Has anyone been able to get REC to fully charge a battery?

[toms]

No luck. REC emailed me an updated firmware version that behaves differently but still won't allow a full charge in a reasonable amount of time.

And now they won't return emails. I take this as a warning that they know there's a problem with certain applications.

What's different with the new firmware you ask? Well, instead of starting to taper the current when the first cell hits the Balance Voltage Start setting, it doesn't start tapering the current until the first cell hits the End Of Charge setting, which is FAR better. All cells must reach the End Of Charge setting for the coulomb and SOC counters to re-calibrate, which never happens.

So now, instead of my pack (highest cell) getting to within 50 to 70 mV of being fully charged, the highest cell reaches full charge. But by the time the lowest cell has gotten to within 15mV of a full charge, the current has tapered down to just 1 or 2 amps.

At just 1 or 2 amps of charging, it would take about 12 hours to move the lowest cell the remaining 15mV, and by that time, the sun will have gone down long ago.

The guy over at OGSS did his best to help, but I don't think he is privy to the proprietary algorithm that REC uses so his assistance is limited to what is in the user manual. I haven't even bought anything from OGSS and he was trying to solve my problem, so that's a big thumbs up for them.

I'm beginning to think that the REC BMS is really for charging applications that have an unlimited 24 hour energy supply, and not really designed to get off-grid batteries charged up quick enough to keep up with the sun. If I were to leave the system plugged into the grid charging, I'm confident it would eventually get to a full charge, but it will never happen using off-grid solar.

A bit disappointed in that regard, but the reality is that in the bigger picture, it's not a show stopper problem. The battery pack will still charge, REC is still going to keep it from going kaboom, and I can always reset the SOC counter manually if needed. I just need to open the user interface, type in a bit of code, and tell it what SOC it should be at.

Thanks for the feedback. Did you ever try changing the battery type to LiFePO4?

I hear what you are saying, it isn’t a common problem with REC though - it’s just you.

To be clear i don’t think you are doing anything wrong - i think something in the REC settings you are using isn’t the same as what everyone else uses. I think if you keep trying you will get the result you are looking for. I’m glad you are sharing this.

 

[Hedges]

Well, instead of starting to taper the current when the first cell hits the Balance Voltage Start setting, it doesn't start tapering the current until the first cell hits the End Of Charge setting, which is FAR better.

I think that's the best that can be done, controlling charging. It couldn't charge your other cells any more fully.
Other part of the picture is balancing. Maybe it has drifted out of balance over time, will gradually rebalance with future use. Or you could manually bleed off high cells.

It it was particularly smart, it could keep track of which cells were out of balance how much, and continue to bleed off a calculated amount of charge even at lower voltages.

 

[MurphyGuy]

Thanks for the feedback. Did you ever try changing the battery type to LiFePO4?

Yes, and no significant difference. It did seem more aggressive, but it might have just been my own interpretation.

I hear what you are saying, it isn’t a common problem with REC though - it’s just you.

I'm finding that hard to believe since I have tested TWO different REC units with two different firmware versions.

I would imagine when REC is working with a DC coupled charge controller with simple on-off controls, that it wouldn't be a problem, but with the Sunny Islands, it seems to be an issue with both grid charging and AC coupling to solar.
I have scanned every page of the Sunny Island manual to search for clues or maybe a setting somewhere and no luck.

Of an interesting note, the REC BMS manual I got with my unit is different than the one on their website today. There are more detailed instructions and I'm going through it as well.
If anyone is interested in seeing the differences, I have the older manual PDF and you can download the new one directly from REC.

To be clear i don’t think you are doing anything wrong - i think something in the REC settings you are using isn’t the same as what everyone else uses. I think if you keep trying you will get the result you are looking for. I’m glad you are sharing this.

I don't know. This thing wasn't that difficult to hook up, just basic wiring. It's not like I had to use a wheatstone bridge to make sure every wire and connection was the exact same. And the menu of settings within the BMS software is also pretty simple. The Sunny Islands are a whole different ballgame, but even so, most of those settings are either irrelevant to lithium BMS's or just left in the default settings.

I've given up on it this past few weeks because of other projects. Maybe one rainy day I'll go at it again.

