52v 14 lithium cells battery simulator for BMS

[dansku]

Hello guys, I am currently testing one BMS hardware, but I don't want to use real lithium cells, instead I want to use a power supply (or multiple suppliers). The BMS can check the status of each cell, therefore I would need 14 power suppliers that could be controllers somehow and stacked into tricking the BMS that we have actually 14 cells connected. Before building this myself, I would like to know if anyone knows if there's such thing on the market already?

Thank you

[dansku]

Explaining it a bit better (forgot to add this info).
I don't really need a powerful PSU as what I am trying to do is mostly communicate with the BMS via CAN.
So there will be almost no current going out of it. I need to read for example the voltage on the cells, etc, which the BMS will provide via CAN. Currently I am checking the possibilities. Maybe I just get 14x voltage regulators and hook to one power supply.

[dansku]

What about having one PSU and 14 zener diodes?

[nctnico]

I have built something like that a while ago for a similar purpose. It's not cheap though (count on a couple of k euro) and if you want to make each cell voltage adjustable it is even more expensive (but still cheaper than 14 seperate power supplies). So I guess something is already on the market 

[dansku]

I have built something like that a while ago for a similar purpose. It's not cheap though (count on a couple of k euro) and if you want to make each cell voltage adjustable it is even more expensive (but still cheaper than 14 seperate power supplies). So I guess something is already on the market 

Can you give any idea on the way you did it?

Thanks   

[nctnico]

I've used power opamps from TI and a resistor ladder fed by another power opamp which is controlled by a DAC. A microcontroller which talks through an isolated USB interface to a PC controls the DAC and does some other checks. The neat thing about the power opamps is that they have thermal and short circuit protection. An output enabe/disable makes it safe to work with as my version goes to 96V max to support up to 20 cells in series.

[station240]

What about using 14 super capacitors instead ?
Of course then you still need a way to charge them to the correct voltage, but that could be done with switches.

[2N3055]

Google "battery simulator".. dozens of them. You would have to give it 14 separate isolated PSUs. And those demonstration designs are analog, with pots to twiddle, so maybe digital pot and I2C isolator could make them adjustable... If you need low current all that wouldn't cost much (few hundred) but is in itself not a beginners project..

You need a PSU that does source and sink, so in other words, a source measurement unit (SMU). These things are not cheap.
Prepare ~$1000 for each, and you need 14!

To be a bit pedantic, technically all you need to simulate battery is two a quadrant PSU.. But of course you are correct, SMUs are usually four quadrant and precision measurements , so they have even more than is needed and they would do the job fantastically.........

[cowana]

You can get isolated DC-DC converters very cheaply, which will output 5v @ about 200mA.

If I were you, I would design a small board that takes in 5v, isolates the control and output voltage via isolated converter, has a small micro and some control circuitry to set the output voltage. You would tell the micro what voltage to output via serial, using an opto to cross the isolation barrier.

Make up 14 of these boards, stack them in series, and use addressing to set the voltage of each 'cell' as desired. If you don't need to sink current (simulating charging), it's a very simple and elegant solution!

[dansku]

You can get isolated DC-DC converters very cheaply, which will output 5v @ about 200mA.

If I were you, I would design a small board that takes in 5v, isolates the control and output voltage via isolated converter, has a small micro and some control circuitry to set the output voltage. You would tell the micro what voltage to output via serial, using an opto to cross the isolation barrier.

Make up 14 of these boards, stack them in series, and use addressing to set the voltage of each 'cell' as desired. If you don't need to sink current (simulating charging), it's a very simple and elegant solution!

That actually sounds pretty good, thinking about using 14x of http://uk.farnell.com/murata-power-solutions/mtc2s1203mc-r7/dc-dc-wandler-3-3v-0-606a/dp/2579612?ost=MTC2S1203MC&selectedCategoryId=&categoryNameResp=Alle%2BKategorien&searchView=table&iscrfnonsku=false  with a potentiometer for the output voltage, this way I can simulate each cell.

Thanks

[cowana]

I was thinking of the super cheap modules - they'll do what you need!
http://uk.farnell.com/murata-power-solutions/cre1s0505sc/dc-dc-converter-5v-0-2a/dp/2564440

A potentiometer will work if your BMS only draws a tiny current - anything significant (talking hundreds of uA) will cause a varying drop. A slightly more robust solution would be to use an adjustable voltage regulator (such as the jellybean LM317), using the potentiometer to set the output voltage. That will let you draw close to the capacity of the DC/DC converter from the 'cell'.

[nctnico]

An adjustable voltage source may work especially if it can sink some current as well. A potmeter may not work depending on the BMS because the internal resistance of a Li-ion cell is tens of milli-ohms. Some BMS check the integrity of the connections to the cells by drawing current from them through the balancing circuit. A large voltage drop could trigger a fault condition. The same goes for a large voltage difference between some of the cells. Still using a stack of adjustable voltages sources may work so it is worth a try but I'm not sure it will be cheaper than using power opamps.