diy solar

diy solar

Large spiky current draws with no additional load... ah crap

jbatx

I make stuff with things
Joined
Jun 8, 2021
Messages
184
Location
Austin, TX
I checked in on my remote system just a minute ago and I was alarmed. One of the two 48v 16s overkill bms's reported some relatively huge current draw downs momentarily. ...and I dont know why. Hoping you all can plant some ideas.

One of the images is of the current chart for both batteries. The orange line is battery 1, which is the only one that spiked. 1 and 2 are parallel.

---System---

Two 16s 48v 100ah LiFePO4 banks with overkills bms's in parallel

Renogy 100a Rover scc

Xyz 5k inverter

5kw in pv from five 134v 9a strings.

----

The first draw spikes of about -40a were at around 6:30pm. The second of about -85a occurred at about 11:30pm.

Notice that battery 2 did not have a draw anywhere close. This was localized to battery 1.

The system is currently functioning.... According the the data flow. Voltage on both batteries is 53.22 currently.

Also, around 6:30, battery one had a large voltage fluctuation down to about 12v and up to about 75v.

For some context, all these data points are coming from the bms's over the serial port to esp32 devices that post to AWS.

Notice also that several of the cells in battery one are showing spikes up to 20v... just exactly 20v. I see this on occasion but never the many so close together.

Is there a short?

Is the bms failing?

Many screenshots attached.



Thanks in Advance.
 

Attachments

  • Screenshot_20211211-000056.jpg
    Screenshot_20211211-000056.jpg
    76.8 KB · Views: 18
  • Screenshot_20211210-235923.jpg
    Screenshot_20211210-235923.jpg
    126 KB · Views: 20
  • Screenshot_20211210-235741.jpg
    Screenshot_20211210-235741.jpg
    270.1 KB · Views: 21
  • Screenshot_20211210-234128.jpg
    Screenshot_20211210-234128.jpg
    77.4 KB · Views: 20
  • Screenshot_20211207-164218.jpg
    Screenshot_20211207-164218.jpg
    195.3 KB · Views: 20
  • Screenshot_20211210-235838.jpg
    Screenshot_20211210-235838.jpg
    76.6 KB · Views: 18
If I could just go look at the system I would. It's 90miles away. But, I suspect I'd see nothing unusual
 
Could quite possibly be erroneous. TTL is subject to noise and if the data gets mangled it will read weird numbers, particularly if the CRC isn't sophisticated.. Keep your serial leads as short as possible.
 
The battery and current fluctuations are not possible.
I agree that it could a bad connector or possibly the bms failing. I would up pack 1 and do so direct measurements before the bms.
 
Oooook... So, this morning finds that there are now +250a spikes. One at 3am and one at 7:30am. Battery one soc is showing 21% whereas battery 2 is at 82%
 
The battery and current fluctuations are not possible.
I agree that it could a bad connector or possibly the bms failing. I would up pack 1 and do so direct measurements before the bms.
I'm not clear on what you mean by "up pack".

I will check voltage before the bms, at the bank terminals and measure the cells and check the connections. It would be baffling if there was a loose connection anywhere.
 
Could quite possibly be erroneous. TTL is subject to noise and if the data gets mangled it will read weird numbers, particularly if the CRC isn't sophisticated.. Keep your serial leads as short as possible.
I'm going to install a simple shunt to validate the flow.
 
SCC is not involved as this is occurring during Non-Solar hours 18:30 & 22:30 hours.
Xyz Inverter - unknown, cannot offer thoughts as this is "Value Gear" but with that having been said there is One point of Suspicion because it may be in process of failing and not well. This requires a deep examination & testing. Please provide full product Name, Model etc and Link to Product Info (Manual preferably).

JBD BMS' are pretty reliable and fairly well known. They are FET Based and therefore subject to the limitations of the tech. For one, if using a 100A BMS do NOT pull 100A from it as that IS the edge of performace specs, always allow a 20% over margin, you would be farther ahead with a 150A Model IF you know you will pull 100A. In Parallel the battery packs with share & divide Load/Charge but you should not count on that, all it takes is one pack to cut off during Hi Load & the other one gets a momentary HIT that can and would damage its BMS.

Battery Packs installed in Parallel need to have the ability to work as StandAlone or "Last Man Standing" in the event that any other pack may cutoff for any reason whatsoever. This rule then plays out that if the Entire System can pull up to 250A then each pack should be able to deliver that, lest there be unpleasant consequences.

