Overkill voltage spike on high cell cut off

[jbatx]

Issue: when one of my bms cuts charge due to high cell voltage, it causes a voltage spike on the dc battery side of things. I have measured this all the way up to 70v. When this happens, my inverter will fault and cut all power until I go out there and reset it.

System: I have three overkill 100a bms's on three 48v packs providing 21.1kwh via a sungold hybrid 6kw split phase inverter (same as lvx6048). Prior to that inverter I had a renogy scc and a dumb 5kw inverter. The issue occurred in both set ups

I know why the cut off is happening, I'm getting cell drift earlier during charging over the last few weeks.

I've lowered bulk voltage to 55.9 and float to 55.8. I'll be going lower since I it cut off yesterday. Previously, 56.8 was possible with drift only getting into the 50mv range. Now, at 55.9 I'm getting 160mv or greater. All packs have capacitive balancers with the overkill balancing switched off.

Question is, is there device I can place on rhe dc battery side which could absorb that spike when and if it happens?

 

[squowse]

Have you tried without the capacitive balancers? Is it possible they could be causing the spike?

 

[jbatx]

Yep, it does it without them as well.

 

[GXMnow]

Your real problem here is that your cells are going out of balance. What cells are you using? Can you do a proper full charge top balance on them?

From the location of your post, I am assuming you have a 16S LiFePO4 arrangement. 55.8 / 16 = 3.4875 volts per cell. That seems reasonable, but is still going into the knee. When you check the cell voltages, is it always the same cell going too high? If it is, can you put a load on that cell and drag it down a bit? What capacitive balancer are you using?

How low are you running the voltage while inverting? If your balancer is always trying to balance, it might be causing a problem if you have a weak cell and you are hitting the lower knee. If you know which cell is going to high, check the cell voltages when the state of charge is getting low. Is that same cell still high, or is it now low? If the high cell after charging, is also the low cell after discharging, then you have a weak cell that may need to be replaced. The whole pack can only be cycled as far as the weakest cell will allow. Ideally, you should never be running down into the lower knee of the weakest cell. Then you top balance, and only have the balance work up near full charge to keep the cells top balanced.

If you are cycling less than 50% capacity, I would suggest dropping your absorb voltage to just 54.4 volts. That is still 3.4 volts per cell, which if you let it hold there long enough, will still hit over 95% charge. And set the float voltage at 54.38 and you will run off the sun until the solar just can't keep up with the load, while the cells continue to stay full and balance at the top.

As for the voltage spike... The charger is putting out a lot of current, and then 1/3 of the load goes away when your BMS cuts off. That is going to cause the voltage to jump a bit, and could very well cause a second BMS to also turn off, making it even worse. The real fix is making sure the BMS is not turning off charge current. One thing that could help would be to ad a large amount of capacitance to the battery buss between the inverter/charger and the battery banks. Because of you seeing 70 volt spikes, use 100 volt rated capacitors. I would think 10,000 uF would slow the voltage rise enough for the charging circuit to catch it and go into CV regulation mode. You might also be able to add an MOV type surge protector, but to keep it from being a load at 54.4 volts, you would want something like a 75 volt MOV, so it won't really clamp a spike unless it is over 70 volts. Using a lower value MOV will start pulling current while your system is at normal operating voltage. The other idea for clamping a spike like that would be an active shunt regulator, but that can get expensive. A high power zener diode circuit could be setup to clamp the voltage, but it will again have a small parasitic draw all the time. Here is a page talking about a circuit like this.

Transistor Zener Diode Circuit for Handling High Current Stabilization - Homemade Circuit Projects

The high power "zener diode" circuit using transistor shunt regulator presented here can be used for obtaining a highly accurate, temperature and voltage stabilized outputs from high current sources, safely. […]

www.homemade-circuits.com

It uses an adjustable shunt regulator IC to drive a power transistor to clamp the voltage to your desired maximum. But you may need to parallel several transistors to have it able to clamp the output of a high current charger. If it is charging at say 50 amps, and the battery BMS goes open, it might have to clamp all 50 amps to say 65 volts. That would mean dissipating 50 x 65 = 3,250 watts of power. Even if it is for a fraction of a second before the charge corrects, we are talking a huge current inrush. You are far better off stopping the BMS from shutting off in the first place.

