Hi @plonkster and others,

What is your advice about floating Pylontechs?

I saw elsewhere that you have Victrons float them at 52v.  You know that Pylontech says to use 53.2v.

I had set my inverter to use 51.5v for bulk and float, and then to turn off once the charge current fell <1A for 30 mins.  I'd like to turn the charger if possible since the inverter fans go off.

My settings get the batteries to 90% SOC according the the info from the BMS on the RS485 port.  I'm not sure if that's good or bad, but I do see that the SOC of the two batteries drift away from each other over a few cycles - 92% and 88%.  So I guess this proves your point about using enough voltage to allow the batteries (and cells?) to balance.

If the charger goes off then the batter SOC drifts down, presumably since the BMS must be powered.

On the weekend I discharged them down to 35% or so and them charged them back up to 53.2v and their SOC is now the same again.

At the moment the batteries are advertising an SOC of 87%, an upper voltage limit of 53.25v, and a max charge current of 10A.  But battery amps is reporting 0A.

So I think I should follow you up to 52v to try to get the batteries to stay in balance.

But - do you float them at that or turn off charging after a while?  And how long a while?

General stuff about LiFePO4 cells says its best not to overcharge, and if they are just to sit they do best only partly charged and without a float voltage holding them at full.  At the moment I'm not cycling the batteries and I don't need their full capacity so I'm not sure whether to leave 

I'm definitely over-analysing this - I suspect that the BMS in any event doesn't actually charge the cells once the SOC is 100% so probable I'm worrying about nothing?

So maybe I should just relax, but these batteries were not cheap!

Thanks,
@Elbow

 

 

45 minutes ago, Elbow said:

I saw elsewhere that you have Victrons float them at 52v.  You know that Pylontech says to use 53.2v.

Right. That's around 0.1V per cell lower than they specify. Really no biggie.

46 minutes ago, Elbow said:

But - do you float them at that or turn off charging after a while?  And how long a while?

Float. And it is no real issue. If you maintain the bus at 52V... and the batteries are at 52V, then no current can flow because there is no potential difference.

Most BMSes use more than one measurement to determine the SoC, so they may report a lower SOC when the cells aren't completely balanced. It can also take up to two weeks for them to be completely balanced (or that was my experience with my own LFPs). Just give it some time and don't worry so much 🙂

15 hours ago, Elbow said:

What is your advice about floating Pylontechs?

Still trying to form an opinion about this, can only share my observations and data collected so far.

After reading your post (very-long-article-about-charging-lfp-batteries) , which was very interesting and informative thank you. I've set to experiment.

I'll share some data screenshots for now, and edit to add my observations. 

tldr; it appears that most productive values for me where in that 52V range.

 

So what is 'float' exactly? Is that just keeping them in a state where they are balanced against their self discharge?

13 minutes ago, DeepBass9 said:

So what is 'float' exactly? Is that just keeping them in a state where they are balanced against their self discharge?

Pretty much. Except that you are also (usually) powering loads while floating the battery; the float voltage is the one for the charging system to aim for while loads come and go. The battery was full at the end of the absorb stage, and in the float stage it stays close to full while there is still solar power available.

Suppose the load exceeds available solar power for a while. When the sun comes back or the load drops off, the charging system keeps charging the battery until the battery voltage reaches the float voltage setting. That's when it knows that it's replaced the deficit, without letting the battery voltage get too high for too long. Keeping the battery voltage too high for hours on end is not good for battery life, and there are only a few percent more SOC at the higher voltage anyway.

I favour a decent gap between absorb (bulk) and float voltages; I use 55.2 and 53.7 V for my 16S LFP battery. The equivalents for 15S (PylonTech and some others) are 51.8 and 50.3 V. So the usual 53.2 and 53.0 are both quite high. 53.2 and 52.2 as above at least have a bit more of a gap between absorb and float voltage settings.

