Hi, 

Has anyone run into a problem where a 4Kw InfiniSolar & PylonTech battery doesn't work well together? The battery bank is 150A and there's 3x 4Kw inverters. The battery is too small, yes, but the client is gradually growing the system and focusing on daytime use. He'll buy more batteries next month or so. 

When running on batteries, the inverter "drains" the battery, yet the SOC LED lights remain on 100%. It takes about 3 minutes to run the batteries flat with a load of about 2Kw. The 7200 Kwh battery bank should last about 3 hours with 2Kw load. According to the inverter, the batteries run down to 48V and then the system shuts down. Took the batteries back for RMA, but supplier says the batteries are fine. In fact, they never been drained below 98%! So it seems the inverter doesn't read the SOC properly. There is no comms between the inverter and battery, like on the Victrons. 

Hi SilverNodashi

What software are you running on the inverter side, or are you referring to another form of SOC coming from the inverter display etc?

The inverters are known for not having accurate SOC, from there most users running BMV's on the older generation batteries.

For accurate results, one need to use the SOC from the Pylon rather than the SOC as reported from the inverter side.

The Pylons have a pretty nifty BMS with loads of info.

Edited June 1, 20186 yr by Riaanh

@SilverNodashi

The Infini has no way of integrating the SoC or any other "data" from the Pylon.

My first pass advise would be, in order to "integrate" the two, you need to use the ICC software. AFAIK, the latest version now "talks" to the Pylon's, and that would make a nifty combo.

I have an infini with 4 pylon tech batteries. There is no connection from the inverter to the batteries apart from the power cables. 

The pylontech bms does all the work and the infini does the charging at at 50 amps irrespective of how empty the battery is. 

The readings on the infini is 52v for the battery and it floats at 1 amp. 

The system was set up by Kurt from Full Circle Solar  

I did a trial once to see by how much the batteries would deplete and it stopped at one green light on the battery outer. 

It seems as if the 3x4kw system is set up as a three phase system with one battery per inverter. 

My feeling is that looking at it from a single phase perspective would be wrong. The best would be to get advice from experts in that field. This would turn it into a different ball game. 

Edited June 2, 20186 yr by Johandup

To the best of my knowledge I only know of ICC that communicate with the Pylon batteries and report actual Voltage, Amps, SOC, Battery Temp, Number of Cycles and the likes.

Edited June 2, 20186 yr by Riaanh
Pylon added

Had similar thing with an axpert, the voltage measurements are as we know not all that good but the pylons cable and multiple connections doesn't help either. You have to compensate for inaccuracy and non linearity of the ADC. i suppose the master inverter in your case, very important that you do this for both the low and high soc V. other issue is as said the cables which you are sharing between 3 inverters, the voltage drop under load is probably too much, maybe get another cable pack take both sets to a thicker one

i could be wrong but I think guys using the BMV and ICC combo don't have any issue because inverter battery volt is overridden by the bmv Battery voltage? perhaps, anyone confirm? ICC guys?

On 2018/06/01 at 2:20 PM, Riaanh said:

Hi SilverNodashi

What software are you running on the inverter side, or are you referring to another form of SOC coming from the inverter display etc?

The inverters are known for not having accurate SOC, from there most users running BMV's on the older generation batteries.

For accurate results, one need to use the SOC from the Pylon rather than the SOC as reported from the inverter side.

The Pylons have a pretty nifty BMS with loads of info.

Any inverters should ready the voltage of any battery. Other similar installations don't show this behaviour, hence asking whether someone else have seen this before. 

 

I don't see why we are forced to buy extra 3rd party components to do a job which the base components should do already. You tell a client "Sir, we need this monitoring system to make sure the inverter reads the voltage properly" - you know what reaction you get? "Take it back to the supplier if it's faulty". Problem is, the Pylontech were RMA'd and tested fine, and so was the Infinisolar inverters. Client refuse to spend more money on a 3rd party tool which wasn't designed by either the battery or inverter manufacturer to make it work. 

On 2018/06/01 at 2:20 PM, Riaanh said:

Hi SilverNodashi

What software are you running on the inverter side, or are you referring to another form of SOC coming from the inverter display etc?

The inverters are known for not having accurate SOC, from there most users running BMV's on the older generation batteries.

For accurate results, one need to use the SOC from the Pylon rather than the SOC as reported from the inverter side.

The Pylons have a pretty nifty BMS with loads of info.

The inverter doesn't run software. It has it's own built-in firmware which manages it. The problem is, the inverter cannot communicate directly to the battery's CAN bus. It should be able to read the voltage on the cables. 

