Victron MPPT 150|45 overcharging issue - 2 x 24v Power Queen 100ah LifePo4 battery

[Miketcq8]

My first post here, so go easy folks.

I have an off grid summer cabin in Finland using a solar array to recharge. The system is new so no much history yet on the system.

Equipment commisioned
6 x 3S2P 195W Ecoworthy Bifacial solar panels
3 x Power queen 25,6v 100Ah batteries
Victron smart shunt 500A
Victron MPPT 150 / 45
Victron global link
Victron Smart Battery Sense used to cut off at 5deg C
Olteanp 3500W 34v Pure sine wave

The 220 V appliances is a coffee machine. About to add a fridge, but need some input on a random over battery voltage issue that rears its head randomly. The batteries are charged at around 28,5 / 28,7V when this happens. Sounds like the batteries are at +-100%

I search the forums and there's a load of similar posts but they don't seem to be the same issues. See the screen shots attached. I randomly get over voltage disconnect on the system. I thought is was the inverter, so I disconnected it. The issue persisted on the 24V DC system where a 24v DC fridge is running full time.
Power queen recommend to set the over voltage disconnect at 15v / 30v. How do I set this on a Victron MPPT? There are no settings I can find in the Victron form.

 

[pollenface]

28.5-28.7 is pretty high, try 27.6-28.0

Over voltage disconnect only applies if you are using loads controlled by a TX digital output cable and 5v solid state relay. The settings can be made under "load output" and "user def algorithm 1"

 

[Q-Dog]

I agree ... 28.5v+ is pretty high. Maybe you are getting a voltage spike when the battery BMSs cut off charging. Try setting bulk charging to 28v.

 

[Miketcq8]

28.5-28.7 is pretty high, try 27.6-28.0

Over voltage disconnect only applies if you are using loads controlled by a TX digital output cable and 5v solid state relay. The settings can be made under "load output" and "user def algorithm 1"

Power Queen recommends
Charge: 28.8V
Float: 27,2V
Max voltage: 29,2V
Equalise: 28.8V

My issue is the spikes ( seen in the images attached) show @ 33.5v and 33.0v.
*** I have added this to the post above

Any idea why this is happening?

 

[Miketcq8]

28.5-28.7 is pretty high, try 27.6-28.0

Over voltage disconnect only applies if you are using loads controlled by a TX digital output cable and 5v solid state relay. The settings can be made under "load output" and "user def algorithm 1"

Firstly thanks for clarifying the over voltage disconnect....

My bad, I omitted some information. I have added it to the post.
Power Queen recommends
Charge: 28.8V
Float: 27,2V
Max voltage: 29,2V
Equalise: 28.8V

My issue is the spikes ( seen in the images attached) show @ 33.5v and 33.0v.
*** I have added this to the post above ***

Any idea why this is happening?
How do I set a max voltage in a Victron system.

 

[Cmiller]

This sounds very much like the BMS shutting down due to cell voltage spike. If the cells aren't completely balanced, then when the pack voltage gets higher, you very easily end up with a cell spiking, which causes the BMS to shut down. If the BMS shuts down while the battery is being charged you will quickly get high voltage spikes like you are seeing!

As others mentioned, try setting the charge voltage lower. Usually when I encounter a battery with this issue, I will set the charge voltage to ~27.8V. At 27.8V the battery will still get fully charged, but it will do somewhat of a slight "taper charge", giving the BMS some time to try to balance the cells.

I would recommend as follows:
Absorb Charge: 27.8V
Float: 27.2V
Max voltage: 29V
Equalize: 28.2V

 

[Miketcq8]

This sounds very much like the BMS shutting down due to cell voltage spike. If the cells aren't completely balanced, then when the pack voltage gets higher, you very easily end up with a cell spiking, which causes the BMS to shut down. If the BMS shuts down while the battery is being charged you will quickly get high voltage spikes like you are seeing!

As others mentioned, try setting the charge voltage lower. Usually when I encounter a battery with this issue, I will set the charge voltage to ~27.8V. At 27.8V the battery will still get fully charged, but it will do somewhat of a slight "taper charge", giving the BMS some time to try to balance the cells.

I would recommend as follows:
Absorb Charge: 27.8V
Float: 27.2V
Max voltage: 29V
Equalize: 28.2V

Thanks. I’ll do the adjustments tomorrow when I go to the cottage. I’ll report back in a few days either way.

I have also asked the manufacturer for direction. Waiting a reply.

 

[RussNM]

In your victronconnect app under your scc (Mokki) 150/45 go to settings/battery and do a screenshot and post it please.

 

[Whats_per_question]

Spoiler: What i said (which was incorrect)

If the BMS shuts down and causes a spike, this would be in the range of a few 10s of millivolts over the charging voltage. That is not what is happening here. In this case, the user says they see 33v, and 33.5v.

What is happening, which is supported by the evidence that it happens at 100% SOC, is that the BMS is throwing back a high voltage to get the charging device to back off.

This happens to our BMS (PACE and AFUPS, and probably many others) and

Edit: All this was totally wrong, thanks to @Cmiller for putting me straight on this.