What @Jaco de Jongh and I are interested in is what kind of SoC the Pylontech batteries show at various voltages. For example, on the weekend we noted that  abattery held at 52V claimed an SOC of 93%, which to me seemed a bit low.

But then I also know that for most BMSes the SOC isn't simply calculated from the overall voltage, it is instead calculated from a number of internal cell-level measurements, that is to say that when it reports 93%, that is the SOC of the lowest cell, and that over time as it balances, it might show a higher SOC at the same voltage.

Could anyone confirm this theory for me, perhaps by simply telling me what kind of SOC the BMS reports at 51.8V, 52V, and so on.

Not certain on the BMS in the Pylontechs, but what is the 100%SOC . Wouldn't the BMS then swing into action in balancing the packs , I don't thing you will get the battery to show 100%SOC unless the BMS has been able to cycle and do its balancing. On the Batrium BMS you can set the battery full voltage per pack where it will go into balancing the individual packs. If you don't reach the full charge on the Pylontechs does the BMS still go into its balancing mode, and then do you run the risk of the packs in the battery going out of balance over time?

7 minutes ago, seant said:

If you don't reach the full charge on the Pylontechs does the BMS still go into its balancing mode, and then do you run the risk of the packs in the battery going out of balance over time?

The balancer activates before 100%, as soon as the overall voltage is high enough. 51.8V is enough for the balancer to start working, that is the info I have, but you're always welcome to check it.

8 hours ago, plonkster said:

Could anyone confirm this theory for me, perhaps by simply telling me what kind of SOC the BMS reports at 51.8V, 52V, and so on.

51.5v gets a report of around 90% - at that voltage my one battery said 92% and the other 88%.

(In my case I was turning the charger off quite quickly once the charge current dropped - so I was probably defeating the balancing - looking at the trend over a few cycles they do appear to drift out of balance)

I'm pretty sure we should take @plonkster's advice and use 52v

Elbow

14 hours ago, Coulomb said:

The equivalents for 15S (PylonTech and some others) are 51.8 and 50.3 V. So the usual 53.2 and 53.0 are both quite high. 53.2 and 52.2 as above at least have a bit more of a gap between absorb and float voltage settings.

Will certainly test out those values next master @Coulomb 

I actually had a perfect test case today. At 52V the BMS was reporting only 94%. So I quickly adjusted the voltage up to 53.2V to see what would happen, and this is what happened: Within seconds the voltage went up to 53.2V and the SOC practically jumped to 100%. Then we dropped the voltage down to 52V... and as the voltage floated down to 52V the SOC dropped to 99%.

So which one is it, dear BMS, 94% or 99%?

We'll be watching it tomorrow, as it cycles down tonight and back tomorrow to see where it settles.

54 minutes ago, plonkster said:

So which one is it, dear BMS, 94% or 99%?

My guess is that it coulomb counts, but there is a little drift in the current measurement, plus all the difficulties of guessing SOC, with temperature affecting it, hysteresis, internal resistance, and so on. 

At a certain voltage, it resets the coulomb counter to zero, and hence the SOC to 100%.

So to answer your question, it was 99% but didn't know it yet 🤔. 

Edit: or it really was closer to 94%, and it expects you to keep it at a higher voltage until the current reduces to a low value, so that the battery really is full and the 100% SOC it reports is justified. 

Edited February 6, 20197 yr by Coulomb

Very similar observation on my end as well, it was reporting 89% than it had a little spike 53.8 and said ok im full now ...

Will do thired day of measure before I try @Coulomb suggestion

15 hours ago, Padwan said:

The equivalents for 15S (PylonTech and some others) are 51.8 and 50.3 V.

 

 

 

 

24 minutes ago, Padwan said:

Very similar observation on my end as well, it was reporting 89% than it had a little spike 53.8 and said ok im full now ...

Will do third day of measure before I try @Coulomb suggestion

> The equivalents for 15S (PylonTech and some others) are 51.8 and 50.3 V.