 

[MurphyGuy]

I think that's the best that can be done, controlling charging. It couldn't charge your other cells any more fully.

Yes, actually it can. The limit I set for maximum cell voltage is 4.2, so if my End Of Charge is set for 4.1, and it has to allow the highest cell to hit 4.115 in order to allow the lowest cell to hit 4.1, then isn't that what it should do?

Other part of the picture is balancing. Maybe it has drifted out of balance over time, will gradually rebalance with future use. Or you could manually bleed off high cells.

When the highest cell is at 4.1 volts, the lowest cell is at 4.085, which is only a 15mV difference.

It it was particularly smart, it could keep track of which cells were out of balance how much, and continue to bleed off a calculated amount of charge even at lower voltages.

REC has two balancing modes: 1) >Balance Voltage Start = it will only balance the cells when charging and 2) >Balance Voltage End = it will balance the cell no matter if its charging or not.

With the new firmware, Balance Voltage Start is pretty meaningless. When my system is charging, its usually charging at 100+ amps, and will reach almost full capacity within two or three hours at that rate. With a BMS that has a 1 amp balancing current on a 480aH pack, that 1 amp while charging is pretty insignificant. IE: 1 amp x 3 hours = 3 amp-hours (on a 480 ah battery) LOL

The Balance Voltage End setting is another story since that will balance the cells regardless of charge current, and that means it has 24 hours a day at 1 amp = 24 amp-hours of balancing capacity per cell.

Down near the bottom of the SOC in the 3.4 to 3.5 area, my cell voltage spread can be as much as 50 to 60 mV, but by the time they've come up into the 3.9 to 4.0 volt range, that spread has shrunk to just 15mV.

I'll keep working on it. The new manual has some very new stuff to look at.

 

[Hedges]

When the highest cell is at 4.1 volts, the lowest cell is at 4.085, which is only a 15mV difference.

I don't know your chemistry, or where the knee is. If those voltages are well into/past the knee, then all should be fully charged. (give or take time at a voltage to reach full charge.)

For LiFePO4, if one hit 3.500V and the other was 3.485V, I'd think that was fully charged.

 

[toms]

Yes, and no significant difference. It did seem more aggressive, but it might have just been my own interpretation.


I'm finding that hard to believe since I have tested TWO different REC units with two different firmware versions.

I would imagine when REC is working with a DC coupled charge controller with simple on-off controls, that it wouldn't be a problem, but with the Sunny Islands, it seems to be an issue with both grid charging and AC coupling to solar.
I have scanned every page of the Sunny Island manual to search for clues or maybe a setting somewhere and no luck.

Of an interesting note, the REC BMS manual I got with my unit is different than the one on their website today. There are more detailed instructions and I'm going through it as well.
If anyone is interested in seeing the differences, I have the older manual PDF and you can download the new one directly from REC.


I don't know. This thing wasn't that difficult to hook up, just basic wiring. It's not like I had to use a wheatstone bridge to make sure every wire and connection was the exact same. And the menu of settings within the BMS software is also pretty simple. The Sunny Islands are a whole different ballgame, but even so, most of those settings are either irrelevant to lithium BMS's or just left in the default settings.

I've given up on it this past few weeks because of other projects. Maybe one rainy day I'll go at it again.

Mine is an SMA AC coupled with a DC charger, but that only bulk charges and is finished by the time the SI starts frequency shifting.

I have setup plenty of purely AC coupled SMA systems, they all work the same as mine.

I have attached a pic of my system at 3pm today. It has been balancing for a short time, but as you can see is at 98% - and will reach 100% today.

edit: added a pic where SOC has gone to 100%

Something isn’t right with your settings. What size is your battery, and what value are you using for your shunt coefficient?

 

[MurphyGuy]

Mine is an SMA AC coupled with a DC charger, but that only bulk charges and is finished by the time the SI starts frequency shifting.

I have setup plenty of purely AC coupled SMA systems, they all work the same as mine.

I have attached a pic of my system at 3pm today. It has been balancing for a short time, but as you can see is at 98% - and will reach 100% today.

edit: added a pic where SOC has gone to 100%

Something isn’t right with your settings. What size is your battery, and what value are you using for your shunt coefficient?

