Lithium gadgetry help (1 Viewer)

[haganap]

So i have the king of all set ups...

The top of the range rolls Royce of gadgets..

I'm used to much simpler things, battery amps out, solar rate in.
Battery voltage left..

Am i reading mine correct? I've got an app with loads of things on, what do they mean?

Earlier the soc was 73% within an hour back to 100 percent..

Anyone want to guide me through the numbers and what they mean.. of you need anything else I'll add it..
But I'm just looking to understand what it all means.

 

[MisterB]

What make of battery etc?

To be honest, after the initial thrill of monitoring everything with an app. I gave up as I only need to know if it provides electricity when I need it!!

 

[haganap]

What make of battery etc?

To be honest, after the initial thrill of monitoring everything with an app. I gave up as I only need to know if it provides electricity when I need it!!

They are all Victron...

 

[Ridgeway]

I would check that notification on the BMV model (I assume it‘s a BMV712). It’s probably just an update

 

[Ridgeway]

Paul is it the factory Lithium pack from NiBi ?

In which case it will be all Victron, BMS, Inverter and batteries etc

^{_{Subscribers  do not see these advertisements}}

 

[autorouter]

Battery: cell voltages are almost equal to each other, so they are well balanced. The cell voltages won't change very much between 30% and 90% charge level, but if one of them deviates from the others a lot, you know there is a problem.

BMV shunt: Ignore the '100%' state of charge, until it's had a few charge discharge cycles and has had chance to settle down. It's a calculated value, not a measured value. It's not actually at 100% because it's taking charge from the solar. It's charging at 13.25V x 4.93A = 65W.

That's what is actually going into the batteries. If the solar is generating 108W, then 108 - 65 = 43W is being used by the system, and is not going to the batteries.

The 'Midpoint Voltage Deviation' is irrelevant. It's an extra voltage reading for people who connect two 12V batteries in series to make 24V. It shows the voltage at the midpoint, ie 12V, so you can see if both batteries are charging OK. On a motorhome with a 12V system, it is redundant, but is often re-purposed to show the voltage of the starter battery.

Solar: The panels are generating 32.80V x 3.4A = 112W, the electronics is changing that to 13.44V x 8.44A = 108W to charge the batteries. There is a small conversion loss, as expected.

I don't know exactly which Victron MPPT solar you have, but it looks like it has a 'Load' output. On a motorhome it's not used, nothing will be connected, so it will make no difference whether it is on or off.

 

[Raul]

The first two pics are you battery. It tells you State of charge percentage, and whats going in the battery or out will be displayed with - minus. Also tells you how long ago was fully charged. Then individual cell voltage, confirming they are balanced.
Third pic is the shunt, the state of charge here is more accurate, and the midd point voltage is no use to your case. You can change that to monitor a second battery like starter battery. You need the + from starter connected to B2 port on the shunt to display that.
Next is your solar charger telling you current solar production at the top and solar voltage and amps, bellow is the power going in the battery after the loads.

Faster fingers above.

 

[Gromett]

My suggestion would be to put it on a mains charger for 12-24 hours to make sure it is 100% full.

The coulomb counter which counts current in and out will drift overtime and a 100% full (proper not as measured) needs to be done to recalibrate it.

 

[Raul]

My suggestion would be to put it on a mains charger for 12-24 hours to make sure it is 100% full.

The coulomb counter which counts current in and out will drift overtime and a 100% full (proper not as measured) needs to be done to recalibrate it.

Not with victron batteries, I trust that info in the app to be correct. The SOC triggers are different set in victron bms es compared to the rest.

 

[Gromett]

Not with victron batteries, I trust that info in the app to be correct. The SOC triggers are different set in victron bms es compared to the rest.

I usually 100% agree with you. But on this I can't see how they can keep a perfect count over time?

^{_{Subscribers  do not see these advertisements}}

 

[Raul]

I usually 100% agree with you. But on this I can't see how they can keep a perfect count over time?

Its hit 100% 18hrs ago, so the trigger parameters have been met no long ago, and cycled very little, then re charged again. It may not be 100% but 99% for sure.
Victrom bms have different methodology in determining the SOC so the triggers are more realistic than the rest. I have personal experience with this, and I tend to believe the data in the app.

