A common limitation with all charge controller and AIO systems

[toms]

Bogart Engineering is the only solar controller that charges according to the battery monitor.

My SMA system charges according to my BMS instructions. I’m sure there are plenty of others. I wouldn’t have it any other way.

 

[Ampster]

My SMA system charges according to my BMS instructions.

And I am sure when a cell or the pack reaches a voltage setting the BMS tells the charger to turn off. Unless you have set the voltage lower on the charger which many people do so the BMS is the final fail safe.

 

[coolbz]

From the thread you link to:
"Even 1 volt drop has been causing Growatt to switch power mode unnecessarily since most settings are based on voltage. "
"If I try full load of 100amps, the voltage drop will be 5 volt, that is tooooo much."

What is the Voltage setpoint you set the GroWatt to go to utility mode when the Voltage on battery is low?
If 1V is the setting window for your 48V system, that is way too narrow.
5V drops means you have problem in your system.

My settings are like this:
49.6v batt -> grid
54.5v grid -> batt

I have been using my system as primary power source of my household, cycle once per day, grid as a backup to fill battery at cheap night hours. My battery is sized at 5kwh to be enough for 1 full cycle per day. I could have "solve" most issues by doubling or tripping battery, but that's not real solution.

As I measured, my battery system has an internal resistance of 0.05 ohms, which is too much if I want to run my system near full watt capacity.

I think the real solutions are:
1, AIO has an integrated shunt to measure and use capacity.
2, AIO allow user to input battery ohms value and dynamically calculate and compensate volt drop.

Both solutions should be very doable from hardware/software standpoint by manufacture.

For me, I need to figure out how to reduce battery resistance. And I'm going to experiment a third solution:
3, Build a 48volt 200ams Headway pack and parallel to my existing battery. That's is going to reduce overall battery ohms. That's like CPU cache memory helping RAM speed in a computer design.

 

[coolbz]

I'd agree you might need to look at something deeper in your set up.

My 48v system was loaded to 2.2kw this evening and the voltage sagged 0.3 volts.

FYI, Schneider advertises using SOC to control battery charge/discharge. I haven't tried it, but that fits your request to use a shunt as best I can tell.

Thanks for advice of Schneider, I will look into it later. I'm curious what batteries people have been using and it sounds like your battery has much lower ohms than mine.

 

[coolbz]

Something?
I have been messing with Lithium batteries for ten years and I like to use physics as a basis for my understanding. I also look at best practices of OEM EV manufacturers, battery manufacturers spec sheets that state a maximum charge current and voltage. All the charge controllers and AIO inverters use two fundamental settings, Constant Current and Constant Voltage.
There is always room for innovation but I am more moved by objective facts if we are going to talk about things in the physical realm or the chemistry of ions and electrolytes.

It is my measurement: 1 volt / 20 amps -> 0.05 ohms

Conclusion so far is that my 16 battery modules have large ohms so it's causing the system not able to perform full watt capacity as originally designed. Lithium battery and AIO system are still far from being a mature product to replace 24 by 7 household power system.

 

[coolbz]

Unfortunately Coulomb counters have drift. I have one in my inverter and another in my BMS and sometimes they disagree. The only time I know they are correct is when they both read 100% and that is because they are programmed to reset when my battery finishes the Constant Voltage stage of charging. However I do pay attention to what my Coulombs counters tell me I have used since a full charge.
I am not trying to argue with you but I just want other readers to see another opinion. I also think charging is a separate issue from discharging. Our batteries are sensitive to over charging and the measure of that is voltage and current. Certainly there are also minimum voltages that the batteries should not exceed. However between the top and the bottom a shunt or Coulomb counter is a very useful tool to see how much energy you have used and how much might be left in you pack.
That is why I do not see the problem stated in the title of this thread as a problem that needs a solution.

Regarding counter drift, maybe a smarter design would be able to auto-calibrate once in a while, based on voltage under 0 load.
For most people, AIO + Lifepo4 battery is for Rv trip or backup purpose, so battery only need to be cycled once per month or year. Using voltage is an easy way to estimate battery status.
Although if this system is running as primary household power source on daily charge/discharge cycle and cycling between battery/solar/grid, it is an issue. Maybe I expected too much and squeezed specs too tight.

 

[coolbz]

BTW, I changed "problem" to be "limitation" in the title since many people do not agree this is a problem.

 

[John Frum]

I used to have a propane car.
The way you fill a propane car is...
Step 1 open the overflow valve on the tank.
Step 2 start transferring propane.
Step 3 when propane comes out of the overflow valve the tank is full.
Step 4 stop transferring propane.
Step 5 close the overflow valve.
The propane tank has a guage but its not used to determine when the tank is full.

I see the high knee of an LFP battery the same way.
When the voltage shoots up into the high knee its time to stop charging.
Shunts drift and they are reset by voltage.

The shunt is a better way to keep track of SOC in the flat part but voltage is a better indicator of SOC at the extremes.

 

[Ampster]

Conclusion so far is that my 16 battery modules have large ohms so it's causing the system not able to perform full watt capacity as originally designed.

Another way of stating that issue is in terms voltage sag. I agree that the Ohms are high. Can you test Ohms at different discharge rates? Could that be a result of a pack that has aged and lost capacity? There is no disagreement about the Ohms. The only hypothesis that has not been tested is the question of capacity. Especially capacity at a one kiloWatt draw.