 

[BradCagle]

Voltage regulation has an overshoot when the load drops, going from high current to low (such as when your BMS drops)
It takes some time for the charger to react to the change in load.

 

[mikefitz]

It may be the capacitive balancers are having the opposite effect to what you are expecting, in reality pulling the cells out of balance.

Mike

 

[Dynoman]

Issue: when one of my bms cuts charge due to high cell voltage, it causes a voltage spike on the dc battery side of things. I have measured this all the way up to 70v. When this happens, my inverter will fault and cut all power until I go out there and reset it.

System: I have three overkill 100a bms's on three 48v packs providing 21.1kwh via a sungold hybrid 6kw split phase inverter (same as lvx6048). Prior to that inverter I had a renogy scc and a dumb 5kw inverter. The issue occurred in both set ups

I know why the cut off is happening, I'm getting cell drift earlier during charging over the last few weeks.

I've lowered bulk voltage to 55.9 and float to 55.8. I'll be going lower since I it cut off yesterday. Previously, 56.8 was possible with drift only getting into the 50mv range. Now, at 55.9 I'm getting 160mv or greater. All packs have capacitive balancers with the overkill balancing switched off.

Question is, is there device I can place on rhe dc battery side which could absorb that spike when and if it happens?

As @BradCagle and @GXMnow stated previously the voltage spike is probably from the Solar Charge Controller as it takes some time for it to adjust.

I agree with @mikefitz that the balancers are throwing off the top balance of the battery bank at lower state of charge.

I posted this in another thread, but it may be useful to someone here too.

The Heltec 5a active balancer is used in my system with two grade B 271ah 8s 24 VDC Lifepo4 battery banks and it works well.

For 4s 12VDC Lifepo4:
https://www.amazon.com/Heltec-Equalizer-Balancer-Equalization-Capacitor/dp/B08CKPZ7PH

For 8s 24VDC:
https://www.amazon.com/Heltec-Equalizer-Balancer-Transfer-Capacitor/dp/B08D9D4NGW

For 16s 48VDC:

https://www.amazon.com/Heltec-Capacitor-Equalizer-Balancer-Titanate/dp/B08M62MVHD/ref=sr_1_2

There are (run) solder pads on the Heltec active balancer that can be used to turn the balancer on and off with a switch.

My system uses the relay on the Victron Energy BMV 712 battery shunt display wired to the run solder pads to switch on and off the Heltec active balancer.

At voltages above 27.3VDC the Heltec active balancer is switched on and balancing begins. As the voltage continues to increase above 27.3VDC the Solar Charge Controllers in the system are programmed to go into Boost (absorb) charge at 28 to 28.2VDC for 10 minutes. During boost charge mode the Heltec active balancer keeps all cell voltages at less than 3.55VDC as the cells balance. At the end of 10 minutes of boost charge mode all cells are within 30 mv or less. After 10 minutes of boost the SCC's are programmed to go into float mode at 27.4VDC. The cells continue to balance and after a few minutes in float the cells are within 15 mv or less. The SCC's stay in float mode while suppling loads during the day. Toward the end of the day (or if clouds build) when SCC's cannot hold float and the battery voltage goes below 27.2 the active balancer is turned off by the Victron shunt relay.

Originally the Heltec active balancers were just wired to the cells and it ran all the time, but I found that cells would balance mid charge and when the voltage approached the boost voltage at high current charge, some of the cells would trigger the BMS cell voltage limit (3.65 VDC) and shut down the battery bank. During mid charge of the battery banks some of the cells can vary 30mv or more and the Heltec balancer would balance the cells at mid voltage throwing off my top balance. I guess that's due to grade B poorly matched cells. Once the Heltec active balance was triggered to only balance at 27.3 VDC and above everything worked fine.

