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

Stielz said:

Could also get current reading through i2c on the supply, but I like the idea of a reliable external current sensor like the Allegro (and no stuffing around with calibration!)

Click to expand...

I was going to suggest tapping into the signal from the internal shunt resistor current sensor but if you can get that get current readings through I2C thats even better, no celebration required.

Click to expand...

I will try both and see which is more reliable. Better than the shunt may be to use the output of the IN-AMP already on board! I am worried that it is +-8V tho 8(

Stielz said:

I've also gone with the arduino controller for my charger, with the LCD-keypad sheild gives you alot of functionality. Coulomb counting is an important one for me so I can see how much energy I'm putting into the battery's.

Also, I found a MUCH more convenient place to get 12V off the power section such that you do not have to remove any of the PCB's except the backplane connector (Even then you dont have to unplug the 5v). I will post a picture soon, 1000 words etc.

Click to expand...

Keen to see this photo, I've got a side project using this PSU to power a bunch of 100W leds. Need 12V for the cooling fans

Click to expand...

Sorry for taking so long onthis, hopefully this afternoon. Relearning embedded c and coding up some examples are taking waaaaay to long!

At least the Proof Of Concept with a POT and a F!ing huge reverse protection diode from an old monster UPS is working. I use the Cycle analyst as my voltage/current/Ah(coulomb counting) meter. The CA is amazing!! Pretty good using a length of the cabling as a shunt too. +-1%ish

Doug found the kit in his glove box! Here we go lets do this.

Does anyone have high-quality inside shots of this supply?

I'm specifically interested in the output capacitance - and possibly the peak amperage (100A-200A peak) that may be used short term (20ms to 100ms).

Prototype board based on the schematics from the German site mostly complete. Now to fox the pwm->voltage (cap to small) and test the code.
I can get more detailed photos of the inside if need be.
Schematic+PCB+code etc available if wanted. Will document it all some time.
2x isolated i2c,
1x isolated rs233 (logic level to ftdi or whatever)
2x opto's for turn on
1x voltage out for PSU o/p control.
1x divider input to internal adc

Keep us updated!

I have been waiting for these to be made for quite a while and am excited to be able to bulk charge soon!

There are 6 on ebay for $70 it is a bid but no one has bid on it. If some one is in Cali and can pick them up It would be worth it. Otherwise $160 for shipping is extremely expensive.

http://www.ebay.com/itm/371161973628

Does anyone know if you can combine a 12 volt power supply with the ESP120 in series to get a higher voltage?

I was thinking about putting in series an ESP120 with two 12 volt power supplies to bring the voltage up to around 74 volts.

Is there any issues with doing this?

I am running 2 HP esp120 server supplies in series with a 12v to get 118v!! Works great. Soon I will build this kit to run the 2 server supplies as a stand alone system and charge up to 55 amps.

You can run any power supply in series you want you just need to make sure the outputs are isolated.

Arlo1 said:

You can run any power supply in series you want you just need to make sure the outputs are isolated.

Click to expand...

Thanks arlo1, that is perfect. How many amps are you pulling from that charger? That is a lot of watts!

Right now I'm pulling 50 amps for 2800 watts from the ESP120. Since I'm using the Max-E I have to keep the voltage below my battery voltage. I wouldn't mind charging even faster sometimes as I have a 2700 watt hour pack. This is why I want to bring the 50 volts up to around 75 volts.

Could you tell me what you mean by making sure the outputs are isolated?

Offroader said:

Could you tell me what you mean by making sure the outputs are isolated?

Click to expand...

Most power supply cases are connected to the negative output. This means that If you put two in series and the cases of each power supply touch then it will short out one of the supplies.

To Isolate a power supply you just have to isolate the area where the ground part(s) of the circuit board mount to the case shown here by Icecube.

You need to test all outputs in reference from one supply to the others with an ohm meter.

Hi all,
I've just started attempting to build a high power charger for my Geo Metro EV using 3 of these PSU's. I picked up 3 PSU's from Ebay for about $40 ea delivered to my home in Reseda, Ca.. My car pack is a 45S CALB SE100Ah LiFePo4 pack. I'm currently using a 10A EMC-2000 charger(2 1/2 years old & still going strong), but would like to charge faster :wink: . I use 8 cellogs to monitor & terminate the charge when any cell reaches 3.5V. At 3.5V per cell*45 cells=157.5 max charge voltage, the 3 PSU’s max voltage of 159.75V is good. However, when the pack is at 70% discharged, the pack voltage will be around 3.1V per cell *45 cells, or 139.5V, which is to low to keep the PSU’s from going into overcurrent shut down. So, the current plan is to Mod the PSU's to increase their voltage range by changing their voltage adjust Pot rating to 10K ohm(53-44.5V), from 1k ohm(50-53V). Then control the output current by varying the output voltage using a 6 channel digital pot, controlled by an Arduino & current sensor. So far I have only tested 2 units in series, with a lite 2 1/2A load. They worked fine, with both units pulling the same input current.
Updates to follow.
Joe

During testing today I found out that when these units are powered up, the DC output isn't completely isolated from the chassis. There is 2.4V DC to the case from either DC output terminal. When the unit is off there is about 36K ohms resistance to the chassis. So, I guess I will be separating the cases with some ABS strips when I assemble the three PSU's .

Nice work JoeG.

Have you tried running them with the chassis touching each other? 36K ohm wont allow much current through so it may not be an issue..

