Victron & Phylontech High Voltage alarms

[Kilowatt]

Hi Guys,

I have a fairly new Victron Multiplus ii installation that has been operational for about 3 or 4 weeks now. Up until 3 days ago it has been operating without any issues but now, i am receiving high voltage alarms on my system. I have not made any changes to the system, so i am unsure why i am receiving the alarms or what the cause could be. Any idea's of what to check? 

The high voltage is set to 52v and the max charge voltage is set to 53.2v although its never reaches that high. It feels to me that the MPPT is maybe not reacting fast enough?  As far as i know i am on all that latest firmware versions on everything. The battery is listed under my device list. 

The voltage spikes to just higher than the 52v setpoint, like 52.16v and it also gives an alarm at just under 52v,  so i don't think i am actually causing any damage to the batteries and i have not noticed any alarm lights on the batteries either. 

Any help would be appreciated. 

[Guest]

@plonkster is your best bet here.

Or whomever you bought it from also an option for help.

[Jaco De Jongh]

5 hours ago, Kilowatt said:

The high voltage is set to 52v and the max charge voltage is set to 53.2v

There is your problem. Under normal conditions for your Pylontechs, to be seen as 100% full, should reach at least 53 volts. What you are saying above is telling me your batteries are not reaching 100%. Please read through the attached document. 

victron_pylontech_us2000_us3000_and_phantom-s.pdf 

 

[deapsquatter]

Although that doc also says this:

Quote

When DVCC is enabled, the battery (via the CAN-bms) is responsible for the charge voltage. The Pylontech battery requests a charge voltage of 53.2V. We have however found that in practice this is too high. The Pylontech battery has 15 cells in series, so 53.2V equates to 3.55V per cell. This is very highly charged and makes the system prone to go overvoltage. It should also be noted that a LiFePO4 cell stores very little additional energy above 3.45V. For this reason we opted to override the BMS and cap the voltage at 52V. This sacrifices almost none of the capacity and greatly improves the stability of the system.

Mine hovers around 52V and reports 99% and sometimes 100%.

Edited July 31, 2019 by deapsquatter

[Jaco De Jongh]

4 minutes ago, deapsquatter said:

Mine hovers around 52V and reports 99% and sometimes 100%.

That is not far off, but having the high alarm set under max charging voltage cant be right, I remember the Pylons sort of "Self Calibrate" itself with Victron to "100%" below max volts. 

[Kilowatt]

5 minutes ago, Jaco de Jongh said:

There is your problem. under normal conditions for your Pylontechs, to be seen as 100% full, should reach at least 53 volts. What you are saying above is telling me your batteries are not reaching 100%. Please read through the attached document. 

victron_pylontech_us2000_us3000_and_phantom-s.pdfUnavailable  

Thanks Jaco, I will go through the recommended settings as per the document. I don't have an issue with the battery not reaching 100% full as i understood this is because of the 52v setpoint. but i am just trying to figure out what's changed or what the cause for the high voltage alarm is now, when this has not been an issue before. Feels like the battery and the MPPT are not communicating properly. I agree with you that the high alarm set under max voltage cant be right so will investigate this. 

 

[Kilowatt]

18 minutes ago, deapsquatter said:

Although that doc also says this:

Mine hovers around 52V and reports 99% and sometimes 100%.

Are your parameters the same as this? 

Parameter Setting Battery type Lithium Charge curve Fixed Absorption voltage 53.2 V Float voltage 53.0 V Absorption time 1 Hr

[deapsquatter]

Pylontech Battery / Parameters shows me:

 

Running ESS - solar panels not connected yet (in case that changes things).

Edited July 31, 2019 by deapsquatter

[stoic]

34 minutes ago, Kilowatt said:

I don't have an issue with the battery not reaching 100% full as i understood this is because of the 52v setpoint

Mine always reaches 100% ... Victron/Pylon setup

[deapsquatter]

What does Settings / System Setup look like? I have:

Edited July 31, 2019 by deapsquatter

[___]

1 hour ago, deapsquatter said:

Although that doc also says this:

I wrote that part of the document, because too many people constantly asked why it stops at 52V!