 

[Ampster]

Build a 48volt 200ams Headway pack and parallel to my existing battery

I have seen several people do that to fix used packs. The Headways have a higher discharge rate and may be able to prop up your voltage sag. Did your vendor give you any specs on the discharge rate or some plots of discharge curves at different current?

 

[time2roll]

Is it certain to be battery ohms? Has the voltage on the cell terminal vs the 48v inverter input been compared? Could the cell to cell connections add to the issue? The drop of 5 volts seems like a lot for LFP. What is the C rate of discharge to get 5 volts drop?

 

[sunshine]

Unless there is ohms-less battery, otherwise this is a common issue.

The difference between 'charging' and 'battery' voltage does come up often.
Just using the system voltage is probably down to simplicity of comprehension and economics.
Charging batteries using capacity and current maybe is more accurate in some cases but I doubt it would be for the random masses and there is also the fact it still relies on voltage readings at the shunt to determine amperage so we are back at the original problem.

 

[Bud Martin]

My settings are like this:
49.6v batt -> grid
54.5v grid -> batt

I have been using my system as primary power source of my household, cycle once per day, grid as a backup to fill battery at cheap night hours. My battery is sized at 5kwh to be enough for 1 full cycle per day. I could have "solve" most issues by doubling or tripping battery, but that's not real solution.

As I measured, my battery system has an internal resistance of 0.05 ohms, which is too much if I want to run my system near full watt capacity.

I think the real solutions are:
1, AIO has an integrated shunt to measure and use capacity.
2, AIO allow user to input battery ohms value and dynamically calculate and compensate volt drop.

Both solutions should be very doable from hardware/software standpoint by manufacture.

For me, I need to figure out how to reduce battery resistance. And I'm going to experiment a third solution:
3, Build a 48volt 200ams Headway pack and parallel to my existing battery. That's is going to reduce overall battery ohms. That's like CPU cache memory helping RAM speed in a computer design.

BTW, where exactly did you measure the battery pack Voltage, right at the battery terminals?
Can you provide the link to the battery spec?

 

[coolbz]

Another way of stating that issue is in terms voltage sag. I agree that the Ohms are high. Can you test Ohms at different discharge rates? Could that be a result of a pack that has aged and lost capacity? There is no disagreement about the Ohms. The only hypothesis that has not been tested is the question of capacity. Especially capacity at a one kiloWatt draw.

I did test at 1kw draw and believe the drop is proportional to discharge amps. Checking BMS real time voltage and amps data proves that. I don't have a battery ohms tester, although this calculated ohms should be close, unless battery resistance heavily depend on amps discharge?

There is one pack that is at lower capacity since brand new, I know because it always peak and tank it's volt during charge and discharge cycle per BMS app data.

I suspect Battery hookup packing/shipping process partially damaged that pack, maybe one or two cells got loose or broken. So I ordered a few more cells of same type and am going to add on to that pack. That will help 16 packs to stay closer and more consistent. Overall I don't see the whole package losing kwh capacity, which is good since they are supose to cycle 2000+ times.

 

[coolbz]

I have seen several people do that to fix used packs. The Headways have a higher discharge rate and may be able to prop up your voltage sag. Did your vendor give you any specs on the discharge rate or some plots of discharge curves at different current?

There was no ohms or curves of discharge amps / volt / capacity on Battery hookup website about this custom built battery pack. They do state that this pack is rated for 1C which is 100amp max continuous discharge.

But if 1C would cause several voltage of drops out of 52volt build, it is not very good.

 

[Ampster]

unless battery resistance heavily depend on amps discharge?

I have not seen a resistance chart but certainly the voltage sag depends on discharge current on the same cells. There is a big difference between cells of different chemistries. Even LFP Headways have one type that has a much greater discharge rate in the same form factor.
Ohms law would suggest that resistance increases with current since voltage is sagging if I remember physics correctly?

 

[coolbz]

Is it certain to be battery ohms? Has the voltage on the cell terminal vs the 48v inverter input been compared? Could the cell to cell connections add to the issue? The drop of 5 volts seems like a lot for LFP. What is the C rate of discharge to get 5 volts drop?

I checked between 1st packs + and 16th packs - connectors, so there should be not other components in between. And I checked 1st packs + and BMS - connectors as well.

It is 1 C rating of 100amps. But so far I haven't tried to reach that. I just don't want risk of a power outage of my whole house. So that 5 v drop is calculated based on smaller scale testing. But I would think it is fairly accurate if I do try.

 

[coolbz]

BTW, where exactly did you measure the battery pack Voltage, right at the battery terminals?
Can you provide the link to the battery spec?

I checked between 1st packs + and 16th packs - connectors, so there should be not other components in between. And I checked 1st packs + and BMS - connectors as well.

It is 1 C rating of 100amps. Battery hookup has pulled off this listing. The pack is made of 20 32700 3.2 cells of 5ah each.

 

[coolbz]

Even with above issues, I still think Battery hookup's $220 4 packs product is very good value. Nothing is perfect especially in this newer domain. I'm eager to try some thing to improve this.

 

[Bud Martin]

I checked between 1st packs + and 16th packs - connectors, so there should be not other components in between. And I checked 1st packs + and BMS - connectors as well.

It is 1 C rating of 100amps. Battery hookup has pulled off this listing. The pack is made of 20 32700 3.2 cells of 5ah each.

Wow! you bought 16 packs of them. Can we see your setup?