 

[Gromett]

Its hit 100% 18hrs ago, so the trigger parameters have been met no long ago, and cycled very little, then re charged again

Ahh, I didn't realise this. Ok that is more clear now

 

[Gromett]

Its hit 100% 18hrs ago, so the trigger parameters have been met no long ago, and cycled very little, then re charged again. It may not be 100% but 99% for sure.
Victrom bms have different methodology in determining the SOC so the triggers are more realistic than the rest. I have personal experience with this, and I tend to believe the data in the app.

To add, I saw the Last Full charge 18 hours ago. But that is full charge as per the BMS I thought. Which is when it thinks it is full and that may not be it actually full if the coulomb counter is out of sync?

 

[Raul]

To add, I saw the Last Full charge 18 hours ago. But that is full charge as per the BMS I thought. Which is when it thinks it is full and that may not be it actually full if the coulomb counter is out of sync?

It doesn't rely on coulomb counter alone, it has additional parameters that needs to be met ( triggers) like cel voltage, pack voltage, tail current, temperature and time. All combined triggers determines when to re set to 100%. The drift will be minimal and adapted with use. A re set to 100% can keep drift away for several cycles.
Even bms es with less triggers can keep the drift away few cycles, untill you need to charge to 100% to have a accurate reference point again.

 

[Gromett]

It doesn't rely on coulomb counter alone, it has additional parameters that needs to be met ( triggers) like cel voltage, pack voltage, tail current, temperature and time. All combined triggers determines when to re set to 100%. The drift will be minimal and adapted with use. A re set to 100% can keep drift away for several cycles.
Even bms ea with less triggers can keep the drift away few cycles, untill you need to charge to 100% to have a accurate reference point again.

That all makes perfect sense.

But if the battery has not been fully charged recently, and has been up and down between say for example 40% and 60%. The coulomb counter could be well out by quiet a large amount as it would be operating on the flat section of the charge/discharge profile.
This would especially be the case if they are using high power equipment for short period in combination with a decent solar set up.

If you then put it on charge, it may reach 100% on the counter well before it actually reaches full real charge. I understand what you are saying about the coulomb counter resetting on cell/pack voltage triggers etc.
My question is. Say the coulomb counter things it is at 60% but the battery is actually at 50% SOC. Put it on charge. The display will show 60,61,62..... 99, and then 100% before the cell voltage trigger is reached.
Or does the victron say get to 95% where it would expect the voltage to start rising off the flat and hold at 95% until the voltage does start rising?

Either way is can be misleading. In the former, you think you are at 100% and turn the charger off and the coulomb counter reset has not been triggered.
In the latter. You see it hit 95% and think 15 minutes left until it is full and turn it off after those 15 minutes when in actually fact it may need to hold that 95% point for an hour or more?

I hope that makes sense. I am not questioning your knowledge on this as I have not looked at this in a very long time and my practical experience is limited. But I do want to understand this.

Thanks for taking the time

^{_{Subscribers  do not see these advertisements}}

 

[autorouter]

The coulomb counter could be well out by quiet a large amount as it would be operating on the flat section of the charge/discharge profile.

The whole point of 'coulomb counting' technology is that it doesn't use the charge/discharge voltage profile at all. It counts the charge in and out of the battery.

Charge is measured in coulombs. But with motorhomes it is usually measured in amp-hours. One coulomb is the charge that passes when one amp flows for one second. So it's an amp-second. Since one hour is 3600 seconds, one amp-hour = 3600 amp-seconds = 3600 coulombs.

As the amps varies, second by second, the BMV calculates the amp-seconds and accumulates the total charge going in or out of the battery. It reports this in amp-hours, because that's what customers want, but it's measuring coulombs and converting.

The amp-hour count depends only on the amps and the time, and voltage is not necessary for this. It remains accurate even when the voltage increases during charging, and when it decreases due to loads.

 

[Raul]

I had few beers, I will come back to this to contribute in a meaningful way. One thing I can tell you for sure is: the current calibration sensor. Coulomb counting relies on accurate flow reading. The victron gear for this reason does not allow current calibration on their bms. I will return later.

 

[Gromett]

The whole point of 'coulomb counting' technology is that it doesn't use the charge/discharge voltage profile at all. It counts the charge in and out of the battery.

Charge is measured in coulombs. But with motorhomes it is usually measured in amp-hours. One coulomb is the charge that passes when one amp flows for one second. So it's an amp-second. Since one hour is 3600 seconds, one amp-hour = 3600 amp-seconds = 3600 coulombs.

