|Jan 23, 2009, 05:51 PM|
Any info on Compaq ESP108 as a Power Supply
I know many folks use the IBM 235 Server Power Supply as a charger supply .
There's a 24 pin connector for all the control circuitry and the top row pins # 2 & 3 need to be shorted to cause the PS to power up using a standard CD/hard drive shorting clip.
QUESTION: Does anyone have any info on the Compaq ESP108 (Model DPS-450CB-1).
It's a similar supply with the same connector. I've pretty much sorted out the various outputs from the size of the board traces, but it goes against years of caution to randomnly short out pins of a 450W device to see what happens!
Any help much appreciated.
|Jan 23, 2009, 06:22 PM|
Joined Oct 2004
the connector is most likely one of these: http://www.smpspowersupply.com/connector_atx_pinout.GIF
and as you can see by a quick google image search it's an industry standard:
Shorting pin 16 (marked as "PS ON" on that picture, pin 14 on the 20 pin connector) to pin 15, 17 or any other of the ground lines should turn the PSU on. However, even though the more recent PSU's are more sturdy, they don't like to be run with no load and will quickly overheat and possibly fail if you do so of any length of time. They might even fail to start giving you the right voltage if they don't have any current flowing through them.
|Jan 23, 2009, 06:33 PM|
Much appreciated. But that connector (which is well documented) is not the one I mean . You need to have seen one of these server supplies to know what I'm about. They use heavy duty blade connectors for the actual current and a small rectangular 6x4 pin connector for all the logic and sense.
I understand the warnings - thanks for that. I've converted a few "ordinary" PC supplies and am familiar with some of the loading quirks.
|Jan 23, 2009, 06:42 PM|
Joined Oct 2004
Ah, sorry for that. When you wrote "24 pin connector" I automatically thought "it's an ATX" and made a lot of wrong deductions starting from that.
|Sep 26, 2009, 02:54 AM|
generally, any PC style power supply these days is a switcher. As such, while they don't like to be unloaded, they usually won't burnout. Think about what happens when the PS get's plugged into the wall but not plugged into the motherboard. Pretty bad design if it burned out. Typical misbehavior when the load is too low is they will fail to reach spec'd voltage on their outputs. The usual trick is to give one of the voltages (usually 5V or 3.3V) a load. You'll need to experiment a bit since they are all different. I've had good luck with a 10 ohm, 5 watt resistor between the 5V rail and ground but you can get by with a higher value resistor. A server supply is probably designed for a heavier load so it might need to see more than an amp. Watch your resistor wattage - use ohms law to figure out the current and then multiply current times voltage to get watts. pick a resistor wattage rating at least 20% over that.
Google has lots of pages on this. Look for "pc power supply as bench power supply".
|Sep 26, 2009, 07:40 PM|
First off, I'd pull the case off and make sure I knew which wires are larger gauge wires that look like they supply major current, and which (if any) are smaller "sense" type wires. No way I'd try shorting out any two wires that are fat enought to carry any real current. But you might be OK trying to short some of the smaller gauge wires to ground.
On the PC PSUs that are so often converted, some wires are marked where they are soldered on the PC board, so you might find some clues there. If you're really determined to figure it out, buy an in-line fuse holder and a handfull of small amperage fuses, like maybe .5A or something, and experiment some. If you get two wires together that really don't like each other, the fuse should save your bacon.
|Sep 27, 2009, 01:52 AM|
Yes shorting wires randomly is a Very Bad Idea.
The only wire you should short to another is power good. There is no shortage of web sites that will tell you how to use a standard PC power supply. Use one of them rather than poking blindly. That usually results in letting the smoke out of the power supply. Electronics run on smoke and if you let it out they stop working.
|Sep 27, 2009, 09:16 PM|
Find an Old AT PSU
I have converted numerous AT PSUs to power supplies by removing the connectors and wiring direct to the 12V outputs. It usually nets 2-3 outputs and I use different connectors for my equipment, Loads can be soldered across the board low voltage power connections to keep a load on PSUs that require it but most AT's do not. I have used old CD library server PSUs and conventional 300watt PSU's. I have purchased old PC's in garage sales for 1-5 dollars with working PSUs. I have sold some of the old components on ebay and recouped my cost. Look for PSU's at least 200 watts. These are very reliable and the older the better. I have a chrome PSU 300watts that has powered my chargers and even some lights for late evening flying for almost 6 years. I have attached 3 multi-chargers and even a 24 volt unit in series.