Since 51.8 V is the voltage where balancing starts, you might want a smidge more than 51.8 V (say 52.0 V) for the absorb voltage. Just so your're sure that balancing happens most days. On a totally cloudless day, an Axpert could exit the absorb stage a few tents of a volt under the setting. Even though most days it will overshoot, you can't rely on that 🙃

1 hour ago, Coulomb said:

Since 51.8 V is the voltage where balancing starts, you might want a smidge more than 51.8 V (say 52.0 V) for the absorb voltage. Just so your're sure that balancing happens most days. On a totally cloudless day, an Axpert could exit the absorb stage a few tents of a volt under the setting. Even though most days it will overshoot, you can't rely on that 🙃

looks like 52 is new 42 ....  😮

I looked in again at this site I mentioned earlier. Today it leveled out at 96%. I slowly pushed the voltage up, and at 52.6V it registered 100%.

I'm actually perfectly happy with any number above 95%.

@plonkster @Coulomb  My pylontechs (4XUS3000) has not been above 88% this week.  And every day the maximum SOC after long periods of "float" at 52V is getting lower.  Today it has been kept at 52V for almost 5 hours, and still the SOC shows only 84%. System: 6000W panels, 250/100 Victron MPPT, MP II 5000, running in ESS mode.

 

There is a longer term graph of the battery SOC:

 

Surely this is not right?  The Pylontechs should be cel-balanced by now?  Why is the maximum SOC getting lower and lower each day? 84% seems pretty low.

 

 

 

Edited November 16, 20196 yr by Ironman

1 hour ago, Ironman said:

Surely this is not right?  The Pylontechs should be cel-balanced by now?  Why is the maximum SOC getting lower and lower each day? 84% seems pretty low.

OK, first off... do not be concerned. The battery is actually full (or very close to full anyway). What is going on is that the SOC is incorrect. It has drifted and the BMS needs to reset its SOC.

I've been pondering this voltage for a while now. I know 52V is too low to trigger the SOC-reset of the BMS, so I've been wondering what the minimum voltage is that would cause the BMS to reset.

My most recent tests did not suggest a sufficient reason to increase the voltage. My hope was that a slightly elevated voltage (52.5V, 3.5V per cell) would do slightly better. But in my tests 52.5V didn't cause a reset. The battery still stopped short of 100%, then drifted upwards slowly and ended up in the high 90s.

But now we have the perfect case of a battery with a large SOC discrepancy, and if you don't mind, I'd like to cash in on it and slowly increase the charge voltage over a few days to see where it ends up. You up for that? If you are, please send me your VRM url. I possibly still have it somewhere... but better to be sure.

1 hour ago, plonkster said:

My most recent tests did not suggest a sufficient reason to increase the voltage. My hope was that a slightly elevated voltage (52.5V, 3.5V per cell) would do slightly better. But in my tests 52.5V didn't cause a reset.

Till a few weeks back the Victron Pylontech document stated 53.2 and 53 volts for charging. Why was it the changed to 52 volts? @Ironman's new Multi is configured according to the new  document, Will it not be better to revert to the old settings?

His batteries were charged at the old setting for more than a month, will this effect where it resets and what it reports now?

11 hours ago, plonkster said:

But now we have the perfect case of a battery with a large SOC discrepancy, and if you don't mind, I'd like to cash in on it and slowly increase the charge voltage over a few days to see where it ends up. You up for that?

11 hours ago, Jaco de Jongh said:

Till a few weeks back the Victron Pylontech document stated 53.2 and 53 volts for charging.

Please do not take this the wrong way - I appreciate your offer to try to find a solution ourselves, but I find myself in a position where It feels like a guessing game.

 

12 hours ago, plonkster said:

I've been wondering what the minimum voltage is that would cause the BMS to reset.