While the BMS does reset the SOC to 100% at around 53V, it isn't absolutely required to take it that high. It will also reach 100% at a lower voltage, it just takes a bit longer.

The problem in this case, however, is that there is a high voltage alarm at a much lower voltage (3.45V per cell), which could possibly indicate an imbalance. One thing you might try is to configure the Multi to 53V, and the solar chargers to 53.2V, then turn off DVCC (and do a VE.Bus reset) and let the Multi/Solar charge the battery nice and high. Perhaps it needs a kick on the high side to balance properly? I was assured that balancing can happen at 52V (3.45V per cell) though.

[Youda]

10 minutes ago, deapsquatter said:

When DVCC is enabled, the battery (via the CAN-bms) is responsible for the charge voltage. The Pylontech battery requests a charge voltage of 53.2V. We have however found that in practice this is too high. The Pylontech battery has 15 cells in series, so 53.2V equates to 3.55V per cell. This is very highly charged and makes the system prone to go overvoltage. It should also be noted that a LiFePO4 cell stores very little additional energy above 3.45V. For this reason we opted to override the BMS and cap the voltage at 52V. This sacrifices almost none of the capacity and greatly improves the stability of the system.

As an owner of 9x Pylontech US3000, I'm charging them to the 52.5V and it works great.
The above quote is exactly what I'm saying for a couple of months already:

• Max charging voltage that is recommended in the Pylontech specsheet is way too high.
• Recommended low-voltage cut-off at 48V is a non-sense for 15S LFP.
• Recommended charging current 0.5C is exactly on the edge of LFP's longevity.

In detail:
Charging the 15S LFP up to 53.5V (3.57V per cell) is hurting the batteries and the internal balancers will produce more heat too.
Low-voltage cut-off at 48V equals to 3.2V per cell, which is exactly the nominal voltage of LFP. Means, that the inverter will cut-off the batteries while there's still a lot of charge in them.
 

Therefore, I do recommend charging to 52.5V, low-voltge cut-off at 46V and 0.5C of current for the Pylons.

Bear in mind that not everything, what's written in a manufacturer specsheets and SoPs, is worth to follow blindly:

 

 

[deapsquatter]

Not sure why mine stays at 99% in Absorption for so long - don't think I've ever seen it in float. It also uses a few Watts from the battery and then puts them back again every now and then (see image). I'm set to Keep Battery Charged and solar panels are not yet connected to MPPT.

 

EDIT: Looking at VE Bus State in VRM and its always in Absorption.

Edited July 31, 2019 by deapsquatter

[___]

3 minutes ago, deapsquatter said:

don't think I've ever seen it in float

There's no such thing as "float" outside the realm of lead acid batteries There's just the charge voltage the battery wants you to use. Some batteries just spew out a single number (BYD == 56.5V, Pylontech == 53.2V, etc), and some will actually move the voltage up and down to compensate for an imbalance or some other condition. The REC BMS can do this for example, also the Lynx Ion from Victron, and I recall seeing Sony/Murata also doing this properly.

[Jaco De Jongh]

5 hours ago, Jaco de Jongh said:

5 hours ago, Kilowatt said:

The high voltage is set to 52v and the max charge voltage is set to 53.2v

There is your problem.

Seems like I was wrong here, although I have never seen a high votage alarm setting in a Victron system, I thought that it should be available somewhere if it is reported to be set at 52 volts. On going through the system, I could not find any place to set a high voltage alarm. After consulting with @plonkster it was realized that the alarm was issued by the Pylontech itself. 

It is still a bit of a mystery why it will complain at a reported low voltage of 52 volts. We ran the AC charger at 53.2 volts for over 30 minutes without triggering the alarm again. I will watch it tomorrow to see if the alarm pops up again.  

