diy solar

diy solar

Cobbled together, never finished

stienman

Mostly Harmless
Joined
Jan 6, 2021
Messages
476
Received the batteries last week, the poron yesterday, and built a temporary testing fixture to avoid bulging. Popped the chargers on two of the cells and was pleasantly surprised when they both completed their charging sequence at the same time with nearly the same AH.

I may end up building an 8 cell harness and testing the remainder of these cells 8 at a time if the discharge test shows these are all that well matched.
E5593DD9-4A9D-4084-B267-E459A1EFBA5E.jpeg

32 of these used (EV pack, Michael) 271AH CATL cells will be added to my current system:

Growatt off grid 6k (SPF 6000T DVM)
Solar city 6k grid tie (delta H6)
64 thunder sky 60AH cells (circa 2008, about 60% capacity)
diyBMS(monitoring and passive balancing only, no control or coms yet)
25 REC 315 panels(7 on solar city, 18 on growatt, half on roof, half not mounted, just laying against the outbuilding)
 
Looks like I'm going to stick with the "temporary" fixture style. Will build 3 more, and then put them inside a metal cabinet as my battery box.

So far the batteries are matched well enough for my needs. I've tested 6 (charge, discharge, charge) and have
266
268
265
270
264
267

These are 271AH cells, but I'm only charging them to 3.45v, and discharging to 2.50v, so the 1.5% lower capacity than rated is expected, and I suspect if I took them all the way to 3.65v I'd see over 271AH. I'll drive them to 3.65 when I top balance.

Anyway, I'll finish the first 8, then connect them into a 24v pack and use that as the power source for the iCharger X8.

The chargers are very nice units, but can only handle dissipating 50w, or 30A, whichever is greater. So while it'll charge at 30A, it will only dissipate a single cell at just below 20A.

Connect 4 cells in series and it'll only dissipate 4A, taking three days to discharge.

I can connect a resistor in series, and it'll dissipate up to its 30A limit, as long as the internal dissipation is under 50W.

Alternately, I can connect the unit to batteries as the power supply, and have it push the energy back into the source batteries (from the batteries under test), dissipating very little in heat, and still allowing for a much higher current draw on the cells under test.

This saves a lot of time, and energy.

The only thing I'm disappointed with is that its power supply input isn't rated high enough to handle a typical 48v battery pack, as it's only rated to 49v max.

So I'll build a 24v pack out of these first 8 batteries, and use that as the power source for testing the remaining 24 cells.
 
Well, my inverter gave up the ghost and let out ALL the magic smoke about a month ago. Signature Solar has been less than helpful, even though the inverter is under 1 year old, so the whole "supported in the US" schtick isn't worth the additional cost.

So my solar has been turned off for the last month, and I'm re-evaluating my overall design, as I now see that I can't depend on quick support or resolution of problems, so I have to build redundancy into the system.

While I really like the low frequency inverters, and their ability to handle large inductive loads, they don't parallel, therefore they can't easily/cheaply be made redundant. Which means I'm leaning towards the 5kW SPF high frequency off-grid inverters from Growatt. At $900 each it's not bad, and if I buy 3 then worst case scenario I'd have 6kw of power available, and with care we'd be fine with that, and generally we'd be able to run the house without concern on the 15kw that'd normally be available.

It kills me, though, that I'm losing $300 of solar energy a month due to an inverter issue that won't be resolved quickly not because it can't be, but because my supplier doesn't provide assistance to fixing the problem myself - I have to ship it back if I want it to be serviced, and that's an expense and waiting period I don't like either. They won't determine whether the fault is under warranty or not without inspecting it themselves, and they don't offer cross-shipping a replacement. The inverter shows no errors, it inverts and charges from AC fine, but it won't charge from solar.

Anyway, I'm in a holding pattern until I get enough money together to buy the first inverter of a parallel set.

:(
 
Oddly enough, I just received an email from them with a pre-paid shipping label, so perhaps they are going to provide the service I was expecting! Still very slow.
 
They've determined the problem is covered under warranty, and gave me several options beyond just repair or replace. I'm very happy to have the additional flexibility, and it looks like I'm going to be able to use the refund towards a more powerful inverter.
 
I chose to upgrade to the SPF 12000T DVM, and they were able to get it shipped to me in a few days. It's installed and tested with battery and solar, I'll configure it and turn it on tomorrow. As this model was on sale, and they fully refunded my previous purchase, the cost to upgrade, including shipping, was just over $800.

I took the opportunity to move the inverters to a different location where the noise won't impact my main workspace, but I'm keeping the batteries inside until I put up walls and insulation inside the solar room and enough heating to avoid freezing.

Anyway, Signature solar honored their warranty, and have been good to work with. Hopefully this inverter won't have whatever problem the previous one exhibited.
 
Batteries:
I assembled one of the batteries months ago, and have been delaying adding the DIYBMS to them due to time.

Just checked the cell voltages today, and they are within 3mV of each other. I'm starting to buy Will's insistance that LiFePO4 batteries don't require a BMS if they are new and matched.

I'll still add the BMS later, but I'm not going to fret over it. Keep in mind I'm running my frankenstein pack in parallel, and due to its limitations I'm running very conservative charge voltages, so that's undoubtedly helping prevent cell overrun and imbalance issues.