As the amps varies, second by second, the BMV calculates the amp-seconds and accumulates the total charge going in or out of the battery. It reports this in amp-hours, because that's what customers want, but it's measuring coulombs and converting.

The amp-hour count depends only on the amps and the time, and voltage is not necessary for this. It remains accurate even when the voltage increases during charging, and when it decreases due to loads.

Sorry if you didn't know this. But I do know exactly what a coulomb counter is, My original trade and qualifications are electrical engineering/electronics.

BUT as you also know you do not get out of a battery what you put in. What you can get out also varies by how fast you take it out.
So coulomb counting on it's own is not a satisfactory even medium term method of keep track of SOC. You need to be able to resync the counters.

My questions here are specifically about how it is implemented at Victron?

 

[haganap]

Paul is it the factory Lithium pack from NiBi ?

In which case it will be all Victron, BMS, Inverter and batteries etc

Its the 10k factory fitted option..

 

[Raul]

Good morning All.
Taking Gromett example above I will try to explain what’s happening with the SOC in the midd range, partial cycle and accuracy of SOC.
Victron bms uses various parameters to determine the SOC at any given time, and also at the top, we all know it resets to 100%, (all bmses) as it knows for sure, due to... voltage. Yes it sounds bizarre as voltage is meaningless in the flat curve, but once you out of the flat curve, voltage IT IS used to determine SOC accurately. For example, at 2.5v soc is 0, and at 3.65 with a 0.05C charge, is full 100%, at 25 degC. Close to this values, is also close for the SOC to be true. Victron changed from 14.4 to 14.2v because, not everyone can supply a 0.2C charge and terminate charge at 0.05C. Many will end up overcharging, by pinning the cell at 3.65v beyond 0.05C down to 1-2amp instead of 5 cut of for a 100ah pack.
Now, coulomb count is used in the bms and bmv, coulomb is acurate and true to 1:1 efficiency, as we increase or decrease charge/ discharge, efficiency is no longer 1:1 and coulomb count does not know or nor can correct that. That’s the job of a Peukert exponent value. The LFP has a peukert almost 0 when new, or in exponent value 1. As the cell ages, this value increases to max 1,05 exponent. What happens in between?
Well, victron sets the gear at 1.05 to account for a wide variation of use and cell age.
As you start with a fully charged pack, you may use a 0.01C or less, and no need for a peukert correction value. But you also discharge with 0.5 or 1C, and the 1.05 exponent is needed to account for that inefficiency. As you guessed, not even a full cycle can be measured accurately by coulomb alone, due to variation of charge /discharge rate. In the lab we set a fixed value and work from that. In real life, can be a phone charge or TV, to as large as a kettle or else. So, in real life we need a non linear peukert value to help the coulomb count. Well, we can’t have it, so we need to set a value to our use, age of cells, and equipment matching criteria. A system with 0.2C charging capability, is no equal to a system that can do 0.5C. Also it’s ok you meet the charging process, but, you screw it by discharging, as many can do 0.5-1C discharge. By having the peukert at 1.05 as set by default, you making the bms bmv calculating you need a bit more than 1:1 to put in, and you actually taking out a bit more than 1:1 when discharging. All of this done by microprocessor , results in a pretty accurate SOC , also given the definite correction by voltage at the top it will make it almost 100% acurate. The voltage and temp is important, as at the top reflects the true state. Temperature is also a factor, take a fully charged rested battery at 13.25v and 25C ambient, and stick it in a fridge/ cooler, the voltage will decrease, and resistance will increase. It’s the nature of the electrolyte behaviour.
To resume, how we determine SOC mid range with partial cycling? We don’t, we use best setting relevant to our use and equipment, and change the peukert as the cell ages. The bms will be out by a bit, but not much. The more you cycle the more data it will have to work with. Victron claim the bmv is learning and adapting, it’s simply having more history ( more cycles) it can calculate a average bases on more values given by various cycles. It’s a complex subject, and not easy to explain, without breaking down, but I hope this can help out to see beyond the settings we put in, and what effect they have.

My settings are as: 3.5v full, absorb 15mins, no tail current, charge efficiency 99%, peukert 1.03 for (high loads), if I would not had a inverter, I would set peukert to 1.02. I did started with 1.02 but it creeped up on my 5 1/2 year old batteries, and the 1.03 now is accurate to under 1% error.
Thank you for taking the time to read a lengthy post.

