Batteries keep dying.

[Ibid49]

I just posted a comment on Will's latest YouTube poll on Battleborn Batteries, and he invited me to post it here to see if anyone can help diagnose this issue.

Short version is my Battleborn batteries (LiFePO4) keep dying after about a year of use. (It's happened twice), even though I think my system is well over-engineered for my usage. The long version can get very detailed, so I'm not sure how much to put here.

My system:

• 4 12v 100Ah Battleborn Batteries in series and then parallel for a 24v 200Ah bank.
• Victron Multiplus 3000w inverter/charger
• Victron 150/85 solar charge controller
• 6 360w PV panels combined in a combiner box for 3 strings of 2
• Bare minimum of 2 awg cabling (this has been changed recently)
• Full video walkthrough of my current system here -

Typical usage for the whole house during the day is somewhere between 200-500 watts. Though my son and I are both gamers and if we're both running our gaming rigs, usage will rise up to about 700-800 watts. We also have a microwave, and for the few minutes that's running a day, it might pull 1500w total for the house.

My current suspicions lie with the Victron inverter, although we just sent it in to someone who's supposed to be the best diagnostician for those devices. And he supposedly ran it through the wringer and gave it a clean bill of health as long as you ignore the weirdness he experienced that then went away.

My reasons for suspecting the inverter charger are that it just keeps acting weird:

• It is the only part of the system that was present during both battery failures. All cabling and all 4 batteries were replaced between the first failure and the second.
• For about 8 months we were using it exclusively to charge the batteries from a gas generator, as we didn't have our PV panels in and working yet, and we still use that method to charge the batteries when we have several cloudy days in a row. So it's had plenty of opportunity.
• It trips the generator's overload protection (a 8750 watt generator) if we are pulling anything over 100 watts when we connect the Multiplus to the generator. We have to shut down everything in the house before we try to charge the batteries off the Multiplus. After the initial startup, we can turn everything back on and it's fine.
• If instead, we switch it from "Inverter Only" mode to "Charger Only" mode after the generator is running and connected, it does not trip the generator overload. It does however lose power to the house for up to 6 or 7 seconds while it switches over.
• It also sometimes trips the generator's overload protection when the batteries reach fully charged voltage.
• Sometimes when batteries reach full charge, the inverter shuts off electricity to the house for a second or two. This might oscillate on and off two or three times before it stays on.
• When these battery issues first started happening for the second time, I noticed the inverter was shorting out the batteries, and I had continuity tone between the positive and negative battery terminals on the inverter, even when nothing else was connected to the Multiplus. That issue resolved itself by disconnecting the batteries for a while and then reconnecting them.
• On another recent instance, I had just individually charged my batteries to full using a separate 12v charger and was hooking them back up to the inverter. Once I had everything connected and tightened and turned the battery switch back on, I measured voltage across the battery bank (of just two of the batteries) and got 4 volts. Finding that exceedingly curious with two batteries that I had just charged to full, I measured them individually, and one of them (connected to negative) was reading just fine at 14.2V, and the other one was reading -10V. That's negative 10 volts. I know what you're thinking, that I had my probes switched. I didn't. I checked it multiple times. (And also, 14 - 10 is 4, exactly what they read together) And when I disconnected the battery switch, they both went back to 14.2V. Somehow, that inverter was pulling voltage from both batteries in opposite directions with nothing but a jumper cable between them.
• When we sent it in to a Victron Service Center and they hooked it up to their batteries, the inverter read low battery alarm even with full batteries. They "fixed" it by doing a factory reset on the inverter and flashing the firmware/software, but they couldn't explain how it happened.

Anyway, I'm sure more detail is needed, but perhaps it would be faster to ask specific questions and I can answer them.

TIA, I know it's a long read.

 

[profesor79]

do you have any monitoring solution that is capable of recording spikes in voltage/current?
in my case my inverter was overloaded and by analizing current curve I found out that my old fridge is dying and givinga spike of 20A.