Due to the load demands on my off grid solar system it may be weeks before the system is able to charge up to Boost (absorb) charge level of 28 to 28.2VDC depending on sunny vs cloudy days in my area. Even with that amount of time in-between boost (full charge) the cells still balance well with the Heltec 5a active balancer. Without the Heltec active balancer one or more of the battery bank cell(s) was prone to over voltage triggering a BMS disconnect on that battery bank.

With my two Overkill Solar 8s BMS set at 3.65 Cell Voltage Limit this works well for both battery banks that are in parallel configuration.

It is my understanding that a Digital low voltage battery cut-off relay module is able to turn on and off the Heltec active balancer like the Victron 712 BMV shunt relay, but I have not tried it since my system has the Victron BMV 712 shunt. If some tries this battery cut-off relay module please post your results here. It looks like there are 12 & 24 Volt units. Don't know about 48 Volt...

https://www.amazon.com/NOYITO-Voltage-Detection-Protection-Discharge/dp/B07DS33JVJ

 

[jbatx]

I think you all are correct about getting to a too-low state of charge and the balancers as the root cause. The batteries were down to near zero soc about 10 or so days ago and that seems to correlate with this new issue. The cells did begin to drift when it down that low. Since then I've set the "back to grid" voltage to 51 where as before it was 46. That should keep it out of the lower curve before inverted switches to grid power. I'll be out there tomorrow and I'll get them charged to 100% gently and let the heltec 5a balancers do their job. After that'll I'll keep the bulk and float voltages even lower since I now have a grid connection and I don't need to rely on the batteries quite so much.

Thanks all!

 

[jbatx]

As @BradCagle and @GXMnow stated previously the voltage spike is probably from the Solar Charge Controller as it takes some time for it to adjust.

I agree with @mikefitz that the balancers are throwing off the top balance of the battery bank at lower state of charge.

I posted this in another thread, but it may be useful to someone here too.

The Heltec 5a active balancer is used in my system with two grade B 271ah 8s 24 VDC Lifepo4 battery banks and it works well.

For 4s 12VDC Lifepo4:
https://www.amazon.com/Heltec-Equalizer-Balancer-Equalization-Capacitor/dp/B08CKPZ7PH

For 8s 24VDC:
https://www.amazon.com/Heltec-Equalizer-Balancer-Transfer-Capacitor/dp/B08D9D4NGW

For 16s 48VDC:

https://www.amazon.com/Heltec-Capacitor-Equalizer-Balancer-Titanate/dp/B08M62MVHD/ref=sr_1_2

There are (run) solder pads on the Heltec active balancer that can be used to turn the balancer on and off with a switch.

My system uses the relay on the Victron Energy BMV 712 battery shunt display wired to the run solder pads to switch on and off the Heltec active balancer.

At voltages above 27.3VDC the Heltec active balancer is switched on and balancing begins. As the voltage continues to increase above 27.3VDC the Solar Charge Controllers in the system are programmed to go into Boost (absorb) charge at 28 to 28.2VDC for 10 minutes. During boost charge mode the Heltec active balancer keeps all cell voltages at less than 3.55VDC as the cells balance. At the end of 10 minutes of boost charge mode all cells are within 30 mv or less. After 10 minutes of boost the SCC's are programmed to go into float mode at 27.4VDC. The cells continue to balance and after a few minutes in float the cells are within 15 mv or less. The SCC's stay in float mode while suppling loads during the day. Toward the end of the day (or if clouds build) when SCC's cannot hold float and the battery voltage goes below 27.2 the active balancer is turned off by the Victron shunt relay.

Originally the Heltec active balancers were just wired to the cells and it ran all the time, but I found that cells would balance mid charge and when the voltage approached the boost voltage at high current charge, some of the cells would trigger the BMS cell voltage limit (3.65 VDC) and shut down the battery bank. During mid charge of the battery banks some of the cells can vary 30mv or more and the Heltec balancer would balance the cells at mid voltage throwing off my top balance. I guess that's due to grade B poorly matched cells. Once the Heltec active balance was triggered to only balance at 27.3 VDC and above everything worked fine.