The Pylontech batteries run on firmware with a set of rules. Somebody with proper knowledge of the programming and operation of the pylontech firmware will be able to tell us exactly how to solve this. Do we, as a market of pylontech consumers, not have a large enough market share to get them to be involved in this?

if we increase the voltage, at some point it might cause a SOC reset, or it could start damaging the batteries. Which will happen first?  And will Pylontech honor my guarantee?  Both might happen - we could hit the SOC reset, but damage the batteries over a period of time.  
Or is there a fault in the Pylontech firmware? Should they not be the ones to fix this issue? If the SOC reset is set too high in the Pylontech firmware, they should solve it.   Let us just think this issue through and try to get pylontech involved before acting.  Like you said - it is just a BMS SOC reset, not a urgent issue. Does not have to be solved today.

These batteries are a large part of my investment and they need to outlast the guarantee. If you, as Victron, have technical specs or guidelines from Pylontech about exactly what voltage will not damage the battery, we can go ahead. Please let me know.

13 minutes ago, Ironman said:

It feels like a guessing game.

Absolutely no guessing needed. Look at the Datasheet from Pylontech, Charging voltage 52.5 to 53.5Volt. Very clear. Now Victron has decided to use lower voltage for their own reasons... I am still of the opinion that is is perfectly safe for your batteries and guarantee/warranty to follow the manufacturers specification. 

If Pylontech say that you can charge at 52.5 V, then surely the SOC should reset when that voltage is reached, with a suitable tail current.

The big problem to me is that the manufacturer, Pylontech, has not specified the charge termination conditions. Or for that matter, what voltage to float the battery at after charging, for maximum life and reasonable energy storage. If they truly believe that you don't float these cells, that's remarkable to me, and they should be documenting why they recommend this.

@Jaco de Jongh Agreed.  According to the Pylontech sheet, the voltage can be higher. Thanks. 

It could be that Victron is setting the float voltage too low since they have to make provision for other modes of operation.  

The multiplus is a flexible inverter and can be used in many system configurations.  Some of these configurations could cause the battery to need a lower float voltage setting than in my case. 

Where is the control voltage being measured? At the output of the MPPT? At the input of the Multiplus? Or from the Pylontech BMS? If they are all available, which one is chosen for the control loop?

@plonkster I will send you the VRM stuff. Please go ahead. I was being over cautious. 

10 minutes ago, Coulomb said:

The big problem to me is that the manufacturer, Pylontech, has not specified the charge termination conditions. Or for that matter, what voltage to float the battery at after charging, for maximum life and reasonable energy storage.

My point exactly.  Maybe they don’t know themselves, since the product has not existed for the expected lifecycle. Maybe it varies with temperature?  
Bottom line, I feel they should give better information regarding this. Here we have two developers that are in charge of (excuse the pun) the charging algorithms of these batteries from different charging systems and they don’t have proper guidelines for the resetting of the SOC.

Edited November 17, 20196 yr by Ironman

Hi guys,

@Jaco de Jongh is right: Pylontech US3000 should be charged to 52.5 at least.

 

Details:

Why 52.5V? Because the internal balancers kick-in at 3.480V per cell which translates to 52.2V. The 0.3V difference is the space that allows balancers to operate safely.
If you charge to 52V only (3.466V per cell) then the balancers won't start and the pack will become ubalanced over a long period of time.
If you will charge to 53.5V, then the balancers will have to burn much more energy and the temperature of the pack will rise. But still okay.
If you will charge to 53.6V, then the internal overvoltage protection will throw an error and the battery will disconnect itself.

Beware,  that even if you set your charger to 53.5V, there still will be a short spikes of power that will push the voltage over, based on the changing light conditions. Therefore, using the upper voltage limit is not a good idea. By the way, these spikes of power are the reason why in the past you've experienced 100% SOC even when charging to 52V only. Looks like the light conditions has changed since, and now there's not enough power spikes to push the voltage over and start the balancers.

 

@Ironman if you want to know more about Pylontechs, just check my LAB, the link is in my signature.