Charger for single cell cycling before battery pack build

[bexander]

I have recently acquired two Tesla Model 3 Long Range modules with the intent to extract the cells and use some of them to build a light weight 9kWh pack for my Clio. In order to do this I need to cycle (charge, fully discharge, fully charge, storage discharge) a lot of cells (~2000pcs) to measure their capacity. In order to do so a charger is needad and there are some available that can do this, for instance SkyRC MC3000. Problem is that the Tesla cells don't have a protruding positive pole and hence the risk of short circuit to the naked negative pole (battery case) is imminent. Also, they cost a lot and only have 4 channels.

So I have designed a simple analog circuit that acts as a charger or as a discharger with some modifications.
My intent is 1,6A charge-/discharge current with 4,2V CV and 3,0V cut-off.

Charger:

Charger

Screenshot_2023-06-22_16-59-21.png (6.21 KiB) Viewed 666 times

Discharger:

Discharger

Screenshot_2023-06-22_17-01-11.png (6.89 KiB) Viewed 666 times

The designs are very similar but differs in how the battery cell is connected. I would prefer if I could switch between charging and discharging using the 328p and not having to physically move or reconnect the battery cell. Switching the voltage reference can be done using analog switches, I think.

Any ideas on how this would be possible at a very low cost? I have considered relays but costs to much (>€2). Or is there a better (simpler, cheaper) design of the circuits that will do both charge and discharge?

One 328p have 8 analog port so the idea is 8 units on one pcb and then maybe 5pcs of pcb for a total of 40 channels...

[Pete9008]

Is the aim to terminate the charge/discharge without using the 328p so that it's safe regardless of what the firmware does?

If you are happy letting the 328p get involved with controlling the charge then I'd be tempted to make a bridge but using your current source circuit top and bottom with the micro deciding which one is active. The advantage is that it could then be left to do multiple charge/discharge cycles.

[bexander]

When acting as a charger I want CC/CV which is the case with the above circuit.
As a discharger it is just CC and then cut-off when 3,0V is reached. Op-amp acting as open-drain comparator.

I think it is possible to have both op-amps connected to a single n-fet since they are behind a diod they are just or:ed together.

[Pete9008]

Yes I get that but would you be happy doing the CC/CV bit in firmware or do you want it all in hardware?

Like the way you have done the hysteresis on the discharger, guessing it was cycling as the cell voltage recovered?

Edit - think I see why you have done it that way now, it allows the micro to measure/integrate current on both charge and discharge without a polarity change.

Edit2 - ignore me, the 328p doesn't have the resources to do the type of thing I was considering (it could only handle 4 cells).

[bexander]

I prefer to do the regulating bit (CV part) in hardware. Less risk of unintended behaviour that way.
Will also try to add a temp switch as an extra protection layer.

I haven't tested the hysteresis part yet but relized that the voltage will increase in the cell as soon as the current draw is removed.

This is a relay solution, but need to add some kind control of the op-amps when going from discharge to charge, as otherwise it will stay off.

Charger/discharger

[Pete9008]

Thinking about it if you lost the hysteresis wouldn't it just do CC/CV on the discharge too?

Could you add a FET across one of the voltage ref biasing resistors to force the op-amp to the inactive state?

[bexander]

Good point!
Actually, removing the hysteresis will not only make it CC/CV on discharge but also remove the need of any extra op-amp control. As soon as the discharge is terminated from the 328p the cell voltage will rise and when switching to charge mode it should not interfere.

Doing it this way should still be safe as the cell can't go below 3V and the 328p can detect current drop to terminate the discharge at any current lower than 1,6A.

Added "store discharge" by adding optional voltage division on the discharge feedback voltage. Should increase voltage threshold to 3,6V.

Charger/discharger

[uhi22]

I use a simpler design to check cells: Charge them to 4.15V, multiple in parallel, not measuring the capacity, but checking the temperature from time to time. Sort out warm cells. Separate the cells, leave them for a day and measure the idle voltage. Sort out the cells which dropped below 4.1V. And then use a resistive discharge circuit, which is switched by a mosfet, and uses two ADC channels to measure voltage and current. Using pcf8591 AD converter with I2C on a Raspberry.

[uhi22]

[bexander]

That's a simple approach but I don't like the idea of connecting multiple cells in parallel that might have different voltages, would need to measure cell voltages before but maybe not to bad in my case as most of the cells should be well balanced. Do have potential to save a lot of time.

It don't know why, but my original plan was:
charge - discharge(log) - charge(log) - store discharge
If I do as uhi22:
charge - discharge(log) - store charge
There is a saving of 3h/cell with no loss in the result as I still discharge from full to empty, which really should be the one to look at.

Very good suggestion!

[bexander]

Latest design, charge them all in parallel via "Cell1" connection (includes cell current fuse) and extrenal RC-charger.

Discharger

Screenshot_2023-06-25_19-07-00.png (3.96 KiB) Viewed 529 times

Basically a copy of uhi22 concept with the exception of a voltage reference as current regulator. Should result in very precise current draw and hence no need to measure the actual current (voltage over resistors) but instead only measure the time taken before trip-off voltage is reached. Only measures cell voltage and 328p controlls trip-off via voltage level as well. If anything goes bad the only damage should be over discharged cell, which shouldn't present a fire hazard.

[bexander]

Final design.

Discharger

Screenshot_2023-06-29_13-33-11.png (7.68 KiB) Viewed 502 times

A copy of uhi22 design. I just rearranged the sens resistor so no need for differential ADC.
Simplest and cheapest solution I could find.
Voltage division on cell voltages is so I can use the internal 1,1V reference in the 328p, for best accuracy when measuring the current.

I also relized, to really cycle the cells as they will be used in the car a constant power load (increasing current with decreasing voltage) is needed. I think a good solution would be a boost converter steping up the voltage to say 5V and then a resistor load. This is complex and costly, but possible.
When I relized this I ditched the constant current discharger. Also more complex than a resistor.

[uhi22]

At the A0 just add a 10nF or 100nF, otherwise the charging of the internal sample&hold capacitor will lead to a voltage drop at the high-impedance divider. This has been discussed here: https://jeelabs.org/2013/05/16/measurin ... index.html