HP ESP120 48V (51.4V) 57A 3KW Server Supply Thread

ZOMGVTEK

100 kW
Joined
Jun 5, 2009
Messages
1,090
Here's a place to discuss the HP ESP120 server power supply. Modifications, uses, sources, etc... Throw it in here!

6244576005_2932d87d55_b.jpg


6244576555_5e5e542171_b.jpg


I got only one to mess around with. These claim to require 200-240V input, at 20A! I will see if I can get it to power on at all off 120, but it would take at least a 30-40A breaker to get its rated power at that voltage...

First order of business is to determine where the power goes. These do not have an IEC on them. I would assume the top terminals are the mains in, probably ground in the center, but I can test continuity between the chassis and pins to confirm.
 
There are diagrams all over for this server supply.. and activating it. you need 220.

http://www.cnczone.com/forums/general_electronics_discussion/112704-hp_server_power_supply_51_a.html

I have esp 115s, and with a small modification to the pins you can achieve an additional 1.3V I think. Maybe the same is true for this guy...
 
Can you check if the In/Out is isolated ?
Because having the possibility to stakc them in serie for HV chareging electric car qould be interesting!

Just the need to add a 100A diode in parallel to each PSU to protect them and you get a cheap FAST EV charger

Doc
 
Yeah, My initial post about this was in my thread on ev charging stations.. I want to put one of these on my bike rack and get J1772 going and 10 minute recharges. ;)

I want to know what this double pole/neutral fusing means:

IMG_5825.JPG

IMG_5828.JPG


"Connect together the three pins inside the red rectangle to activate the DC output."
 
I assumed these supplies would be well documented, just didn't look around yet.

The DC ground goes not go to the chassis on these supplies. It looks like you should be able to get away with running them in series.

I can't see any way to raise/lower the output voltage. There is probably a way to do so...
 
sorry, I meant esp114.. I get 50A out of it at 13.2V I think

Sometimes there are some 0 ohm resistors that remove the ground to case.

I can show you the diagram to the esp 114.. it has the same pin config..

From RC groups regarding esp114:

" Basically, it takes 2 things to max out the voltage increase. The max I could get is about 13.3V before it shuts down for over voltage. Below is a pin matrix of what to do. Basically, you short pins D1 and D2 together and put a 230 ohm resistor between C5 and B5. If you don't short D1 and D2 together, you will only get a voltage increase to 12.67V.

--1 2 3 4 5 6
D S S 0 0 0 0
C 0 0 0 0 R 0
B 0 0 0 0 R 0
A 0 0 0 0 0 0

S = Short
R = 230 ohm resistor

With a 30A load on the power supply, it holds a steady 13.3V"

Hope this helps. I guess it is a littl ebit different config.. but here is the enable:

DL580PowerSupplyPinouts.jpg


someone posted these instructions on rcgroups so find the pins:

"Finding the voltage control pin is a process that also has the added benefit of revealing other important pins.

Attach a voltmeter to the output of the PS.

Record the base voltage reading. For example lets say +12.2v(+V).

Through a 750Ω - 1kΩ resistor, connect the +V output to each pin one at a time. This saves more time than having to find all ground, +3v and +5v pins.

Watch your meter as you connect to each pin to see if the +V output changes.

----------------------------------------------

If you get a +V change from only two pins than you've found both the

A: The +12v current share pin

and

B: The Voltage control pin. (+12 remote sense return)

The pin that changes the voltage the most will be the +12v current share pin.

The pin that changes the voltage the least will be the voltage control pin.

-----------------------------------------------

If you get a +V change from only a single pin then you've found either the

A: The +12v current share pin.

or

B: The Voltage control pin. (+12 remote sense return)

Repeat the process with a lower value resistor. Lets say from 500Ω to 749Ω.

If the +V change is the same as when connecting the first resistor than you've found the +12 current share pin.

If the +V change is different than when connecting the first resistor than you've found the Voltage control pin.

Always use one of the +3v or +5v pins when finalizing the main PS output voltage adjustment as this enhances voltage regulation.

-----------------------------------------------

In some instances you may also find the fan speed control pin. A minority of power supplies have this pin internally connected to ground.

Introducing a positive voltage to it will cause the fan to speed up.

The +12v current share pin can be used the fully regulate a PS as in post 470 or to vary fan speed in proportion to power

output as in post 461.

Hope this helps.."
 
For 12V I really like the ESP135's... Theyre rated at 47A @ 12.15V, and will go up a few volts with mods, if need be. Simple IEC in, bend 3 pins into each other to turn on, and simple outputs.

6199800836_cb3bce1736_z.jpg


6199798968_f4d1f45d04_z.jpg


I've managed to get 55A out of these for a while, without any issues. People claim they shut down if you run them past 47A for a while, but they DO have big exposed shunts you can solder up and probably get 60A safely.

I have 2 more on the way to throw power at a big mess of lead off a cheap generator. The lead will have a 100-200A peak load on it, so the supply would need to be capable of some serious current to keep up. I will see how well these supplies can handle the VERY messy power off the generator... I wouldn't be plugging any high end equipment into it, but a $15 PSU? Why not!
 
no crap ZOMGVTEK TYVM I just grabbed me one of ebay.. will check it out then probably get some more.
18 bux free shipping can not beat that. Your setup is pretty sexy.
 
They sure are cheap, and they sure are great. Very high quality supplies in a compact form factor, and they happen to be REALLY cheap. I got mine for $12/ea shipped, but I got 8 at once. These supplies have substantial output capacitance, more than any Meanwell or similar. And they handle rapid transients VERY well. All around, impressive supplies that were clearly designed for hard use.
 
I would like a high voltage, smaller form factor psu I can put on my bike rack. maybe 900-1500W 51.4V That would be wonderful. sub 8 pound...

post a picture of the esp120 near where your rack would go on your bike.. to get an idea of the size, please.
 