^{_{Subscribers  do not see these advertisements}}

 

[Gromett]

Excellent post and exactly as I would imagine how it operates. I am a huge fan of Victron gear and it is what I use when doing new builds. scotjimland got one of my old multipluses when I upgraded. I can't remember the year but think it was 2007? I have been using Victron since 2004. They are simply the best.

To resume, how we determine SOC mid range with partial cycling? We don’t, we use best setting relevant to our use and equipment, and change the peukert as the cell ages. The bms will be out by a bit, but not much. The more you cycle the more data it will have to work with. Victron claim the bmv is learning and adapting, it’s simply having more history ( more cycles) it can calculate a average bases on more values given by various cycles. It’s a complex subject, and not easy to explain, without breaking down, but I hope this can help out to see beyond the settings we put in, and what effect they have.

However, still doesn't answer my question? If the battery is operated in the middle of it's range for an extended period of time with varying charge rates and varying discharge. What happens when it is being charged up after having the SOC counter pushed out of sync with reality.

You mention here that "the more you cycle the more data it will have to work with". But I think this only applies where a cycle is terminated with a full charge? Otherwise there is no way to do those calculations and get better with time. Without a set point where values are known there is no way to do the calculations to become better with time.

I will give an example. Someone buys a motorhome, installs lithium and large amounts of solar. They store it at a storage facility without hookups. When they use it they never go on hookup they are freeloaders . So this van never sees an electrical hookup and relies completely on solar and a B2B. They carry a generator just in case. Their use case is they travel all over Europe from the Norway down to Spain. They do two big trips a year, a winter trip and a summer trip. Sometimes the charge rate will be the maximum of their solar panel and the B2B. Other times when parked they may have very little coming in from their solar. They have varying usage. Normal steady usage such as TV, lights, water pump. But they also have an induction hob, wife's high power hairdryer etc. So there is no consistent charge rate, and no consistent discharge rate with high power surges on the discharge side.
If this is all done on the flat of the curve during an extended 3 months 120 day break in the EU. So the battery has for an extended period of time been down as low as 14% and as high as 89%. They are careful not to run it below 10% but when they see it getting up towards 90% they use that opportunity to get jobs done such as charging their e-bikes, wife's hair dryer use and using electric to heat the hot water to save diesel. They have an 800W 10L propex 240v water heater this takes 30 minutes to get up to temperature.

You may think the is a made up scenario to make a point, but it isn't far from the truth of a few fulltimers I know.

So my question is. The SOC gauge on this system will drift over time I am sure you would agree? If this is the case. How does it handle and display the data to the user when charging up to full after this extended trip. Will it go from say 50%, 51%..... 99%, 100% then keep charging silently. Or does it starting hitting the curve and halt the displayed 95%, 96%.... holds here until a set voltage is achieve then starts counting up again as the current falls below a certain level? Or is it using the second part of the CCCV to dynamically predict the last part of the charge as it is happening. So the display will be roughly in line? But then risks having a period where it holds at a specific percentage for quite a while.

I hope I am explaining my question well enough? And I do appreciate the time you are putting in to answer this. Thanks

 

[Raul]

Excellent post and exactly as I would imagine how it operates. I am a huge fan of Victron gear and it is what I use when doing new builds. scotjimland got one of my old multipluses when I upgraded. I can't remember the year but think it was 2007? I have been using Victron since 2004. They are simply the best.


However, still doesn't answer my question? If the battery is operated in the middle of it's range for an extended period of time with varying charge rates and varying discharge. What happens when it is being charged up after having the SOC counter pushed out of sync with reality.

You mention here that "the more you cycle the more data it will have to work with". But I think this only applies where a cycle is terminated with a full charge? Otherwise there is no way to do those calculations and get better with time. Without a set point where values are known there is no way to do the calculations to become better with time.