IF the cells, connections (cell busbars) etc were having faults you would see off numbers at cells being reported by the BMS'. A Shorting BMS' lead for example would present very erratic cell readings from that lead.

You have not mentioned TEMPERATURES, are the battery packs & cells within normal temps ? Have you observed ANY temperature issues or inconsistencies ? IF 2 BMS temp Sensors are in place and one varies by more than 1 degree from the other, it is an indicator of potential issues.

Get BATTERY 1 OFF the bank till you sort it. Those Voltage Spike are VERY BAD and far outside acceptable limits... It does appear the BMS is BAD as it allows such to occur. There is a Good Possibility that the other BMS has been damaged as well.

LASTLY on the Inverter.... They CAN fail as well and sometimes not well at all. If for some reason it is creating that demand spike (due to a short or so) it will NOT end well. You must determine if IT is OK and that needs tests.

PLEASE I urge Caution & Care ! Do have a Dry Chem extinguisher on hand.
Hope it helps, Good Luck. * Will Watch this thread *
Steve
 
I'm not clear on what you mean by "up pack".

I will check voltage before the bms, at the bank terminals and measure the cells and check the connections. It would be baffling if there was a loose connection anywhere.
LOL apologize for that. You cannot have two packs in parallel and one is putting out that much voltage and current while the other is perfectly normal. It's either that the data cable is bad or the BMS is sending out bad info.
I was trying to say that you should separate the two batteries and just test the one that says it's putting out 70V separately using a Volt meter and Amp meter. I suspect the readings will be normal even though the software may be saying differently.
 
If I could just go look at the system I would. It's 90miles away. But, I suspect I'd see nothing unusual
Crap I did not realize that. I would make the drive, it's not a good idea to have something like this just left running with a possible issue.
 
Answering questions from above:

Yes, temp spikes that make no real sense have shown up. That on going. Started a couple months ago or more. I contacted overkill about this a weeke ago. No response yet. I attached a screenshot of the recent spikes of temp.

The 5k inverter was installed about 3 weeks ago and it replaced a 3k. 5k on Amazon

XYZ INVT 5000 Watt Pure Sine Wave Inverter 48V DC to AC 110V 120V, 5000W Power Invert Surge 10000W Power Converter for Solar System. (5000W 48V) https://www.amazon.com/dp/B089F8G98...abc_FCK5HVG4XATKDMT162K0?_encoding=UTF8&psc=1

The system is housed in shipping container, so it won't burn down a house or anything.

There is nothing on the system that would -90a unless it was shorted. The most I've ever seen is -30a draw when the ac, cooktop and well pump are all on. Further, that -90a reading was only on battery 1 while 2 stayed at about 1a.

as for the +250a spike, there is no input from the scc that would cause that. Again, that reading was only on battery 1 while battery 2 was somewhere around +2-5.
 
well... shits going crazy on battery 1. I've been watching the data flow and it's throwing almost all the various error conditions and protections.

I'm heading there now to test and maybe swap out the bms. ...I have a new one for a different battery. Here's a screenshot of the latest protections that have toggled on/off since this all started yesterday.

Note: the graph is the sum of the bool values for the protection parameters by minute.

also, I didn't begin capturing the cfet and dfet booleans until just a few minutes ago. Had I been, you'd see them toggling like crazy.
 

Attachments

  • batter_1_protection.png.png
    batter_1_protection.png.png
    339.9 KB · Views: 7
All the signs point to a faulty BMS. You may have to make the trip and swop the two BMS’s around and observe.
 
My opinion: some bug in the software or a hardware fault on the ESP or BMS. What software are you running on the ESP32 - buffer overflow maybe? Race condition? Parse error?

I think you may be onto something...

The esp code is covered in this thread: https://diysolarforum.com/threads/anyone-working-with-the-overkill-solar-arduino-lib.27811/page-3

...that particular post has the JBD protocol spec sheet in it.

Anyhow, I have the esp device pulled off, the BT module connected and I'm watching the log in the app (android version). So far, nothing is a miss.

I guess It's possible that the esp32 device was causing some issues with the BMS... cold weather contact shrinkage and a short maybe ...I'm using breadboards which isn't great. I'm going to sit here a while and see what goes one with the app.
 

Attachments

  • IMG_20211211_165729.jpg
    IMG_20211211_165729.jpg
    138.2 KB · Views: 16
Last edited:
Those breadboards are not good for long-term use as they develop bad connections on the chips due to not being soldered or gold plated but are good for prototyping and proving it works. If you unplug the chips and just put them back in your problems may go away.
 
Back
Top