Its a VERY large, rather heavy supply. You do not want it on a bike. 3KW continuous rated power is no joke.

I can take a pic of it next to my bike, but I don't think you will like what you see.
 
If it is like my 3KW-capable Sorenson DCS 55-55, it is about as big as my front hubmotor wheel on CrazyBike2, and quite possibly twice as heavy. :(
 
Sweet time for this thread :mrgreen:
 
If you painted the casing black, people might not even notice this thing is there!

6246900520_fc97c32d3f_b.jpg


My bathroom scale claims it is 13.5Lb, which might be on the low side. It's definitely larger than it is heavy...
 
Especially if I made a wooden rear crate out of it and hid it inside. Thanks for the picture. It's huge, but on my 29er, it will fit right in.

Please disassemble it an post pictures of actual dimensions of components. Maybe it can even be separated into two smaller halves (???) or modified to take up smaller footprint.

I bet there is a 5-7mm gap going all the way around to separate board from metal enclosure.

I might lose a half a mile of range, but I will no longer need to lug a 70-100 pound bike up 4 flights of stairs (to apt.) to charge, and I'll theoretically be able to charge in 10-20 min.
 
Well, it looks like there is allready a good start on this supply over here...

https://www.endless-sphere.com/forums/viewtopic.php?f=14&t=29316

I'm tearing it down now. I will be powering it up and tripping some breakers shortly.
 
I should be getting Arlo's and mine in on Monday, so I'm keen to see what you guys come up with. I'm gonna use it for charging at work on 220v in like 10 minutes. :mrgreen:

Can you show us how you got the voltage up to 51.4 volts? Have you verified that it charges with no hiccups?

I need more batteries to spread the bitch slap out that this charger promises to deliver to any lame cells. :lol:
 
6248037807_7ba4c751b8_b.jpg


6248559434_1742e79264_b.jpg

That little pot appears to do nothing to the 48V output.

6248557490_1f382a753f_b.jpg

These fans look like they sure can get loud...

6248034755_8d07e6f413_b.jpg

AC in board. Fuses on each phase, caps going to ground on the two screws to the right. The only 'grounded' portion of the case is the metal section this board screws into, which in turn grounds the chassis to AC ground only.

6248038721_942755d02c_b.jpg

DC out board. The thicker red/black wires with the ferrite appears to be the 5V rail. There's spots where caps can go to run the DC ground to the chassis, but they are absent.

6248036873_7a5ddec5f6_b.jpg

Lots of stuff going on here. The pot on the top was covered in some silicone and did nothing to the output.

6248562916_b55493c18c_b.jpg

Fully unloaded output voltage. A few hundred mV of ripple unloaded.

The lighting in the basement always throws the white balance off, these will have to do.

The AC inlet happens to be conveniently labeled L G N right on the case.
I don't think I have any real use for a 48V supply, especially one that needs 240, so progress might be slow on modding this unit.
Overall design appears to be very good. There's probably a way to at least marginally adjust output voltage. It probably only goes up, however.
 
Bear with me guys, so say you supply is cc/cv and it tends to taper off near the end of the charger. I wonder if a control board might be made that could just trip the device to shut off prematurely at 50.4 v? Could an arduino chip module be made to do such a task? It seems like a work around that I've never seen used, but it would work for my application.

Great pics by the way, thanks for posting those, that thing is HUGE too huh? I will have to get cracking on wiring up 220v if we find a way to adjust voltage so I can safely charge at home.
 
It looks like there is a separate board for DC and AC. Past those two big transformers the board is separated, so you probably could reconfigure it a bit. However, the fans sure look like serious business, so this thing must really need some airflow at high power.

I don't have any 220V outlets here as well, I just shoved the romex into two adjacent breakers that had useless stuff on them. Probably not the best idea for long term use.

If you had a CC CV charger, you definitely could configure something to kill the supply at a preset voltage. However, there is no real use, since it would be best to have the output voltage set to exactly what you want your cells to be charged up to. I would imagine this supply would definitely not drop voltage to maintain its current limit, and would need something external to control the current for direct battery charging. If you had a CC CV charger, there would be little need to shut it off immediately after reaching HVC, so it wouldn't be really even needed. It would be dangerous to rely on using an external board to turn the power off when you are charging at such high power, and the output of the supply is above your intended HVC.
 
Thats great that it separates. Couod u measure each section.. Lxwxh?

Yeah,, I was looking to use arduino. I would have it check voltage based on the result of a function that determined state of charge. Charging would have to be stopped to take measurements. Easiest way is to measure one cell...hooked up straight to board. I would only go to 90-95% soc

I could also use common sense and tell the arduino to check more often as the voltage rises and stop when it hits 3.45V No need for function.. Just case statement

For 14s 51.4 would be ideal to get a higher overall charge time.. Something higher might even be better

It would be pack specific, and there would have to be a scalar in to change every 200 cycles when pack capacity has dropped.. (to change increments between voltage checks)

I will try this with 7s/25.5V supply and let u know how it goes.

$50 3kw cut-off charger?!
 
The problem here is you would need a pack voltage very close to the output voltage of this supply. It is designed to NOT drop voltage under load, so you would need something like thin wire that takes the abuse. You might be able to get enough voltage drop through wire so that it will work, and somewhat taper off towards the end of the charge which would help you read the cell voltages. Ideally you would want to monitor every cell individually, if not, just use the pack voltage. Something is going to be dumping a lot of power as heat to make this happen, but its possible and works OK.

Definitely shoot for 85-90% SOC at these charge powers. It would be really easy to overshoot your charge voltage pumping 3kW into a <<3kWh pack... Things can get dangerous at these power levels.
 
Back
Top