My battery is LiMnO2 (from chevy bolt ev) set up in a 14s8p config for a total capacity of 480 Ah. I did checked the shunt coefficient and it is set to 0.0078125 which corresponds with my REC invoice for a 200 amp shunt.

I can also (mathematically?) verify the shunt being at the correct settings and accurate because when I use Grid Charge Mode with the Sunny Islands, the number of amps being pushed into the battery corresponds with the Sunny Island settings for what I have allowed it to draw in on the AC side.
(Sunny Island output not connected) For example, if I set the Sunny Islands for 15 amps, then 15 amps * 242v = 3630W and 3630W divided by the present battery voltage of lets say 53 volts = 68.49 amps going into the battery. Of course, the actual battery amps are never exactly spot on, but the math will always work out to within 1 or 2 amps of what it should be. You might have noticed the 242 volts, our grid voltage here is usually just a bit high with each leg at 121v and can go as high as 123 on some days.

At the end of the day, what my problem boils down to is the REC BMS tapering the current too far too soon. The manual says it will taper it down to 1.1 amps x number of devices. What is "Number of devices" ? Number of Sunny Island's? Number of Cells in the system? Number of REC BMS's connected?

 

[Baxter2120]

At the end of the day, what my problem boils down to is the REC BMS tapering the current too far too soon. The manual says it will taper it down to 1.1 amps x number of devices. What is "Number of devices" ? Number of Sunny Island's? Number of Cells in the system? Number of REC BMS's connected?

Are you sure. Just repeat myself over and over, you can not be sure if the BMS is tapering the current or your chargers without knowing the actual current limit your chargers are using. That information is still missing using SI.

Regarding number of devices, thats the parameter SISN within your BMS configuration. It is described in the manual.

 

[MurphyGuy]

Are you sure. Just repeat myself over and over, you can not be sure if the BMS is tapering the current or your chargers without knowing the actual current limit your chargers are using. That information is still missing using SI.

The battery charge current limit in the sunny islands is set to 100 amps and it hits that when we are charging with solar.
When using grid charge mode, the SI's are set to draw 10 amps (I reduced it from 15a), and that is what it draws. The 10 amp grid current draw ends up working out to 10a*242v = 2420W / (bat voltage) = what it should be.

Is there some other current limit setting in the Sunny Island setup that you are looking for?

Also, as far as if its the BMS doing the tapering, I can confirm that as soon as the highest cell reaches set point, the current begins to taper within one to three seconds. It will remain relatively steady within 1 or 2 amps up until the highest cell reaches the set point, and as soon as that happens, it will start rapidly tapering.
The Sunny Islands have no idea what cell voltages are, it only does what the BMS tells it to do.

Regarding number of devices, thats the parameter SISN within your BMS configuration. It is described in the manual.

I get that part, but what is a "device"? Is it the number of SI's, the number of cells, BMS's ??

 

[MurphyGuy]

Mine is an SMA AC coupled with a DC charger, but that only bulk charges and is finished by the time the SI starts frequency shifting.

I have setup plenty of purely AC coupled SMA systems, they all work the same as mine.

I have attached a pic of my system at 3pm today. It has been balancing for a short time, but as you can see is at 98% - and will reach 100% today.

edit: added a pic where SOC has gone to 100%

Something isn’t right with your settings. What size is your battery, and what value are you using for your shunt coefficient?

Quick thought. You're using Sunny Islands and REC BMS. Your REC obviously has a shunt, but do your Sunny Islands have their own shunt as well?

Mine do not. The only shunt in my system comes from REC for the BMS.

 

[Baxter2120]

The Sunny Islands have no idea what cell voltages are, it only does what the BMS tells it to do.


I get that part, but what is a "device"? Is it the number of SI's, the number of cells, BMS's ??

The SI does what the internal CCCV algorithm of SI is telling the SI to do. The BMS is just telling the upper limit to the SI AND you still do not know what the limit is because SI is not telling you what it receives as limit from the BMS (that what you have to look for).