[Kilowatt]

I just want to say a big thank you too all those who are so willing to pass on advise and help, but in this case especially to Jaco De Jong. I am blown away at the commitment and time he spent last night trying to resolve a problem for me ( An absolute Stranger). 

Thank you very much. 

[___]

12 hours ago, Jaco de Jongh said:

It is still a bit of a mystery why it will complain at a reported low voltage of 52 volts. We ran the AC charger at 53.2 volts for over 30 minutes without triggering the alarm again. I will watch it tomorrow to see if the alarm pops up again.  

My theory is that the high voltage alarm applies on the cell level, not on the overall voltage. This means there was a high cell. Unfortunately we can't check the per-cell voltages (well you can, but then you need the RS485 cable and software for that, or at least need to know how to extract that info via the serial console).

So why did we develop an imbalance? Maybe (at 52V) the balancers might not be as effective as they are higher up. Now I was told by the manufacturer (or at least someone that works there) that 51.8V is high enough to start balancing, and there are certainly many here who run with 52V without any problems. So this was a shot in the dark, a wild guess, and so the plan was simply to push the voltage up and see what happens (either the problem gets worse, or it goes away). What I did, because @Jaco de Jongh asked me, is patch the 52V limit up to 53V. The idea is to let it run like that for a few days (I assume you cycle the batteries?) and see if the problem (or others) returns.

Note that the 53V patch will disappear if you upgrade the Venus firmware. Just so you know.

Also, if this gets worse, we need to channel it through proper Victron support channels (in that case you can DM me your email address @Kilowatt). This is not am approved support platform (but we do want to help people)

Edited August 1, 2019 by plonkster

[Youda]

48 minutes ago, plonkster said:

Unfortunately we can't check the per-cell voltages (well you can, but then you need the RS485 cable and software for that, or at least need to know how to extract that info via the serial console).

Actually, it's easy. I've put the diagnostic software and the guidelines for creating RS232 cable here:

https://powerforum.co.za/topic/2322-youdas-off-grid-lab/page/2/?tab=comments#comment-53641

 

48 minutes ago, plonkster said:

This is not am approved support platform (but we do want to help people)

Too late buddy!

[Kilowatt]

27 minutes ago, plonkster said:

My theory is that the high voltage alarm applies on the cell level, not on the overall voltage. This means there was a high cell. Unfortunately we can't check the per-cell voltages (well you can, but then you need the RS485 cable and software for that, or at least need to know how to extract that info via the serial console).

So why did we develop an imbalance? Maybe (at 52V) the balancers might not be as effective as they are higher up. Now I was told by the manufacturer (or at least someone that works there) that 51.8V is high enough to start balancing, and there are certainly many here who run with 52V without any problems. So this was a shot in the dark, a wild guess, and so the plan was simply to push the voltage up and see what happens (either the problem gets worse, or it goes away). What I did, because @Jaco de Jongh asked me, is patch the 52V limit up to 53V. The idea is to let it run like that for a few days (I assume you cycle the batteries?) and see if the problem (or others) returns.

Note that the 53V patch will disappear if you upgrade the Venus firmware. Just so you know.

Also, if this gets worse, we need to channel it through proper Victron support channels (in that case you can DM me your email address @Kilowatt). This is not am approved support platform (but we do want to help people)

Thank you plonkster, 

Unfortunately today is raining and overcast, not sure if we will get to full charge or not.

I do cycle my batteries, and also run them them on the optimized without battery life setting. There are also times when the battery has a large discharge current then immediately followed by a large charge current, so i am not sure if this confuses it or not. 

I won't make any changes to the firmware, so we can let it run as set and hopefully no more alarms. I will give you feedback, hopefully it will just be to confirm all is in order. 

thanks for all the help

[Kilowatt]

Another question, I normally have scheduled charge set for 5pm - 7pm while we cook, normally the batteries are at fully charged by this time of day so they stay at the maximum 95% i can select. This works well as i don't use electricity to charge the batteries and obviously they are not subjected to heavy battery usage while we cook. But what i noticed the other day when i selected keep batteries charged is that the battery went to 100% followed by absorption for the duration of the period i chose. Correct me if i am wrong but i don't think they can go into absorption when scheduled charging is selected, at least i have not noticed that. 