On 2018/06/01 at 9:11 PM, Johandup said:

It seems as if the 3x4kw system is set up as a three phase system with one battery per inverter. 

My feeling is that looking at it from a single phase perspective would be wrong. The best would be to get advice from experts in that field. This would turn it into a different ball game. 

No, I never said it is setup like that. All 3 Pylontech batteries are in the same bank. The 3 inverters are paralleled on a single phase grid. 

Who are the "experts" you referring to then? If neither the inverter of battery supplier has an answer to the question, who else would? The manufacturers haven't shed any light on the matter either.

11 hours ago, Weasel said:

Had similar thing with an axpert, the voltage measurements are as we know not all that good but the pylons cable and multiple connections doesn't help either. You have to compensate for inaccuracy and non linearity of the ADC. i suppose the master inverter in your case, very important that you do this for both the low and high soc V. other issue is as said the cables which you are sharing between 3 inverters, the voltage drop under load is probably too much, maybe get another cable pack take both sets to a thicker one

i could be wrong but I think guys using the BMV and ICC combo don't have any issue because inverter battery volt is overridden by the bmv Battery voltage? perhaps, anyone confirm? ICC guys?

The 3x 50A Pylontech's are in a single battery bank, forming 150A/53V (7200 Kwh) battery bank. From there, 3x 25mm cables are installed, one set per inverter.  On the longest run, of about 2.5m (probably a bit less), mathematically the voltage drop is 0.09V. Not a heck of a lot to worry about. 35mm wire would give a voltage drop of 0.06V. This is well within recommended specs. 

10 hours ago, SilverNodashi said:

Any inverters should ready the voltage of any battery. Other similar installations don't show this behaviour, hence asking whether someone else have seen this before. 

 

I don't see why we are forced to buy extra 3rd party components to do a job which the base components should do already. You tell a client "Sir, we need this monitoring system to make sure the inverter reads the voltage properly" - you know what reaction you get? "Take it back to the supplier if it's faulty". Problem is, the Pylontech were RMA'd and tested fine, and so was the Infinisolar inverters. Client refuse to spend more money on a 3rd party tool which wasn't designed by either the battery or inverter manufacturer to make it work. 

Whilst I had LA batteries I was able to make a workable "solution" to the Axpert's SOC using the data from the Axpert itself as collected by ICC so inverters' inability to get a accurate SOC is just laziness on the manufacturers part. Client in your case is missing the bus as few inverters have that capability. Victron does but this capability falls away once you add a MPPT without CGGX. I think with @plonkster on board Victron's intergration (which is pretty good) will improve.

On 2018/06/01 at 9:18 AM, SilverNodashi said:

Hi, 

Has anyone run into a problem where a 4Kw InfiniSolar & PylonTech battery doesn't work well together? The battery bank is 150A and there's 3x 4Kw inverters. The battery is too small, yes, but the client is gradually growing the system and focusing on daytime use. He'll buy more batteries next month or so. 

When running on batteries, the inverter "drains" the battery, yet the SOC LED lights remain on 100%. It takes about 3 minutes to run the batteries flat with a load of about 2Kw. The 7200 Kwh battery bank should last about 3 hours with 2Kw load. According to the inverter, the batteries run down to 48V and then the system shuts down. Took the batteries back for RMA, but supplier says the batteries are fine. In fact, they never been drained below 98%! So it seems the inverter doesn't read the SOC properly. There is no comms between the inverter and battery, like on the Victrons. 

Could this not be due to voltage sag?

13 hours ago, Chris Hobson said:

Could this not be due to voltage sag?

...or a cabling issue - measure the voltage at the battery terminals and at the inverter under constant load (like a kettle).

2 hours ago, pilotfish said:

...or a cabling issue - measure the voltage at the battery terminals and at the inverter under constant load (like a kettle).

I've seen this once. With a large number of racks, the cabling that comes with the batteries become a little too thin and the voltage drop increases. But it was less than a volt at full power in that particular instance. The trouble was that for the blue inverters, when running ESS, we use something called dynamic cut-off that makes the low voltage cut-out a function of the current draw. An extra 0.6V at full power was enough to drop it under the cut-out and the inverter would go back to grid cyclically. The fix in this case was a) install voltage sense cabling (if the hardware has it), or b) drop the dynamic cut-off curve by 0.6V or so at the top end. The customer applied the latter advice and it solved the problem perfectly.

So one thing to do would be to place your DMM across the cable and measure the voltage drop, multiply by two to get the total... or just measure either side across the cables. Adjust your low-voltage downwards by that much and see if it improves.