 

[Q-Dog]

You don't have enough battery capacity.

Also, are your batteries shutting down due to the cold?

 

[TacomaJoe]

When you measured -10V on the battery it was because that battery was shut down. The two batteries and the load form a series circuit. Replace the battery with a resistor about double the resistance of the inverter and you will get those same measurements. A BMS has some resistance when shut down giving you your results.

 

[Mattb4]

I would say you have an issue with a faulty BMS that led to issues with your Victron. But it is just a WAG. It could be elusive to detect if the fault is intermittent.

 

[Ibid49]

do you have any monitoring solution that is capable of recording spikes in voltage/current?
in my case my inverter was overloaded and by analizing current curve I found out that my old fridge is dying and givinga spike of 20A.

I don't think so. I don't even know what would look like. Best thing I can think of would be a full-fledged Oscilloscope , but I don't have one and I think that would have to be inline.

 

[Ibid49]

You don't have enough battery capacity.

Also, are your batteries shutting down due to the cold?

I realize I have smaller than would be typically recommended. That's by design however. I bought these batteries two years ago. Batteries were way more expensive per year of life than PV panels, and so I bought enough PV panels that even on cloudy days, we can still almost but not quite match our usage. Figured we could limp along or run the generator if need be. Plus I knew that battery prices would be coming down, so wanted to wait until they did (Still think they will come down further)

Is there a reason why my battery capacity is "wrong" vs just "not what I'd do"? I'm not exceeding any of the specs of the batteries that I'm aware of, but I also recognize I don't fully understand all the chemistry or electric circuitry stuff. So I could be missing something.

And no, that battery box is insulated with 2 inches of foam board all the way around. When I built it, I actually bought heating elements and a thermostat thinking I would need to provide heat in there. But it turns out the electronics themselves provide all the heat I need. Even in -10F weather, that box doesn't get below 50F.

 

[upnorthandpersonal]

You've got about 4.8kWh storage total. Your loads are at a minimum (if I take your 200W as a continuous base load) of 4.8kWh over a 24 hour period. This means that, in order for your batteries to get fully charged you need to generate twice this amount per day (charge the battery + load) - so about 10kWh.

You have 6 360W panels. Let's assume these are optimally placed, deliver their nameplate power (which they won't) and you have 5 hours of direct sunlight - this means a 10kWh generation. In other words, your system is undersized. If you can charge for example the battery to full after 5 hours, you then have to pull a load of 200W minimum for the remaining 19 hours - that's 3.8kWh (or essentially 80% of your capacity) - and that's without any of the other loads you've mentioned.

 

[Ibid49]

You've got about 4.8kWh storage total. Your loads are at a minimum (if I take your 200W as a continuous base load) of 4.8kWh over a 24 hour period. This means that, in order for your batteries to get fully charged you need to generate twice this amount per day (charge the battery + load) - so about 10kWh.

You have 6 360W panels. Let's assume these are optimally placed, deliver their nameplate power (which they won't) and you have 5 hours of direct sunlight - this means a 10kWh generation. In other words, your system is undersized. If you can charge for example the battery to full after 5 hours, you then have to pull a load of 200W minimum for the remaining 19 hours - that's 3.8kWh (or essentially 80% of your capacity) - and that's without any of the other loads you've mentioned.

Yeah, I'm actually rather proficient with a calculator myself. And having lived full time with the system for about 2 years, I'm probably more versed than you are in its actual capabilities and shortcomings. Especially since you're operating on numbers I've given you that are generalized and averaged, but don't take into account things like intentional human behavior, such as keeping our high-draw activities focused during high-sun hours when we're producing 1500-1600 watts. When it's a cloudy day, we can bring our total house draw down to 20-30 watts in order to conserve batteries, and we frequently do so. And yes, that still means that we are frequently using the generator to charge the batteries, as I said from the beginning, sometimes effectively charging the batteries twice per day.