I find these older PSUs much more reliable than the newer ATX units.
Causality is the Root of all Life!
|Sep 28, 2009, 04:32 AM|
Update on this
3d-harrier, sorry I wasn't watching this thread and missed your question. In short I never got this one working but it was possibly a faulty supply in any case. Had become "surplus to requirements" for a good reason at work I suspect.
You may be interested in a thread I started on a supply I have got to work which is much better. It's a currently easy to find HP supply that will maintain a bit over 12V at 118Amps with excellent regulation (a bit less in the US and other 110V supply countries). I included some thoughts about how to guess at startup configurations which might be useful. Latest postings are at
There are also several other threads around about different models of IBM and HP supplies that work.
There is a lot of confusion about this topic:
1) Most folks assume we are talking about PC AT/ATX type supplies for which there is a huge amount of info on the net. The comments below are quite appropriate for such supplies. The ones I have are proper Server supplies, not out of PC's sold as servers - quite different.
2) Poking about shorting pins is actually safe enough on the logic connector as long as you know what you are doing, use an appropriate current limiting resistor and sensible anti-shorting techniques. As always, if you don't know what you are doing it's best not to muck around with these beasts or open them. They are 1300Watt devices with mains potential and some nasty capacitors inside. The fans are also fairly noisy. :-)
Sorry can't be more help.
|Oct 01, 2009, 03:18 AM|
Joined Oct 2009
I have this power supply that I removed from a retired server today, hoping I could make myself a nice little bench supply. (Not related to RC stuff, but I found this forum Googling for how to make the PS work. So I figured I'd give back the info I have.)
After reading this thread and one other in this forum, I decided to try and figure it out. The logic that made the most sense to me had to do with the short pin. I think this is a common element on these server power supplies. The short pin is likely the "enable" pin and it needs to be grounded along with the "power on" pin.
If you're looking at the contact end of the supply with the contact row on top and the label facing up, there are 6 blade connectors on the left, then a group of 24 pins, then another row of 5 blades.
Pics for reference:
The pinout as far as I have tested is as follows:
Numbering from the left side of the 6 blades:
Blades 1 and 2: +3.3V
Blades 3 and 4: +5V
Blades 5 and 6: Chassis Ground
Numbering from the RIGHT side of the 5 blades:
Blade 1: No apparent connection.
Blade 2: +12V
Blades 3, 4 and 5: Chassis Ground
The grid of pins, numbered from top left to right:
1 Enable (Short pin. Ground this one...)
2 Power on (...and this one to make the unit work)
5 Chassis Ground
7 +5VSB (Stand By, always on)
Regarding unknown pins: I run a lot of these servers and I would expect the other pins are health and diagnostic pins. For example, fan rotor lock or RPM sensor and power supply temperature are both very likely to be in there somewhere. I also would expect possibly a feedback system to keep the main power feeds calibrated.
I can get my hands on a few more of these power supplies since we are retiring a lot of servers and recycling them. If anyone wants one, let me know and I'll see what sort of cost I can give you, just to cover shipping. There are a few different models, so I'll have to figure out each of the pinouts first.
|Oct 01, 2009, 04:16 AM|
Joined Dec 2004
Sorry I can't be of much help in your question on this supply BUT I see from your other thread that you found a real beast!!
|Oct 01, 2009, 06:00 PM|
Thanks, Mark. It's nice to have someone here who actually knows what he is doing! I got the HP 1300W supply working (see other thread) by trial and error and the same assumptions about the short pin. The nice thing about the HP supply is that it puts out virtually only 12V and the other rails are small, so the available current is huge (120A @ 12.4V on a 240V mains supply). These Compaqs seem to be fairly common though so for a 25A supply are a great find.
|Category||Thread||Thread Starter||Forum||Replies||Last Post|
|any info on REALLY small foamies? KP00 engined-30g or less?||natasha.b||Foamies (Kits)||6||Jun 18, 2009 06:04 PM|
|Can Anyone give me any info on the WINDY from NSP||CHRIS_00||Electric Sailplanes||3||Jan 29, 2002 09:21 PM|
|Any info on two Dymond planes?||DaleM||Electric Sailplanes||13||Sep 21, 2001 08:05 AM|
|Any info on this motor? Viper 650 neo||Heather||Power Systems||6||Jun 30, 2001 08:45 AM|
|battery charger as 12v power supply||sandman||Power Systems||6||May 09, 2001 04:15 AM|