In my experience, I can easily drop a 12.8V Lithium battery to under 12.2V if I discharge it at 1C. The problem might very well be that 48V is just too high.

Edited June 4, 20186 yr by plonkster

So, revisiting this issue.

At another installation, with 3x new 5KW Axpert inverters, each with 2x Pylontech batteries, the inverters run quite long (depending on the load) and drained the batteries to 80% SOC until theuy shut down

Yet, the 3x Infinisolar Inverters run for about 7  minutes and then shut down. The screens were completely off. Yet, on the batteries, all the LED's are still lit. The inverters shut off at 48.3V / 29% SOC.

Could it really be a case of the ADC not reading the Volt, or Ampere of the batteries properly? Would a firmware update fix this, if one is available?

I've just read the whole thread and I have to agree with what most of the guys are pointing to.

My 2 cents:

1) 35mm2 cable for 3x InfiniSolar 4K is way too small. Use 50mm2 for all the connections.

2) Also, check the voltage drop on DC fuses, while under heavy load.

3) 2.5m cable between (small) battery pack and array of 3 inverters is too long. Build a shortest possible battery connection and keep in mind that all the inverters must have the same length of cable between the battery and inverter itself. This is explicitly noted in the manual for parallel operation.

4) Lower the cut-off voltage setting in the Infini settings, don't forget to SYNC.

5) Measure the battery voltage and compare that with what inverters are reporting. If there's a difference in readings, you can calibrate the InfiniSolar's ADC in the same way like Axpert.

The link bellow is for 3K version, which is different from 4K. For 4K ask your dealer for "battery calibration SOP".

http://www.ostrovni-elektrarny.cz/support/Voltronic/AccessPort_and_SOP_battery_calibration.rar

Edited June 21, 20186 yr by Youda

1 hour ago, Youda said:

1) 35mm2 cable for 3x InfiniSolar 4K is way too small. Use 50mm2 for all the connections.

35 is fine if each inverter has it's own pair to the busbars. 4kw at 48V is less than 100A. You can get away with 25mm^2 in a pinch. Of course if any part of the current path is shared, then you got to beef that bit up.

14 hours ago, Youda said:

I've just read the whole thread and I have to agree with what most of the guys are pointing to.

My 2 cents:

1) 35mm2 cable for 3x InfiniSolar 4K is way too small. Use 50mm2 for all the connections.

2) Also, check the voltage drop on DC fuses, while under heavy load.

3) 2.5m cable between (small) battery pack and array of 3 inverters is too long. Build a shortest possible battery connection and keep in mind that all the inverters must have the same length of cable between the battery and inverter itself. This is explicitly noted in the manual for parallel operation.

4) Lower the cut-off voltage setting in the Infini settings, don't forget to SYNC.

5) Measure the battery voltage and compare that with what inverters are reporting. If there's a difference in readings, you can calibrate the InfiniSolar's ADC in the same way like Axpert.

The link bellow is for 3K version, which is different from 4K. For 4K ask your dealer for "battery calibration SOP".

http://www.ostrovni-elektrarny.cz/support/Voltronic/AccessPort_and_SOP_battery_calibration.rar

1. There's a separate set of cables to each inverter, connected to a very thick copper bas bar. 

2. The voltage drop was about 0.06v from bus bar to inverters

3. Unfortunately there isn't room to build shorter cables, or keep them the same length. This is one of those cases where we have to make work with what we can. 

4. Although I used the recommended settings from the manufacturer, this is definitely something worth trying, when I go there again. 

5. The voltages were out by a couple millivolt, can't remember what it was. 

 

As a side-note, at another installation, there's 2x Pylontechs to a 5Kw Axpert with about 3m 25mm cable, and the inverter ran for 1hour 40 minutes at 2.1Kw load. This is much much longer than the 7 minutes on this particular installation. 

Hi SilverNodashi,

did you had a chance to check the site? I'd really like to educate myself on what was the root cause and the solution.

Thanks beforehand,

Youda

What are your cut off voltages, when grid is available and when grid is unavailable? 

Without dynamic cut off voltage, like the Victrons offer, the Pylon battery can be thwarted by high loads. 

I'd be interested to see whether the voltage calibration program rectifies the issue, if there is an existing voltage discrepancy. 

While the warranty terms dictate setting the low voltage disconnect to 48V, the internal BMS will always protect the battery from excessive charge or discharge. 

Unlike lead acid batteries, you could set the low voltage disconnect as low as you like, and the Pylontech BMS will simply disconnect the battery on low SOC. 

The only caveat to forcing the battery BMS to disconnect itself from the load and switch off is that you will need to manually switch the batteries back on for the next intended cycle.