But even if we charged/discharged the batteries twice every day (we're not), that should still only shorten the battery lifespan by half of it's life expectancy, not by 90-95% which is what we're seeing. Nor does it matter if we're using 80% of our capacity, since LiFePo4 batteries are purported to be able to draw their full capacity. Even if we were taking them down to zero (again, we're not), as long as we're taking them back up to full and letting them balance, that should not be enough to kill brand new batteries in a year's time. So I don't see how your explanation holds any water.

Honestly, this brings up another issue. I've seen this before in other solar forums. It seems like the types of people who frequent these forums tend to be more sanctimonious and judgmental than vegans. The common thread seems to be "Don't question the group think." Instead of actually trying to solve the problem you just want to preach your own personal opinions and preferences, rather than actually understanding the underlying physics and trying to help.

I came to this community from an invitation and offer to help, but the only thing I've learned here is that your opinions and expertise aren't worth more than "Here's what worked for me, and I don't understand anything deeper than that." You really should try to focus on helping people with where they are, rather than just trying to tell them what to do with a condescending attitude. Anyone who's only solution to a problem is "throw more money at it" isn't an expert and shouldn't portray themselves as such.

 

[Ibid49]

If anyone else wants to help, the only thing I can figure is that I've seen simple AC-DC conversions that use four diodes to create a full-bridge rectifier. This produces a DC current, but it still pulses with the AC source. In fact, if I understand it correctly, I think it would pulse twice for each AC cycle, so closer to 120Hz. If the Multiplus is using that methodology (I have no idea if it is) and if it's reporting pushing 1500 watts into the batteries, that could mean that the peaks would be at around 3000 watts, which is just over the 1C rating of the batteries. Though, I would think that any internal wiring in the batteries would only heat up (or fail) based on the average amperage, not the peak. I don't have any idea what that would do to the BMS, however.

Either way, though, that seems like it would be an engineering flaw in the charger, and I bought the Multiplus because Victron seemed like a good brand, and it seems fairly well engineered from the outside.

 

[krby]

Honestly, this brings up another issue. I've seen this before in other solar forums. It seems like the types of people who frequent these forums tend to be more sanctimonious and judgmental than vegans...

This is a common problem on a lot of "niche" online forums where new folks turn up asking questions. Sometimes it is folks being preachy and wanting to be right. But, IMO it's also because it is often difficult to tell "new person without much background looking for basic info" from "new person with some experience in the niche asking for specific help". Most forums focused on a single topic (like this one) have far more of the former than the latter.

The calculator advice you got is often appropriate for someone being totally new to solar and electricity and batteries who hasn't thought through the issues around generation, capacity, load sizing etc. That's a really common situation here.

Now to actually try and help:
Regarding the AC->DC conversion and the pulse peaks. Yes, you'll get 2x the AC input frequency from just a full-bridge rectifier (with peak voltage being about 0.7V below the AC input peak) but that is not the only thing the Multiplus has in the AC->DC conversion path. At a minimum, it will also have some capacitors to give a much smoother DC output. Also, it's not really the total watts into the battery that matter, it's the voltage and the current (each with separate limits). Ya, I know: V x A = Watts. But the battery has voltage limits and current limits, not a watt limit. The current limit is really related to heat buildup in the cells, which is what really kills things.

I don't have a good answer for you. If it were me, I would talk to Battleborn and see what their folks suggest. It's probably going to involve monitoring the power in and out of the batteries, which you could do with one of several battery capacity tracking thing (Victron makes an expensive, but good one) or you can get a "display screen" for the Multiplus on your laptop or phone if you buy the correct USB->??? interface from Victron (the same thing you'd use to program the Multiplus). But, what I would want is automatic logging so I could look at every moment of several days, and I don't think the Victron apps will do that.

