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

I'm sure Arlo can post photos of the kit and its operation to help guide you to make your own version of it specially suited towards the needs of being a wide range of use high power battery charger.
 
liveforphysics said:
I'm sure Arlo can post photos of the kit and its operation to help guide you to make your own version of it specially suited towards the needs of being a wide range of use high power battery charger.
That's the plan. I will need help from the forum. :)

I got my 50 amp supply and it just keeps resetting. So now I will all need this.
 
Arlo1 said:
liveforphysics said:
I'm sure Arlo can post photos of the kit and its operation to help guide you to make your own version of it specially suited towards the needs of being a wide range of use high power battery charger.
That's the plan. I will need help from the forum. :)

I got my 50 amp supply and it just keeps resetting. So now I will all need this.
Cant wait for the pics etc! <fidgets excitedly>
Having huge trouble finding this PSU in New Zulund.... nope! just found a whole stack by fluke, a bit more than $20... but in NZ cant complain 8)

Want to get started on control logic for a charger based on this, but need to see how we are interfacing with the cc/cv module. Is there an existing project that could be modded to this purpose (best option methinks) or start from scratch? (New thread???)
We are starting a project here at our elec. group trying to figure out if we can use a mini ralink module (http://www.aliexpress.com/item/Mini...ap-ethernet-port-wifi-hlk-rm02/798875078.html) running openwrt as a realtime (ish) controller in place of an arduino as it has wifi, ethernet, I2C etc etc... great for phone and remote interface. But thats OT...

chur.
 
Buddy Doug is trying to find them... He lost them in all his stuff he hauled for racing and had to deal with a motard bike in a prius with a broken collar bone when he got home.... :)
 
Poor bastard! Broken collar bone sounds sh!t, less fidgeting, more sympathy

Thanks to Stielz:
http://elweb.info/dokuwiki/doku.php?id=hp_3kw_netzteil_als_lader
I think its all here except the source code! Sooo sweet! I2C with a huge address space (Up to eight devices!) and a full data sheet <3

Does anyone read German natively, google translate is not that great.
If the source code is not there I think we should contact them and ask for it, they claim its all available under BSD license. I can see the I2C code, from prelim (very) scim looks like max current is set here, yes, yes it is... too good! (I love I2C! My solar hotwater controller is all I2C peripherals, on 5m of cable, and it still works, even though it practically shouldn't)

... and this was meant to be an early night 8)
 
izeman said:
@spmp: post whatever text you like and i'll trabslate it. i just don't have the time to translate everything as i'm on a ipad and this is no work environment.
Many thanks for your offer!
At this stage annotating in English (I think Foxit reader is available for iOS) the circuit diagram PDF's and if you can have a quick look at the wiki page and see if you can see reference to
a) Current control
b) Links to actual source code for the uC

PDF's:
Isolated
http://www.inexess.com/bilder/hp3kwlader/hp_3kw_lader_prototyp_v2_schaltplan.pdf
non-isolated
http://www.inexess.com/bilder/hp3kwlader/hp_3kw_lader_prototyp_ohne_wandler_schaltplan.pdf

Wiki (as above message)
http://elweb.info/dokuwiki/doku.php?id=hp_3kw_netzteil_als_lader

Could you please be the diplomatic leiason and write to the author of the ewiki page requesting the source code for v1 and v2 shoudl the source not be available in the wiki?

I am interested as to why they dropped the I2C support in the isolated version. Isolation is a damn good idea, and nesc. if we are to gang more than one in series. We will need a board design that allows for extenisbility at least up to eight supplies.
Personally I think adding I2C back is a good plan, using an I2C isolation chip.
I want to use these as a stand alone charger, but I see that many of you are using them as sources for th iduo etc. so I hope I am not on a one man crusade here. But to date this looks like the cheapest and most reliable route to LiFePO charger of descent magnitude!

Cheers!
 
I have translated with the help of google the best that I can one of the circuit diagrams (I believe V1)
 
The other circuit diagram, V2, which is probably what Ario1 has.
Please correct my errors if you can. I have used Xournal which I believe has a windows version as well. Can post the xoj files if nesc.

Some comments on the schematic, please correct me if I am wrong:
Should R7 and R8 be between K7 and K2, and the output go through an isolation amplifier for true isolation?? Otherwise the reference is floating right?