Regarding device, it is in the manual: "Number of inverter devices on the bus" -> So number of SIs in your case

 

[Hedges]

Mine is an SMA AC coupled with a DC charger, but that only bulk charges and is finished by the time the SI starts frequency shifting.

I have setup plenty of purely AC coupled SMA systems, they all work the same as mine.

I have attached a pic of my system at 3pm today. It has been balancing for a short time, but as you can see is at 98% - and will reach 100% today.

That looks like 8.0H SI. Does that use the same REC BMS as our 6048US SI? SMA has a different list of compatible batteries for it.

Anybody know if REC is compatible with Sunny Boy Storage? It does support a high enough battery voltage, but protocol could be different.

"I've tried several setting changes in the REC and nothing is making a difference. My SOC meter thinks my battery is at 60% charge while at 4.02 volts because I have never been able to get the battery to a full 4.1 for the SOC counter to reset. Not once in 3 years."

If not full in 3 years, SoC of cells have had more time to diverge. If you have now corrected the situation, could take quite a while for SoC to rebalance.
Is there a problem with actual state of charge of battery & individual cells, or just reporting? What if you told it to reset at 3.95V?
The difference between 4.02V and 4.10V seems awfully small to trigger balancing and distinguish SoC. For LiFePO4, there is a knee (around 3.3V?), people top-balance to 3.65V, operate to maybe 3.4V or 3.5V. Enough above the knee to enable balancing and reset SoC counter.

 

[toms]

Quick thought. You're using Sunny Islands and REC BMS. Your REC obviously has a shunt, but do your Sunny Islands have their own shunt as well?

Mine do not. The only shunt in my system comes from REC for the BMS.

I only use the REC shunt.

Another possibility is that your cells diverge too much when they reach their balance point - and the REC reduces current to prevent overheating.

My cells are quite old, and when i haven’t seen 100% SOC for a while they will take a few days before the balancing algorithm returns to 100%.

As you can see from my screenshots, they are still a bit imbalanced but still get to high 90’s before SI frequency shift.

As Hedges points out, my experience is only with SI8.0H. As far as i’m aware the REC is the same, i use the REC SI BMS. I keep forgetting that the US Sunny Island is different, i have no experience there i’m sorry.

 

[MurphyGuy]

I only use the REC shunt.

Another possibility is that your cells diverge too much when they reach their balance point - and the REC reduces current to prevent overheating.

My cells are about 15mv spread when close to the 4.1v End Of Charge setpoint. Over heating? You mean the BMS or the cells? My battery bank barely goes up 1c to 2c in temperature, but the REC does hit 40c when balancing.

My cells are quite old, and when i haven’t seen 100% SOC for a while they will take a few days before the balancing algorithm returns to 100%.

As you can see from my screenshots, they are still a bit imbalanced but still get to high 90’s before SI frequency shift.

As Hedges points out, my experience is only with SI8.0H. As far as i’m aware the REC is the same, i use the REC SI BMS. I keep forgetting that the US Sunny Island is different, i have no experience there i’m sorry.

I'm doing another test right now. I made one minor change to the physical wiring that wasn't specified in the original REC manual. After reading the new manual, I changed the #16 wire from being jumpered (2 inches) to the #14 cell (pack positive) on the green BMS connector, to being routed directly to the pack positive terminal.
My thought is that if the BMS is balancing that #14 cell, whatever voltage drop is in the 18ga wire over the 3ft length, could somehow affect the BMS. That's a leap, but they didn't specifically put that in the new user manual for no reason. They even show diagrams for "this is right" and "this is wrong", and I was set up under the "this is wrong" diagram. Like I said, it wasn't addressed in the older manual.

I suspect it won't make any difference, but then again, we are dealing with voltage measurements in the millivolt range so it could be possible. Perhaps, when the #14 cell is being balanced and drawing close to 1 amp, the voltage drop across the wire causes some reference voltage in the BMS to be inadvertently offset causing it to see a false pack condition.

Here's a screenshot of the new manual page I'm referencing. I was set up under NO, but have now changed to the YES configuration were the #16 wire goes directly to pack positive.

 

[MurphyGuy]

Well, fixing the wire to their approved format made no difference at all.