No I haven't tried running them with the chassis touching, and you are right that with 36K ohm there won't be much current flow. However, it won't be hard to isolate the chassis when I bolt them together, so I'm going to isolate them just to be safe. I hope to have some time to work on this project this weekend to do more testing.

Another thing to look out for is if the negative terminal is tied to earth.. If it is then you will get a short through the mains earth wire when running in series.

My ESP120 has a zero ohm connection from negative terminal to chassis and from earth to chassis.. Maybe there are different versions of this PSU

Edit: just re-read your post, you've already isolated the negative output and earth from the chassis?

Stielz said:

Another thing to look out for is if the negative terminal is tied to earth.. If it is then you will get a short through the mains earth wire when running in series.

My ESP120 has a zero ohm connection from negative terminal to chassis and from earth to chassis.. Maybe there are different versions of this PSU

Edit: just re-read your post, you've already isolated the negative output and earth from the chassis?

Click to expand...

Weird... I just wired mine up (after making sure it was safe) and run them in series. I never had to isolate anything. It just works.

On my ESP120's the mains earth is tied( 0 ohms) to the chassis. Both positive & negative DC terminals show 36K ohms to the chassis. This is how I received the units. As noted in a prior post, not much current will flow with that much resistance. It is good to know that they will run in series while touching cases. Thanks for the info.

I think if the 12V is on the low side, it doesn't matter as long as the unisolated 12V case isn't touching the esp120 case.. high side might be the same case it's just you have now exposed 100V pretty openly

Hi All,
I mounted the 3 PSU’s together using some 1” alum angle, 1” dia wood doweling for carrying handles, ¾” x 1” pine base, & 10-24 threaded rod to tie it together. I didn't end up isolating them, they work fine touching cases like Arlo noted in a prior post. They are pretty heavy at 13 pounds each, or 39 pounds for 3. I spaced the units apart with some 1/8” thick ABS for airflow to aid cooling. I wired the units in series and wired the AC power connections. Then I tested the inrush arching with 25 Ohm, 50W ceramic power resistors on each leg of the AC input. It tested with a very small spark, but the resistors got pretty hot, since they are undersized at 50W for the initial inrush of about 1156W each. There is about 7000 uF total input capacitance. I need to find a couple of 100W resistors for the final build of the pre-charge circuit. Did a low power test with the 3 PSU’s attached to my 500W stove top element resistive load. With no load they draw 1.5A AC. With a 4.5A DC load the AC draw is 3.9A. At 4.5A DC, the DC voltage drops from 160.6V to 160.0V. I also tested the anti-back drive diodes by turning off the center PSU during the 4.5A load test. It passed with no smoke or sparks, just turning back on and when I turned in on, and it continued working. During battery charging, if one supply drops out, the voltage will drop to 106V, and the current will drop to nothing. Because of this, I think I will skip the high Ampere resistive load turn off testing, just so I don’t stress the diodes(70A, 200V) needlessly. I'm working on the Arduino program to control PSU's and need to finish building the 3 ohm load for high Amp testing.
Updates to follow.

Nice work Joe.

Does anyone know if there is a difference between the Compaq ESP120 and the HP ESP120? They both have the same number for the HP part number and Compaq part number, same spare parts numbers, and the specs are the same too. The only thing that I see that is different is that they say Compaq instead of the HP logo.

I just want to make sure that this one will also work with a limiter board.

Thanks,
Scott

should be the very same one. sometimes they don't even distinguish between the two brands that become one.

Sorry I didn't write sooner, Firefox does not let me login on the forum for some reason...
I have completed a pretty basic proof of concept to get the PSU's charging my car batteries. It does no error checking etc. at this stage, but has been working remarkably reliably for the last few months!!

JoeG hopefully there is something in here you can use. It would be of great interest to me to work collaboratively on this -

The circuit I made was designed to be fairly general and test various things. I2C worked well enough that I did not want to go pulling out the voltage from the shunt or the opamp on the PSU. Basically aiming for minimal inteference with the ESP 120.

Attached is the code as is ... it is still a WIP. If you want to contribute I can give you my personal mercurial repository address etc..
The platform is Arduino Nano using avr-gcc.
The TODO list is something like:
Better LED indication of state
Error condition checking (under voltage, etc)
BMS interface
GUI/bluetooth-android integration



I hope this helps somebody...

Brilliant work 5pmp!

Now any power level of off board charger is dirt cheap.

liveforphysics said:

Brilliant work 5pmp!

Now any power level of off board charger is dirt cheap.

Click to expand...

Cheers man, Thats the idea 8)
And help would be appreciated to refine this to a really solid, extensible controller. I think the principle is at least sound.
SOmeone on here was building their own controller and gave up, but had a really nice we GUI thing going on, maybe a frankensteins monster is in order. CAN would be great 8) and currently working on a bit banging serial reciever so the charger can get data from my Cycle Analayst (v2.3) such that if charge ends it can check whether this is premature.

There is one oddity: Every now and then (three to four times in total, so quite rare) I have not been able to get a high enough voltage to complete the charge. I think this is easily overcome by cranking the voltage on the 'static' psu.

BTW, there is a MASSIVE diode on a big heat sink protecting the PSU's up near the battery connection, so obviously the voltage drop has to be taken into consideration etc, and is a noticble function of A at charge currents... but, well, not worth worry about methinks 8)

Happy New Year!