power supply with the floated ground? Or have I done something wrong?

Please help!

Got a photo of that wiring setup currently?

 

[Chapinb]

Something is wrong. I did this same thing a long time ago, and they put out a rock steady 25v regardless of amp draw, right up till a breaker trips.
Other then the pin mod, the only other thing I remember doing is removing all grounds from the slave. The slave dosent ground to the outlet, nor the master. Just to its own case. I'm pretty sure the ground from slave and the "pos" from master were isolated from each other too somewhere. And the cases are isolated from each other of course. I do remember distinctly scratching the ground trace in the back corner of the board to remove that ground.

It kind of sounds like your slave is getting a signal that it's not expecting, causing a fault and turning off that power supply briefly. A ground that should be floating, but its seeing it at 12v or 110v for example.

 

[Rocketzx1]

Got a photo of that wiring setup currently?

Yep. The last photo is the slave power supply with the floated ground. Voltage is all over the place.

Something is wrong. I did this same thing a long time ago, and they put out a rock steady 25v regardless of amp draw, right up till a breaker trips.
Other then the pin mod, the only other thing I remember doing is removing all grounds from the slave. The slave dosent ground to the outlet, nor the master. Just to its own case. I'm pretty sure the ground from slave and the "pos" from master were isolated from each other too somewhere. And the cases are isolated from each other of course. I do remember distinctly scratching the ground trace in the back corner of the board to remove that ground.

It kind of sounds like your slave is getting a signal that it's not expecting, causing a fault and turning off that power supply briefly. A ground that should be floating, but its seeing it at 12v or 110v for example.

From my understanding, you float the two ground points in the front and leave the single ground point in the rear so the slave will output 24v through the positive side. The earth ground or case ground, am I supposed to leave that unattached?

 

[Rocketzx1]

I’ve re-read the guide I’m going off of 4 times, and nothing is said about reattaching the earth ground or case ground. I’m going to remove it and see what happens.

Edit- Made no difference. Voltage is still all over the place. I’m not sure what else to check.

 

[Tex Koder]

I’ve re-read the guide I’m going off of 4 times, and nothing is said about reattaching the earth ground or case ground. I’m going to remove it and see what happens.

Edit- Made no difference. Voltage is still all over the place. I’m not sure what else to check.

Are they plugged in to the same wall outlet?

Try plugging each PS to a different outlet.
Perhaps something with the grounding in your house's AC circuit.. perhaps?






*************

unfortunately, my backplane (dell) is different from yours..

 

[Rocketzx1]

Are they plugged in to the same wall outlet?

Try plugging each PS to a different outlet.
Perhaps something with the grounding in your house's AC circuit.. perhaps?






*************


View attachment 392407

View attachment 392408

View attachment 392409

View attachment 392410

That’s one big power supply, the ones I bought are small.

And ill give the different outlet idea a try.

I pulled the board out and the voltage is still all over the place, sitting on the rubber work pad.

I just noticed that these aren’t the exact same power supplies, I bought them as a bundle from eBay. The one that works is an HP HSTNS-PL29 643955-201, while the one giving me issues is HSTNS-PD29 643955-101. Would that possibly be an issue?

Ugh, this was supposed to be a simple project and it’s turned into a headache.

 

[Tex Koder]

That’s one big power supply, the ones I bought are small.

And ill give the different outlet idea a try.

I pulled the board out and the voltage is still all over the place, sitting on the rubber work pad.

I just noticed that these aren’t the exact same power supplies, I bought them as a bundle from eBay. The one that works is an HP HSTNS-PL29 643955-201, while the one giving me issues is HSTNS-PD29 643955-101. Would that possibly be an issue?

it might (could be)

Ugh, this was supposed to be a simple project and it’s turned into a headache.

So, may I ask -what are you trying to power up with that setup originally?

 

[Rocketzx1]

it might (could be)





So, may I ask -what are you trying to power up with that setup originally?

I bought an iCharger Dx6, its a dc only lipo charger. Instead of tossing out 150 bucks on a 1500 watt power supply, I read up on roughly 100 threads and posts over about 2 weeks, on how to do the conversion yourself and save some money. I spent 56 bucks total for everything to do the conversion myself. The server power supplies were 26 bucks.