Its solved by reducing the charging voltage by 0.1v at a time until the issue goes away.

Once the issue goes away, you might be able to bump that charging voltage back up again over time, especially if the cells were unbalanced and then the battery has enough time at high SOC with enough solar coming in to do the balancing. Some BMS will only balance with 100mA or there abouts, so a badly out-of-balance pack will take a long long time to bring into balance. While it is unbalanced, the pack won't come up to full voltage without causing the highest cell to hit the alarm point. For example, if 5 cells are at 3.33v while 3 are at 3.53v, you've got 27.24v but the battery is "full" (because the 3 high cells can't go much higher). If all cells were balanced, you would be "full" at 3.53x8 = 28.24v.

To reduce the charge voltage, in victronconnect, settings, battery, "Absorption voltage"

 

[Cmiller]

If the BMS shuts down and causes a spike, this would be in the range of a few 10s of millivolts over the charging voltage. That is not what is happening here. In this case, the user says they see 33v, and 33.5v.

No, this is not accurate! I have seen 80V+ spikes on 51.2V bank when the BMS shut down! Which is same as 40V+ on 25.6V bank!

What is happening, which is supported by the evidence that it happens at 100% SOC, is that the BMS is throwing back a high voltage to get the charging device to back off.

The BMS WILL NOT "throw back" high voltage to the charging device to get it to shut down! That is simply not how BMSs work! When there is an overvoltage the shut down, and the charging device usually just simply doesn't respond quickly enough to keep the voltage from spiking!

"Throwing back" high voltage would potentially cause serious issues such as "frying" said charging device! We have had inverters and charge controllers get toasted from this overvoltage spike!

This happens to our BMS (PACE and AFUPS, and probably many others) and its solved by reducing the charging voltage by 0.1v at a time until the issue goes away.

Cell imbalance will cause this effect in any battery with a BMS.

Once the issue goes away, you might be able to bump that charging voltage back up again over time, especially if the cells were unbalanced and then the battery has enough time at high SOC with enough solar coming in to do the balancing. Some BMS will only balance with 100mA or there abouts, so a badly out-of-balance pack will take a long long time to bring into balance. While it is unbalanced, the pack won't come up to full voltage without causing the highest cell to hit the alarm point. For example, if 5 cells are at 3.33v while 3 are at 3.53v, you've got 27.24v but the battery is "full" (because the 3 high cells can't go much higher). If all cells were balanced, you would be "full" at 3.53x8 = 28.24v.

This is correct, all except the battery isn't "full". It is just seeing cell overvoltage, and therefore shutting down!

Again, the BMS can not and will not "throw back" high voltage! That is simply not how BMSs work! They shut down to protect the battery when there is cell overvoltage, overcurrent, etc. When current is flowing and the BMS opens, you get a voltage spike before the charging device has time to respond!

 

[Miketcq8]

In your victronconnect app under your scc (Mokki) 150/45 go to settings/battery and do a screenshot and post it please.

Ok this is the default setup as per the Power Queen manual. Today I’ll reduce the parameters as per the recommendations in the post above;
Charge: 27.8V
Float: 27.2V
Can’t set max Volts on my Victron gear
Equalize: 28.2V

 

[Whats_per_question]

"Throwing back" high voltage would potentially cause serious issues such as "frying" said charging device! We have had inverters and charge controllers get toasted from this overvoltage spike!

We can debate the cause , but the solution is clear; reduce the charging voltage.

 

[Whats_per_question]

Ok this is the default setup as per the Power Queen manual. Today I’ll reduce the parameters as per the recommendations in the post above;
Charge: 27.8V
Float: 27.2V
Can’t set max Volts on my Victron gear
Equalize: 28.2V

You set the "max volts" (ie charge voltage) by changing absorb voltage - 28.8v is 3.6v per cell, far too high in my opinion, 3.55v would be much better (i.e 28.4v) but 3.5v might be best for now to solve the BMS disconnecting issue.
Absorb voltage is the point the battery will charge to in bulk mode (constant current, with the voltage of the battery slowly increasing), then it changes to constant voltage (and decreasing current) at the absorb voltage.
Once the current drops below the tail current, the battery (for LiFePO4) is considered full. This is the default charging pattern ("3 stage") for Victron LiFePO4. Lead is slightly different.
A sustained voltage of 3.6v per cell during absorb would very likely cause issues, as some cells would be above this unless the battery was perfectly balanced

 

[Miketcq8]

If the BMS shuts down and causes a spike, this would be in the range of a few 10s of millivolts over the charging voltage. That is not what is happening here. In this case, the user says they see 33v, and 33.5v.

What is happening, which is supported by the evidence that it happens at 100% SOC, is that the BMS is throwing back a high voltage to get the charging device to back off.

This happens to our BMS (PACE and AFUPS, and probably many others) and its solved by reducing the charging voltage by 0.1v at a time until the issue goes away.