Growatt inverter:
One of the MPPT converters in the Growatt SPF 12000T DVM is broken, though. It has an 1.8kW array attached, and in full sun it never produces more than 1kW. Looking at the voltage graph it looks like it never goes above pack voltage, and the maximum power point should be around 113v.

I haven't even bothered contacting Signature Solar about it - they made it pretty clear last time that they don't allow users to troubleshoot or repair their units, and require them to be shipped back for warranty service at the customer's cost if they determine the fault is not covered by the warranty. With a unit that requires a pallet jack this is not an inexpensive option. They used to say "Buy a spare controller so you're never down!" but when I asked about buying parts and dealing with it myself, at my own cost, they indicated they can't sell parts, and can only perform warranty service.

So I bought the 12k knowing that if I had problems (like this) I'd pretty much have to work around them unless they were so significant that sending it back at my expense was worth it.

So I'm buying cheap 60A MPPT solar charge controllers, and letting the inverter just be an inverter/AC charger/bypass switch. I'll use the remaining MPPT inside the growatt until it, too, gives out, but it's been an overall lackluster experience.
 
Batteries:
I assembled one of the batteries months ago, and have been delaying adding the DIYBMS to them due to time.

Just checked the cell voltages today, and they are within 3mV of each other. I'm starting to buy Will's insistance that LiFePO4 batteries don't require a BMS if they are new and matched.

I'll still add the BMS later, but I'm not going to fret over it. Keep in mind I'm running my frankenstein pack in parallel, and due to its limitations I'm running very conservative charge voltages, so that's undoubtedly helping prevent cell overrun and imbalance issues.

Growatt inverter:
One of the MPPT converters in the Growatt SPF 12000T DVM is broken, though. It has an 1.8kW array attached, and in full sun it never produces more than 1kW. Looking at the voltage graph it looks like it never goes above pack voltage, and the maximum power point should be around 113v.

I haven't even bothered contacting Signature Solar about it - they made it pretty clear last time that they don't allow users to troubleshoot or repair their units, and require them to be shipped back for warranty service at the customer's cost if they determine the fault is not covered by the warranty. With a unit that requires a pallet jack this is not an inexpensive option. They used to say "Buy a spare controller so you're never down!" but when I asked about buying parts and dealing with it myself, at my own cost, they indicated they can't sell parts, and can only perform warranty service.

So I bought the 12k knowing that if I had problems (like this) I'd pretty much have to work around them unless they were so significant that sending it back at my expense was worth it.

So I'm buying cheap 60A MPPT solar charge controllers, and letting the inverter just be an inverter/AC charger/bypass switch. I'll use the remaining MPPT inside the growatt until it, too, gives out, but it's been an overall lackluster experience.
I think you misunderstood Will's recommendation: I never heard Will say not to use a BMS. I think that your mixing up with an active balancer. I've heard Will say that if the cells are well matched and balanced and if you charge/discharge them with low C rates, than you will probably be fine without active balancing.
 
Yes, yes, yes, I understand, I don't intend to kick a hornet's nest with the BMS vs no conversation - I intend to install one. For those interested in BMS discussions there are thousands of posts on dozens of threads on this forum, and we don't really need another one.

My post was only intended to express surprise and happiness that these cells are so well matched.
 
No BMS? That's a terrible idea.
I think it depends on how hard you drive your packs and how good they are to begin with. I have an 8-cell pack that has been running for six months. Voltages are within .01 of each other. No BMS... Even my old BYD packs are within .1 volts. per cell. I run them 25v-27.1v so there is no stress on the packs.

I think BMS's have their place, but you can definitely run packs without them.
 
If we are going to have a BMS discussion here, please 1) define what a BMS is for the purposes of your post and 2) provide evidence or at least experience based information. Some people think of a battery protection system as a "BMS" while others think that a BMS includes balancing (active or no), and still others have something else in mind.

Battery protection (no over or under voltage) is an absolute must, but most equipment that connects to the battery have this built in. Adding it to the battery pack is a redundant protection that is well worth the cost in time and resources, but not always done - for instance in vehicles they don't duplicate this aspect of it since it's considered one integrated system.

Battery balancing is necessary as the pack goes out of balance, otherwise you eventually have runaway cells which will either trip a cell-level battery protection or swell/vent/etc. But battery production and binning (matching) has come a long way in the last two decades, and a battery built from well matched cells and not used close to its ends may only need to be balanced (often a top balance) once a year. Always on balancing, whether active (moving energy from cell to cell) or passive (loading cells that are above the pack average) is an additional protection, useful when packs have unmatched cells, or are heavily utilized and nearly fully charged and discharged frequently.

Cell level monitoring and reporting provides advance warning of runaway cells, or changes in IR which can indicate potential future failure or reduction in storage.

Temperature monitoring not only indicates when a pack is overheating, but when a pack is too cold to be charged without damaging the cells.

There are many other features and aspects of a "BMS", and not only would we need to discuss how critical it is, but also where it should be handled - given the advances in inverters, charge controllers, etc a lot of the minimum protections are available without having a purpose built BMS.

Further, LiFePO4 can handle a lot of abuse - nowhere near as much as lead acid, sure, but a lot more than people realize before they pose a safety issue. Most BMS usage for this chemistry these days is intended to extend the life of the batteries, and goes beyond just protecting the user/structure from dangerous battery failure.

There is no real "yes/no" discussion about BMS, there's only risk/reward and feature discussion.
 
Back
Top