I will give an example. Someone buys a motorhome, installs lithium and large amounts of solar. They store it at a storage facility without hookups. When they use it they never go on hookup they are freeloaders . So this van never sees an electrical hookup and relies completely on solar and a B2B. They carry a generator just in case. Their use case is they travel all over Europe from the Norway down to Spain. They do two big trips a year, a winter trip and a summer trip. Sometimes the charge rate will be the maximum of their solar panel and the B2B. Other times when parked they may have very little coming in from their solar. They have varying usage. Normal steady usage such as TV, lights, water pump. But they also have an induction hob, wife's high power hairdryer etc. So there is no consistent charge rate, and no consistent discharge rate with high power surges on the discharge side.
If this is all done on the flat of the curve during an extended 3 months 120 day break in the EU. So the battery has for an extended period of time been down as low as 14% and as high as 89%. They are careful not to run it below 10% but when they see it getting up towards 90% they use that opportunity to get jobs done such as charging their e-bikes, wife's hair dryer use and using electric to heat the hot water to save diesel. They have an 800W 10L propex 240v water heater this takes 30 minutes to get up to temperature.

You may think the is a made up scenario to make a point, but it isn't far from the truth of a few fulltimers I know.

So my question is. The SOC gauge on this system will drift over time I am sure you would agree? If this is the case. How does it handle and display the data to the user when charging up to full after this extended trip. Will it go from say 50%, 51%..... 99%, 100% then keep charging silently. Or does it starting hitting the curve and halt the displayed 95%, 96%.... holds here until a set voltage is achieve then starts counting up again as the current falls below a certain level? Or is it using the second part of the CCCV to dynamically predict the last part of the charge as it is happening. So the display will be roughly in line? But then risks having a period where it holds at a specific percentage for quite a while.

I hope I am explaining my question well enough? And I do appreciate the time you are putting in to answer this. Thanks

Understood, the partial cycles are cumulated to count full cycles. A full cycle is the value you enter as discharged floor, or max ah capacity . If you have a discharge floor setting that will be used, if not, then total capacity is used.
In the event of 100ah capacity, and cycled between 50-75% SOC, a 25ah a day over 4 days will cumulate one full cycle. This cycle will inherit its errors or accuracies based on settings and the use rate of that energy.
When you are in a position described above, and not able to meet a full charge for a very long time, accuracy WILL drift, by how much? Its a function of settings and type of use. The coulomb count is king in this situation and discrepancy can only be out just by the rates outside of peukert value. Not one will be the same. What you are left with, is one reference : the low knee. As the top is accurate, so is the low voltage coming out of the flat curve. The cell will deep drastically below 3v. At 3v you should stop discharging and reserve that energy left for must needed tasks until you can recharge.
At 3v per cell you have about 5% SOC. enough for critical equipment at low discharge rate.
Also, in partial state of charge, the bms, will sample cell and pack voltage to adjust its SOC value, you can observe this by charging a pack, and see the the SOC value changing specially at low voltages. Some bms does better than others, the industrial type is more focused on high power, and some are more accurate due to components and multi trigger settings. Some will trigger SOC on voltage alone, some with a combination of data. A 200a bms will be pants at low amps a curacy 1-2A, but pretty close at 100-150A.
And the reverse is the same. A 100A bms is best calibrated at 50A , so it has same consistency on both ends. If you calibrate a 200a bms at 5A, it will be out at 100-120. If you calibrate it at 150 it will be out at 5A.
Large bms has large tolerances and you will need a shunt bmv to help out.
Multiple smaller bms tend to be more accurate. Note they can only account for coulomb 1:1, and the absence of peukert correction is compensated by pack and cell voltage. A shunt will have all on board and calculate in real time, within settings values. Anything outside settings values, will contribute to error cumulation. A outside setting value, can be a discharge of 1C with a peukert exponent of 1.02, for that load, you will need a 1.05. But the reverse is also outside value, so we operate with a compromise, as not all energy departs the battery at same rate. Its more important the percentage of energy how its used: 20% at 1C and 80% at 0.01C and bellow, will ned a low peukert like 1.02-1.03. And the reversal is also true: 80% at 1C with 20% at 0.01 will need a high peukert 1.05.
Generally LFP below 0.2C is almost 99.9% efficient when new. It degrades with use.

 

[haganap]

So, am i fully charged this morning or not?
Is that what the 100pc means?

 

[Raul]

So, am i fully charged this morning or not?
Is that what the 100pc means? View attachment 896240

The voltage does not reflect full charge. To be full at idle needs 13.25 ish minimum. Your settings are out, giving a wrong estimate. Also you have a notification for update, don't ignore it.

 

[Lenny HB]

I'm feeling sorry for Paul haganap I think you lot might have fried his brain.
I bet he wishes he had never asked the question now.

^{_{Subscribers  do not see these advertisements}}