I was thinking that perhaps some of my issues i have experienced as mention above, could possibly be from the batteries not being allowed to go into absorption. Possibly not allowing the cells to balance often enough? 

And lastly how often should the batteries go into absorption if i cycle the daily from 100% - 25%? 

[___]

16 hours ago, Kilowatt said:

the maximum 95% i can select

If you turn off the "Stop on SOC" switch, it goes to 100%. That is literally what happens internally (the limit is set to 100%). That is also why the max you can select is 95%, otherwise there'd be two ways to get to 100% and customers would get all confused and ask too many questions

Additionally, charging never stops when the limit is 100%. There is also a good reason for this, and that reason is that your battery monitor (the BMV, or the BMS in an intelligent battery) might be out of calibration and it might reach 100% early. Many battery monitors needs to reset at the top, and most of them will reset when the voltage goes high enough for long enough (eg Pylontech resets around 53V).

Further, scheduled charging will always prevent discharge, but it tries to let any surplus PV through to the loads. I had a very long discussion with @The Terrible Triplett yesterday about the hardships that goes with this, because inverter efficiency is not a constant and it is close to impossible to balance the incoming PV with the loads in such a way that the battery component is exactly zero, you will always either slowly charge or slowly discharge the battery. There is some trickery involved to make it even out at the SOC limit you selected, but if you're battery is at 100% and you set the limit for the charge schedule to 95%, you should see a slow discharge (I can explain how the regulation works if anyone is interested).

16 hours ago, Kilowatt said:

Correct me if i am wrong but i don't think they can go into absorption when scheduled charging is selected

Scheduled charging is implemented on the CCGX/Venus-GX and simply tells the Multi to charge. The charge state is decided by the Multi and though you can manipulate it from the CCGX... why would you? So it is certainly possible to go into absorption at any point, if the Multi thought you were in Bulk and then you moved into a scheduled charge slot... the next state will be Absorption.

[___]

Oh, and one more thing. With LFP batteries there isn't really such a thing as absorption. That's a  lead acid thing. Lead acid batteries are like sponges that get really stubborn towards the end (the last 15%) and the only way to get them full all the way to the top is to soak them in electrons, in other words, to hold them at a high enough voltage (but not so high that you make too much gas) and wait for the tail current to drop.

With LFP, there is none of that. When the cell gets full, it will naturally spike up to above 3.55V. And then you know you are done. No point in having a prolonged absorption phase. Because of the requirements of the BMS -- some will only balance at a higher voltage, some only reset the SOC at higher voltage -- it is sometimes still recommended that you follow a two-voltage charge cycle, but just as often the battery will just have one voltage (the charge voltage, corresponding to 100% SoC) and you'd just hold it there permanently.

[Guest]

5 minutes ago, plonkster said:

Scheduled charging ...

After our chat my conclusion on Scheduled charging is that it is a good feature to have and also the reason why it was added i.e. to recharge batteries using utilities at specific times at night when power is sold very cheap.

But we don't have that in SA.

But because the feature is there, it has been "forced" to mimic other types of scenarios, scenarios which it was not intended for and therein some challenges are presented.

So unless it is used as designed, or the challenges presented when use not as designed, scheduled charging is not something one can really use.

Ps. I just think it must be said, calling a spade a shovel and all that.  

Unless I am missing something?

 

[___]

43 minutes ago, The Terrible Triplett said:

So unless it is used as designed, or the challenges presented when use not as designed, scheduled charging is not something one can really use.

To stop discharge at certain times of the day is an allowed (designed-in) abuse. It's even documented as part of the freakin' annoying questions in 4.4.4.

[Guest]

1 hour ago, plonkster said:

(designed-in) abuse.

Yes, I saw that. Unintentional consequence with a excellent end result.  

Maybe I must do that late afternoons to not discharge the batts for better after 12 evening use ... will see.