I just wanted to have a good, clean dc power supply putting out 24v and have a fun project to do. As of right now all I'm getting is 12.3v

 

[Tex Koder]

I bought an iCharger Dx6, its a dc only lipo charger. Instead of tossing out 150 bucks on a 1500 watt power supply, I read up on roughly 100 threads and posts over about 2 weeks, on how to do the conversion yourself and save some money. I spent 56 bucks total for everything to do the conversion myself. The server power supplies were 26 bucks.

I just wanted to have a good, clean dc power supply putting out 24v and have a fun project to do. As of right now all I'm getting is 12.3v

Hmm..
Okay.

Would you have a link to that conversion in using your chosen server PS's?

Even if it worked as described..
You might want to find a way to cover up those exposed copper contacts on the backplane.. because that could have a short.

 

[Rocketzx1]

Hmm..
Okay.

Would you have a link to that conversion in using your chosen server PS's?

Even if it worked as described..
You might want to find a way to cover up those exposed copper contacts on the backplane.. because that could shortout.

This guide is the one I followed.

https://blog.seidel-philipp.de/hp-dps-1200fb-power-supply-hack-for-charging-lipos/

 

[Rocketzx1]

@Tex Koder @Chapinb

So I found out the issue maybe?

The case has 3 contact points for the board, and each one has copper on top and bottom to ground the board. Since I took out the 2 in the front, the guide I was using said to leave the one in the back left. I also noticed that their specific model power supply DID NOT have an earth ground, and mine do.

So I put a nylon nut and screw through the back left mount, and the voltage jumping around is less now? It’ll stay steady at 12.2 vokts, drop to 10, jump up to 12.3 drop to 8.4, but it’s not instantly dropping or jumping voltage, it’s more intermittent.

Still doesn’t solve the issue of voltage being all over the place. I think I’m just going to order another power supply.

 

[Tex Koder]

@Tex Koder @Chapinb

So I found out the issue maybe?

The case has 3 contact points for the board, and each one has copper on top and bottom to ground the board. Since I took out the 2 in the front, the guide I was using said to leave the one in the back left. I also noticed that their specific model power supply DID NOT have an earth ground, and mine do.

So I put a nylon nut and screw through the back left mount, and the voltage jumping around is less now? It’ll stay steady at 12.2 vokts, drop to 10, jump up to 12.3 drop to 8.4, but it’s not instantly dropping or jumping voltage, it’s more intermittent.

Still doesn’t solve the issue of voltage being all over the place. I think I’m just going to order another power supply.

It's got to be the grounding issue internally..
Perhaps ground both PS cases together?

 

[Jerold]

You know what might be happening is the common ground of the chassis to the AC outlet and they are probably non-Isolated the DC ground.

So when you put them in series you're might be grounding the 12V output to the ground of the second one, which would probably just shut down at that point. . . hopefully.

I would check out the ground to see what's shared.

Most likely the Chassis ground is the AC ground and possibly the DC ground too.

Turn off the supplies and check the continuity from the DC ground to the AC ground and the chassis all three ways. Either one or both could be a common ground. If that is the case not sure how to fix it unless there is another pin for a floating ground on the supply?

This guide is the one I followed.

https://blog.seidel-philipp.de/hp-dps-1200fb-power-supply-hack-for-charging-lipos/

In this he isolates the chassis ground from the DC ground with the nylon hardware. Once you do that the ground is floating.

 

[Chapinb]

Start probing with a multimeter. There should be no continuity between ac grounds on plug side - slave should have ground wire chopped off from power cable. There should not be continuity between cases. Slave ideally should have no continuity between ac ground and dc ground internally.
Master might have ac and dc ground connected.
Its possible to have some continuity between any of the 4 dc outs.

Glad to see your makin progress, but don't use it in that state. There's still a fault and likely won't put out any amperage to back up that voltage. Even if it does, fluctuating voltage is bad for the slave and charger.

 

[Rocketzx1]

Start probing with a multimeter. There should be no continuity between ac grounds on plug side - slave should have ground wire chopped off from power cable. There should not be continuity between cases. Slave ideally should have no continuity between ac ground and dc ground internally.
Master might have ac and dc ground connected.
Its possible to have some continuity between any of the 4 dc outs.

Glad to see your makin progress, but don't use it in that state. There's still a fault and likely won't put out any amperage to back up that voltage. Even if it does, fluctuating voltage is bad for the slave and charger.