I still cannot see how this circuit achieves a CC mode on the supply, the potentiometer in the circuit just creates an analogue voltage for the adc on the micro (ie one way to set an input).

Any quick methods for setting the voltage digitally and maintain the isolation? Isolated DAC? digital potentiometes (but isolated??)
 

Attachments

  • hp_3kw_lader_prototyp_v2_schaltplan_Annotated.pdf
    383.3 KB · Views: 272
izeman said:
spmp, sorry i can't help at the moment. i haven't found a way to edit/annotate those pdfs on my ipad. i tried to open the ones you made and can't read the red text. it's only red rectangles, seems to be a non standard font.
Sorry, I did it in Computer Modern font from LaTeX, I changed them both to Times New Roman, should work now.
 
Izeman... damn, It should have worked, will try again some time soon.
Whilst I am trying to figure that our could you please annotate the jpg circuit diagrams into english 8)

I shall try to sum up what I have understood from the wiki page:
The output voltage of the ESP-120 can be adjusted from 24-63V by changing the potential where the 'trimpot' should be, with the rough transfer function:
Code:
12V on Potipad ⇒ 0.4V output voltage 
 6V at Potipad ⇒ 24V output voltage 
 5V at Potipad ⇒ 29V output voltage 
 -0.6 V at Potipad ⇒ 56V output voltage
(Thanks google translate 8)
I am unclear yet what must be done to get the output voltage to 63V, and whether any extra modifications need to be done.

Constant current mode is achieved by modifying the output voltage to maintain a constant current, so its not CC in the sense of a benchtop power supply :(

There is a full I2C spec for the PSU.
You can get full information including ouptut current, inlet temp, outlet temp, input voltage etc etc...
And you can turn on/off the output and put the fans into 'fast' mode.

For more than one supply in series they are a little uncertain, except to say that you need to protect against reverse voltage with a 60A Schottky diode.
There is not a lot of info on this. They suggest that the cases may need to be ungrounded, but we have seen that this is not the case (right?)
There also is a link to a pdf 'Load sharing signal' which looks interesting but does not load for me:
http://www.inexess.com/bilder/hp3kwlader/single wire sharing bus.pdf
Can anyone find the original or know what this is???

So I think there really is everything there to get started on making a LiFePO4 charger out of these things.
 
I have a preliminary idea for how I want to make the charger, FEEDBACK please!!!
If anyone else is interested in contributing I will post a link to the bitbucket repo.

I will do something very similar to the V1 board using Arduino Nano and C/avr-gcc etc.

I will utilise I2C as much as possible as this gives the ability to access multiple devices behind one isolation point.

Each supply will have two I2C isolators, one for powering on the output and talking to the processor (ie getting temp, current etc etc), and one for modifying the voltage at the potentiometer using an I2C DAC connected to the power side 12V rail.

For my batteries (and I am sure yours) 100% DOD is 3.0V, so for my string (28) I need 42V minimum per supply so a +-5V DAC is fine
I dont need the -ve voltage as all I have to go to is 100.8V, but may be good for others, or to push the full 6kW into the batteries 8)
Any ideas how to get a slightly negative voltage simply without a dc-dc?

I will use the Allegro hall sensor for current measurement and well... probably an I2C ADC for voltage measurement via a divider on the first supplies I2C and 12V.

Interface via serial for now as I suck at LCD's, so probably want a push button start/stop/ and probably a switch at the power socket for 30A or 20A charging depending on where I plug in (need to be under 2.4kW on a normal house circuit in NZ)

With uC control we have heaps more freedom on charging program, is there a better way to do bulk/absorb/float charge to prolong the battery life?

... and is this the right thread, should I delete this message and repost somewhere else?

Cheers
 
Nah its all good.

The whole design is a bit overkill with the microcontroller, I've designed current limiters for the HP supplies that use the stock shunt (so no need for expensive current sensors) and a cheap optocoupler for isolation. Using a sledgehammer to crack walnuts etc.
 
heathyoung said:
Nah its all good.

The whole design is a bit overkill with the microcontroller, I've designed current limiters for the HP supplies that use the stock shunt (so no need for expensive current sensors) and a cheap optocoupler for isolation. Using a sledgehammer to crack walnuts etc.
OK, this is what is being sold as the cc/cv kit:
http://www.elweb.de/coppermine/albums/userpics/10635/HP-Lader_ver_1_1.png All analog and pretty simple.