Due to the load demands on my off grid solar system it may be weeks before the system is able to charge up to Boost (absorb) charge level of 28 to 28.2VDC depending on sunny vs cloudy days in my area. Even with that amount of time in-between boost (full charge) the cells still balance well with the Heltec 5a active balancer. Without the Heltec active balancer one or more of the battery bank cell(s) was prone to over voltage triggering a BMS disconnect on that battery bank.

With my two Overkill Solar 8s BMS set at 3.65 Cell Voltage Limit this works well for both battery banks that are in parallel configuration.

It is my understanding that a Digital low voltage battery cut-off relay module is able to turn on and off the Heltec active balancer like the Victron 712 BMV shunt relay, but I have not tried it since my system has the Victron BMV 712 shunt. If some tries this battery cut-off relay module please post your results here. It looks like there are 12 & 24 Volt units. Don't know about 48 Volt...

https://www.amazon.com/NOYITO-Voltage-Detection-Protection-Discharge/dp/B07DS33JVJ

regarding turning on/off the balancers; I think what I'd do is add a relay to my esp32 based monitors that are on the bms's now. it'd be a simple add-on and simple code mod.

 

[GSXR1000]

put this between SSC and battery

 

[740GLE]

What SCC?

As others have stated BMS should never stop the SCC from charging the cells. The SCC should ramp down current as system voltage is raised.

Good luck with balancing your cells.

 

[jbatx]

What SCC?

As others have stated BMS should never stop the SCC from charging the cells. The SCC should ramp down current as system voltage is raised.

Good luck with balancing your cells.

Sungold hybrid 240 split phase 6wk. It's the exact same as the lvx6048 that mpp and the others sell. Maybe a more expensive scc would ramp down, but this one doesn't. It just faults with code 8. According to sungold support, it's caused by voltage spike on battery.

The BMS "shouldn't" do that you're right - until something changes and it does - which is what it's designed to do.

Then you have to adapt, somehow. As I had mentioned in the first post, this is a new issue.

It's likely due to the near 0 soc state and the active balancers screwing things up. I'm sure I'll get around to adding automation to turn them on/off as needed.

 

[740GLE]

Agreed the SCC is not the root cause, but it identified your issues, out of balance bank caused by balancer.

Are you monitoring cell voltages near high voltages, you might have caught this unbalance starting to happen before it got to the point of high voltage disconnect.

 

[Dynoman]

regarding turning on/off the balancers; I think what I'd do is add a relay to my esp32 based monitors that are on the bms's now. it'd be a simple add-on and simple code mod.

Yes, I find turning on the Heltec 5A active balancer around 27.3 to 27.4 volts just before the voltage ramping up into the knee of the Lifepo4 battery bank works well. Then, running the voltage up into the knee at approximately 28 volts or so for a short period of time (10 to 20 minutes) really appears to help the top balance of the cells.

The active balancers are then turned off just below the float charge value at 27.2 volts. That way the Heltec balancers are turned on all the time the SCC is in float around 27.4 volts.

The balancer of the Overkill solar BMS(s) is also programmed to turn on at 27.4 volts or higher, but it's unlikely the little blead off resistors does much on 271 AH cells.

Programming the esp32 to switch the Heltec 5A active balancers on and off sounds like a good idea.

 

[jbatx]

[SOLVED]

not battery related - I assumed it was a battery v spike issue for historical reasons.

it was n-g bond related

details in my inverter review thread: https://diysolarforum.com/threads/sungold-split-phase-6k-hybrid-inverter-scc-review.36334/post-533456

 

[740GLE]

So have you been able to charge the battery and monitored cell divination?

It’s been proven time and again having an active balancer connected 24/7 unbalanced the cells.

 

[jbatx]

So have you been able to charge the battery and monitored cell divination?

It’s been proven time and again having an active balancer connected 24/7 unbalanced the cells.

I'm not saying that it's not an issue, in general. it is. But, that isn't causing my fault code 8 problems this time. ...though, it has in the past. When I checked my data records for all the bms units (stored in Blynk) none of them were hitting hvc for any of fault and shutoff situations. I was able to charge to 57v yesterday and the balancers did their job.