Given your description, it sure feels like the Multiplus isn't acting like it should. Your comment about the inverter mode tripping the generator while charging only mode doesn't has me wondering whether you have an AC load that is pulling a ton of current on startup. Maybe something that when it is plugged in pulls a ton of current for very little time? It's a guess, though.

 

[Mattb4]

Post deleted.

 

[Ibid49]

Thanks @krby.

but that is not the only thing the Multiplus has in the AC->DC conversion path. At a minimum, it will also have some capacitors to give a much smoother DC output

Yes, I thought of that after I posted. In the vein of "that's what I would do if I were designing this device, so Victron probably has done something at least that smart." The next thought I had was, I wonder if this particular unit has a bad soldering job or some other weak connection on one of their big boy capacitors? Maybe not a full break, but an intermittent fiddly connection or nicked wire. That could possibly avoid visual inspection if you weren't looking too hard. Could that possibly cause this kind of issue? In fact, couldn't it also possibly explain all the other issues as well?

Also, it's not really the total watts into the battery that matter, it's the voltage and the current (each with separate limits). Ya, I know: V x A = Watts. But the battery has voltage limits and current limits, not a watt limit. The current limit is really related to heat buildup in the cells, which is what really kills things.

That makes sense.

If it were me, I would talk to Battleborn and see what their folks suggest.

I have been. In fact, their customer service has been incredible, which is why I posed on Will's Battleborn Poll in the first place. I bought four of these batteries, they have sent me 10 total so far, warrantying out 4 of them the first time this happened, and another 2 the second time. They have been very friendly and go above and beyond to help, but they've been focusing almost entirely on the cabling, which, by my calculations, at amperage I'm using can only account for a difference between 99.1% voltage delivery vs 99.5%. And the cables are never warm to the touch, so I can't see how they could be losing appreciable amounts of electricity as heat. They've been loathe to look at the Multiplus as the culprit, and my only guess as to why is maybe because Victron isn't as good to work with, so maybe they're chasing the easy answers instead of the likely ones.

Maybe something that when it is plugged in pulls a ton of current for very little time? It's a guess, though.

We have a regular full-sized fridge (2 years old) that would certainly qualify, but A) it's happening every time, even when we can hear that the compressor isn't running, and B) I would only expect the true spike to occur right when the compressor kicks on, which would have to coincide exactly with when we were trying to switch the charger on. In any case, the fridge is the first culprit we thought of too, but even when we turn it off (via the breaker box), and have only lights and computers pulling(which have their own AC-DC conversion and filtering circuitry), we still trip the generator. The only thing we've found that works is turning enough stuff off so we can get below that "magic" 100w number.

 

[Ibid49]

Funny but I did not do any of that. I proposed a possibility that you did not either comment on or did not read. Instead of thinking it is your Victron based on the symptoms you gave. I suggested it is a BMS problem in your BB batteries.

Try to avoid group condemning when answering to individuals.

I appreciate that you didn't do that. But my language used words like "tend to" and pointing out a pattern I've seen, not painting everyone with the same brush. Generalizations necessarily imply exceptions. I didn't respond to your particular suggestion because we're talking about a failure of four different batteries at approximately 12-15 months after initial installation. So a problem with four different batteries' BMS seems probabilistically unlikely. Unless you're suggesting that it's an endemic design flaw in Battleborn's BMS?

 

[krby]

Re Battleborn, that's good to hear.

Tripping the generator is weird when pulling only 100W of load. How have you verified that the load is 100W? If I were in your shoes, I'd try to find some way of capturing and recording either:

a) Zero in on that "trips generator when over 100W" issue and capture the current between generator and multiplus and/or multiplus AC output and all loads.

b) Focus on the DC power to/from the Multiplus and get moment to moment (maybe every few seconds?) current and voltage between batteries and the Multiplus while demonstrating the problem (or just a typical 24hr day, starting from batteries in a fully charged state, charged by something other than the Multiplus)

I'm grasping at straws here, but since you've replaced the batteries a few times, it sure feels like you're "bleeding" power somewhere and if I were in your shoes I'd probably end up trying to monitor all the power ins and outs of the Multiplus (AC in, AC out, DC in/out) and hope that leads to some conclusion. The ideal case is you could hook up enough gear to record frequent values over a long period, then import those into something and graph them, looking for problems. Off the top of my head, I don't know what gear will do this for sampling and recording for you.