Once the issue goes away, you might be able to bump that charging voltage back up again over time, especially if the cells were unbalanced and then the battery has enough time at high SOC with enough solar coming in to do the balancing. Some BMS will only balance with 100mA or there abouts, so a badly out-of-balance pack will take a long long time to bring into balance. While it is unbalanced, the pack won't come up to full voltage without causing the highest cell to hit the alarm point. For example, if 5 cells are at 3.33v while 3 are at 3.53v, you've got 27.24v but the battery is "full" (because the 3 high cells can't go much higher). If all cells were balanced, you would be "full" at 3.53x8 = 28.24v.

To reduce the charge voltage, in victronconnect, settings, battery, "Absorption voltage"

Very helpful, much appreciated. I took the advice and dropped the charge parameters down to 27.8v and float to 27.2v

You set the "max volts" (ie charge voltage) by changing absorb voltage - 28.8v is 3.6v per cell, far too high in my opinion, 3.55v would be much better (i.e 28.4v) but 3.5v might be best for now to solve the BMS disconnecting issue.
Absorb voltage is the point the battery will charge to in bulk mode (constant current, with the voltage of the battery slowly increasing), then it changes to constant voltage (and decreasing current) at the absorb voltage.
Once the current drops below the tail current, the battery (for LiFePO4) is considered full. This is the default charging pattern ("3 stage") for Victron LiFePO4. Lead is slightly different.
A sustained voltage of 3.6v per cell during absorb would very likely cause issues, as some cells would be above this unless the battery was perfectly balanced

thanks for the insights. I followed Power Queen’s manual to setup the MPPT. But I’ll take the advice given here and move to 28v and move increase it by a decimal point over time to find the sweat spot. Now I have the 220v fridge drawing 50w and the systems stable. A few good sunny days and I’ll have clear direction.

 

[Whats_per_question]

Good stuff, great to see your spikes have stopped.
There is almost no capacity gained between 28 and 28.8, much like the level in a fuel tank when you get to the filling tube - i.e the fuel level rises fast in the filling tube, but doesn't contribute all that much to the total fuel capacity.

 

[RussNM]

Victron has a very good conservative preset charging program (rotary dial 7) for lifepo4 batteries. IIRC its 28.4 absorb and 27.0 float. ..when it reaches 28.4 it stays for 2hrs so all cells can catch up and balance. It might be a good idea to use the preset for a day to see how your battery reacts.

 

[Cmiller]

We can debate the cause , but the solution is clear; reduce the charging voltage.

I am not debating the cause. I am simply informing you of what is actually happening. I have been heavily involved with Rubix Battery and Ark Battery (Ark is now part of Rubix). I do not work for them, but for a sister company, and as such, I saw them start up and saw quite a bit of the early R&D etc. I was also heavily involved in designing the Rubix test lab's power system that powers their cycle testing machine, as well as their load bank for surge testing BMSs in their batteries and also 3rd party batteries. A BMS simply does not (and can not) "throw back" high voltage.

In a scenario like OP is seeing, the BMS goes into OVP (OverVoltageProtection) mode due to individual cell overvoltage, and shuts down the charge FETs (they have separate charge and discharge FETs), therefore depending on the charge current and the connected DC load (this could even be an inverter) you will get a variable voltage spike before the charging source has time to shut down. This variable voltage spike is also dependent on how quickly the charging device can respond to voltage rise. Sometimes there is hardly any voltage spike, other times it is quite extreme!

Regardless, you are correct that lowering the charge voltage "target" will avoid seeing cell voltage spikes as quickly, as well as give the BMS time to balance the cells. And also, as you mentioned, with time the charge voltage can often be gradually bumped up to a higher level once the cells are balanced. And also, I agree 100% the 28V vs 28.8V will result in such a minimal capacity gain that it is not really worth going higher than 28V if the battery has any potential for cell imbalances!

Hope my explanation makes sense to you @Whats_per_question As I said, I am not debating, I am simply informing! I have worked on a few LifePo4 batteries over the last few years... (a few hundred, that is... )

 

[Rocketman]

Do you have BlueTooth on the batteries to see the cell voltages?

If so as the battery gets full watch for cell imbalances - any cells being higher than others. ( you won’t see it until the batteries are almost full).

Since you lowered the absorption in the charger - over time the cell imbalance should go away- but keep watching it to verify.

 

[Miketcq8]

Do you have BlueTooth on the batteries to see the cell voltages?

If so as the battery gets full watch for cell imbalances - any cells being higher than others. ( you won’t see it until the batteries are almost full).

Since you lowered the absorption in the charger - over time the cell imbalance should go away- but keep watching it to verify.

I don’t believe that these models have Bluetooth.

 

[Miketcq8]

***Update. It’s all settled down now. Thanks for the advice from all the contributors, it was a good education for me.
Absorption: 28,0
Float: 27,2
Equalize: 28,0

 

[Q-Dog]

You need to turn equalization OFF.

 

[HarryN]

Glad to hear that it is working. Some people will add a small load- example 24 volt fan or light - that is on all of the time. It helps to deal with pulsing / spikes in power systems while being charged.