Same behavior. Gets to within 10 or 15 mV of a full charge and stops.

It would seem that the BMS has sent a "STOP" command to the sunny islands because that -1.00 amps is the result of the current the sunny islands pull when they're sitting and doing nothing.

The system has been in grid-charge mode since around 11am this morning. The highest cell reached. 4.1 volts at around 1:15 to 1:30 pm, the lowest cell was about 15mV behind it. Once the BMS started tapering current at 4.1v (End Of Charge) setting, the charging hysteresis of the cells fell about 8 to 10 mV but the BMS did not bother to compensate. It just keeps ramping down the charge to zero like its on a timer and doesn't care about cell voltages.

 

[MurphyGuy]

This is interesting. Not sure how this affects anything.

In the REC Manual, it lists an ADC reference voltage under command REFC? (see photo)

The spec is supposed to be 5.000v +/- 0.003 (Read Only) but mine is 4.992

I'm wondering if the fact that it is 8mV off the mark has anything to do with the fact that my highest cell settles at about 8mV below the End of Charge setting. Maybe just a coincidence .

 

[MurphyGuy]

I only use the REC shunt.

Another possibility is that your cells diverge too much when they reach their balance point - and the REC reduces current to prevent overheating.

My cells are quite old, and when i haven’t seen 100% SOC for a while they will take a few days before the balancing algorithm returns to 100%.

As you can see from my screenshots, they are still a bit imbalanced but still get to high 90’s before SI frequency shift.

As Hedges points out, my experience is only with SI8.0H. As far as i’m aware the REC is the same, i use the REC SI BMS. I keep forgetting that the US Sunny Island is different, i have no experience there i’m sorry.

If you type in REFC? to your communicate box, what does your reference voltage return as?

 

[toms]

If you type in REFC? to your communicate box, what does your reference voltage return as?

I’ll let you know tomorrow. (i don’t have a PC permanently connected)

 

[bdserv]

I did set my END OF CHARGE to 4.16 once, then waited until all cells were over 4.1, then reset the END OF CHARGE to 4.1 while it was charging. It immediately stopped charging and reset the counters to 100%.

Failing discovery of a setup or firmware solution I would automate this sequence using something like an attiny85 or esp8266. Would be simple enough to monitor the REC for cell voltages and charge cycle status via the serial port or wifi, and make the setting adjustments when appropriate.

 

[Goboatingnow]

No luck. REC emailed me an updated firmware version that behaves differently but still won't allow a full charge in a reasonable amount of time.

And now they won't return emails. I take this as a warning that they know there's a problem with certain applications.

What's different with the new firmware you ask? Well, instead of starting to taper the current when the first cell hits the Balance Voltage Start setting, it doesn't start tapering the current until the first cell hits the End Of Charge setting, which is FAR better. All cells must reach the End Of Charge setting for the coulomb and SOC counters to re-calibrate, which never happens.

So now, instead of my pack (highest cell) getting to within 50 to 70 mV of being fully charged, the highest cell reaches full charge. But by the time the lowest cell has gotten to within 15mV of a full charge, the current has tapered down to just 1 or 2 amps.

At just 1 or 2 amps of charging, it would take about 12 hours to move the lowest cell the remaining 15mV, and by that time, the sun will have gone down long ago.

The guy over at OGSS did his best to help, but I don't think he is privy to the proprietary algorithm that REC uses so his assistance is limited to what is in the user manual. I haven't even bought anything from OGSS and he was trying to solve my problem, so that's a big thumbs up for them.

I'm beginning to think that the REC BMS is really for charging applications that have an unlimited 24 hour energy supply, and not really designed to get off-grid batteries charged up quick enough to keep up with the sun. If I were to leave the system plugged into the grid charging, I'm confident it would eventually get to a full charge, but it will never happen using off-grid solar.

A bit disappointed in that regard, but the reality is that in the bigger picture, it's not a show stopper problem. The battery pack will still charge, REC is still going to keep it from going kaboom, and I can always reset the SOC counter manually if needed. I just need to open the user interface, type in a bit of code, and tell it what SOC it should be at.

Interesting I’m building my own bms this winter so this is all interesting info