One other thing: You said Victron reflashed firmware and reset everything. Did you re-program it with the settings you want? The Multiplus line has a TON of settings, things for max charge current, bulk/absorb/float voltages, etc. What are those things all set to?

UPDATE: I *just* watched your video. I should have done that earlier. You have the smart shunt thing...I wonder if anyone knows how to get data out of that and record it? Maybe try searching for that on these forums? What is is showing for Ah in and out of the batteries when you're seeing them dead? TBH, since your having trouble, I would not trust the shunt reading until you confirm the voltage and capacity it shows with a multimeter across the whole 2S2P battery pack (maybe at the bus bars, or where the pack connected to the Multiplus)

 

[upnorthandpersonal]

@Ibid49

I was trying to establish a base line based on the power numbers you stated and no other details: trying to eliminate certain aspects. Maybe you think that's not helping solve your problem, but eliminating things is part of the process.

Honestly, this brings up another issue. I've seen this before in other solar forums. It seems like the types of people who frequent these forums tend to be more sanctimonious and judgmental than vegans. The common thread seems to be "Don't question the group think." Instead of actually trying to solve the problem you just want to preach your own personal opinions and preferences, rather than actually understanding the underlying physics and trying to help.

Thanks for that one...

 

[upnorthandpersonal]

When these battery issues first started happening for the second time, I noticed the inverter was shorting out the batteries, and I had continuity tone between the positive and negative battery terminals on the inverter, even when nothing else was connected to the Multiplus. That issue resolved itself by disconnecting the batteries for a while and then reconnecting them.

No fuse between the inverter and the batteries? Shorting out as in continuous or just to charge the capacitors on the input? Do you have a pre-charge circuit?

What are your Multiplus settings?

On another recent instance, I had just individually charged my batteries to full using a separate 12v charger and was hooking them back up to the inverter. Once I had everything connected and tightened and turned the battery switch back on, I measured voltage across the battery bank (of just two of the batteries) and got 4 volts. Finding that exceedingly curious with two batteries that I had just charged to full, I measured them individually, and one of them (connected to negative) was reading just fine at 14.2V, and the other one was reading -10V. That's negative 10 volts. I know what you're thinking, that I had my probes switched. I didn't. I checked it multiple times. (And also, 14 - 10 is 4, exactly what they read together) And when I disconnected the battery switch, they both went back to 14.2V. Somehow, that inverter was pulling voltage from both batteries in opposite directions with nothing but a jumper cable between them.

Talking about physics, this is impossible. Even with a dead short, it would pull both batteries down - or the BMS protection would kick in, no current would flow, nothing to pull down.

 

[Culchee]

“More sanctimonious than a vegan”.... I don’t think that is possible..

 

[krby]

When you measured -10V on the battery it was because that battery was shut down.

Is that what the Battleborn BMS does when it disconnects? I thought it showed just a low voltage, like 4V?

 

[upnorthandpersonal]

Is that what the Battleborn BMS does when it disconnects? I thought it showed just a low voltage, like 4V?

All About Low-Voltage Disconnect

Low-Voltage Disconnect is a feature of our Battery Management System to protect the battery from damage. Learn how to "wake up" your battery.

battlebornbatteries.com

Low-Voltage Disconnect is a feature of our internal Battery Management System (BMS), which comes inside all Battle Born batteries and has lots of other safety features programmed into it to protect your investment and preserve the life of your battery. When fully depleting a Battle Born battery, the BMS detects when the battery’s voltage falls below 10V, and it will disconnect the battery.