So I’m green around the ears with testing continuity and trouble shooting problems like this, and have a few questions to make sure I am understanding this correctly.

When I connect the test leads to the AC ground and the case ground, the reading is 000 with a continuous beep. That indicates there is no continuity? And if there is continuity it’ll show me a number?

 

[Rocketzx1]

I just checked all 3 board ground points and the back left is giving me intermittent readings, the front 2 are dead as they should be, reading 0v to “float” the ground, correct?

So if the back left board ground is giving me intermittent readings, that’s the issue??

I’m shooting in the dark here.

 

[Jerold]

So I’m green around the ears with testing continuity and trouble shooting problems like this, and have a few questions to make sure I am understanding this correctly.

When I connect the test leads to the AC ground and the case ground, the reading is 000 with a continuous beep. That indicates there is no continuity? And if there is continuity it’ll show me a number?

If you are using the diode setting your multi-meter that would be correct. (Green below)

It makes a noise when there is a low impedance path. So if you checked the paths (below) and all three make noise, you have a common ground between the two power supplies and you central (floating ground) is not actually floating.

Personally, because I like numbers, I would us the Ohm "Ω" (Pink) setting and check each path. There are three tests in all

• Chassis-DC Ground
• Chassis-AC Ground (third terminal)
• DC Ground - AC Ground

Make sure it is powered OFF
It doesn't matter what probe is on which terminal, Ohms are non-directional, but you can certainly try it both ways.

If you see near 0Ω (~1Ω or less) readings for all, then you have are your connecting the DC 12V power of the first supply to the ground of the second one. Which is, as we say in electronics no bueno. Then I would look at the tutorial that you noted earlier. He isolates the ground. But since you have slightly different supplies so it might be a bit different procedure. DPS-1200FB (his) vs the HSTNS-PL29 643955-201 and HSTNS-PD29 643955-101 (yours). You may have to look for more screws or ground connections to the chassis.

BTW I didn't find any reason why the supplies wouldn't work with. They "seem" to be compatible except for the server family. With that said we won't get HP datasheets with pin outs and schematics.

If you see any of the paths have higher values post it back.

 

[Rocketzx1]

If you are using the diode setting your multi-meter that would be correct. (Green below)

View attachment 392507

It makes a noise when there is a low impedance path. So if you checked the paths (below) and all three make noise, you have a common ground between the two power supplies and you central (floating ground) is not actually floating.

Personally, because I like numbers, I would us the Ohm "Ω" (Pink) setting and check each path. There are three tests in all

• Chassis-DC Ground
• Chassis-AC Ground (third terminal)
• DC Ground - AC Ground

Make sure it is powered OFF
It doesn't matter what probe is on which terminal, Ohms are non-directional, but you can certainly try it both ways.

If you see near 0Ω (~1Ω or less) readings for all, then you have are your connecting the DC 12V power of the first supply to the ground of the second one. Which is, as we say in electronics no bueno. Then I would look at the tutorial that you noted earlier. He isolates the ground. But since you have slightly different supplies so it might be a bit different procedure. DPS-1200FB (his) vs the HSTNS-PL29 643955-201 and HSTNS-PD29 643955-101 (yours). You may have to look for more screws or ground connections to the chassis.

View attachment 392510

BTW I didn't find any reason why the supplies wouldn't work with. They "seem" to be compatible except for the server family. With that said we won't get HP datasheets with pin outs and schematics.

If you see any of the paths have higher values post it back.

I am getting readings all over the place with the chassis dc ground. The other 2 ground are a constant 0.2 reading.

I should also note that I bought 2 more power supplies, and left a note to the seller to make sure BOTH are the exact same to try and avoid any more issues.

I’m not giving up on this power supply yet, it has to be a simple issue that I just don’t kno enough about electronics to figure out.

 

[Jerold]

I am getting readings all over the place with the chassis dc ground. The other 2 ground are a constant 0.2 reading.

~0Ω = Connected
---Ω = Open, no connection

All over the place says there are might components in the the path and it makes it much more difficult to figure out without a schematic.

Which ones have the 0.2Ω?

• Chassis-DC Ground
• Chassis-AC Ground (third terminal)
• DC Ground - AC Ground

Try reversing the probes on the flaky paths, see if we can get some more data.

Was the power supply modified like in the link you sent?