But I am worried with two supplies in series that it may not work so well. So for the simplest SAFE option for charging my precious LiFePO's (need 100.8V@30A+)
I figured go micro as its the most extensible and to me seems the safest 8)

Just killed one of my supplies 8( not sure how. Took it apart to get 12V, and remove trimpot and get I2C out... now when I connect all the pins to tunr on the green blinking light turns to the solid orange exclamation 8( I guess this is why I bought three. But weird as I cannot see anything that I did wrong. removed all the wires I soldered on, checked for solder balls etc etc... Maybe the dreaded ESD 8( 8(
 
Yeah thats pretty much what I would have expected.

Wont work well for series ones, since they do weird crap when they fight each other. You need to use the design I have with the optos to get them in sync. Also use the internal shunts so you don't have to use the expensive allegro current sensors - I should get one of these supplies to do some experimentation with.
 
I run them in series strings all the time.

I cut the ground wire on the AC input and keep the device chassis at different potentials from shorting against each other with a sheet of plastic between them.

For your pack you don't need CC/CV of both supplies. One can be stock and simply cut the ground wire and isolate chassis.
 
liveforphysics said:
I run them in series strings all the time.

I cut the ground wire on the AC input and keep the device chassis at different potentials from shorting against each other with a sheet of plastic between them.

For your pack you don't need CC/CV of both supplies. One can be stock and simply cut the ground wire and isolate chassis.
I run my in series with the grounds connected :) In fact all three have the grounds connected.
 
Arlo1 said:
I run my in series with the grounds connected :) In fact all three have the grounds connected.
Thanks Arlo1, that is what I had heard you can do too, its a massive plus 8)

Thanks Liveforphysics, I just assumed that I needed both supplies exactly matched (ie Vout/2), but of course I dont... duh! yes that simplifies things massively, now only need three i2c Isolators and one DAC. I2C bipolar voltage output dacs are still eluding me.
For my LiFeP04's what VPC will I need to go to in order to pull 60A?

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!)

Just discovered the existence of resistor-divider-IC! 0.1%, if only they were not $8.

And the shamelessly lazy question: Is there a definitive link or info to The One True community tested charging cycle to keep these babies alive and kicking for 15 years?

---
In relation to the German Wiki, The guy who made V1 boards with Atmel micro says the source code would be useless to us as its written in perl for atmel (Luna), and Chris says that the circuit on the wiki is the one he is using (I think it is V2.1, not V2.2) with no component values, so what you have Ario1 may still be worth breaking down.

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.
Thought I broke one of mine as the orange light came on after I fiddled with it a bit... didnt check the output, just assumed it was dead, so I really got into it 8) got the board with the pot on it out, without damaging the main PCB... aluminium tape and a paint stripping gun 8) (I was trying to find a 12V near the top edge... failed)
Anyway it turns out the next supply I played with (this time no disassembly required) works fine, but the yellow light comes on (fault), so I probably didnt break the first one, just never checked the output :oops:
 
I tested all with a multi meter and found the grounds and + and - to be all isolated from each other.
You should never trust what I say and check for your self :)
But in the video and pictures all the grounds are tied together to the ground of the wall outlet I did this as a safety thing and I don't know a lot about it. But my youngest brother give him shit for not having an extension cord with ground on it to run the mean wells I put together for him to charge his bike.
 
heathyoung said:
Nah its all good.

The whole design is a bit overkill with the microcontroller, I've designed current limiters for the HP supplies that use the stock shunt (so no need for expensive current sensors) and a cheap optocoupler for isolation. Using a sledgehammer to crack walnuts etc.
Do you have a link to the current limiter you designed?

Guys I'm interested in understanding how these work better.
Maybe we can design our own boards.
 
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!)

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.

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.

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
 
The thread I referenced - http://www.endless-sphere.com/forums/viewtopic.php?f=14&t=47415

Originally designed for the 12V units, since their adjustability is poor at lower voltages (the LVC on them is very sensitive).

These 48V supplies have a greater range, but if you use a chemistry that has a high differential between charged/discharged - you need to adjust more than one.

I'll have to find my notes, but I made something to detect single /3 phase power input to these when using 3 of them and adjust the current output accordingly.
 
Thanks for posting the link. I will be learning about this. As I bought this And found It doesn't have amp control just a CV supply. :( Not to mention it would have not worked < 105v anyways.
Time to do some digging.
 
Back
Top