That was raised a couple of days ago, can you please raise a bug on github.
Thé system was originally designed for 48V usage, but does work with higher voltages.
Many also talked about the problem of turning the relay on and off at the voltage limit values. In many systems, this is solved by two values. The first value is disable, the second is enable. Thus, by spreading these values among ourselves, we get rid of the problem of relay bounce.
Did you ever solve your issue with the CRC errors?
Similar setup to you, works fine when running on mains USB PSU, but get a few CRC errors when running on a stepdown converter connected to the batteries. Tried multiple converters and still same results.
Jon
As others have pointed out it could be your batteries, but yes, the rules need to have a recover value to prevent the oscillation. I made a version that had that with the rules, then rewrote that to be less generic. I don’t have rules, so much as I have max/min and recover settings for cells and the pack and temperature. Then you designate your relays to be either load or charge, and the logic of the code does the rest. You can also set the relays to respond to state of charge, but that’s only possible if you are monitoring current, which I am doing.
I don’t think I have the rules style code anymore, but I committed the one I am using to https://github.com/johntaves/diyBMSv4Code
Thank you. I’ll try
Can you track down where the CRC error is being introduced is it in the modules or the comms back to the controller?
What does the configuration web page show, are there high error counts on the modules?
So the chip used by JLPCB was different and ended with a “T” instead of a “P” I had to change the code to correct the error.
Just a note to say thanks mainly to Stuart and also others on here for all the work that has been put in over the last few years. I’m glad I hung in there as the v3 BMS nearly drove me nuts when it just wouldn’t communicate. V4 is so much better and improving all the time with the updates.
I started my battery storage project almost exactly 2 years ago when I collected my first cells from the local tip and started processing them. I now have just over 10kWh of storage in a 14s100p configuration though I’m only cycling about 40% of the capacity to prolong the life of the cells (all 1600mAh+ and 80%+ of original capacity still).
In the last 24hrs I’ve only imported about 0.25kWh hrs from the grid, roll on next summer!
My project is still a work in progress as I’ve not got current control of the Flatpack power supply implemented yet. This is also only a temporary location.
Hang in there everyone still working on their own build it is well worth it in the end.
6 Likes
@Keith great powerwall. Really interested to see you using 1600mAh cells. I’ve been looking at 1500mAh for my own powerwall, working with multiple 7s18p banks.
Hi.
Some findings on our packs.
Don‘t use shared wires, even 5cm longs for the cell modules,
Always use 2 extra wires from each cell module to the battery tap/pole.
If we use - from first cell and + from first cell for the first module, but the same + wire from cell 1 as - for module 2 the meassured voltage on module 2 will rise when balancing starts on module 1.
The reason is that the minus voltage is dropping with the cable resistance on 800mA balance power but the + voltage stays where it was, because module 2 is not balancing yet.
Then we had a chain reaction up to the last cell, even if they were balanced first, all 16 modules start to burn energy
After cabling every cell module directly to the battery this „feature“ has gone.
1 Like
I just looked at the price for the mouse bites PCB here at Open energy monitor…
21.18 pounds or 25 euro / 28 dollar
18 august I ordered 30 x cell pcb at JLC all parts assembled (include attiny841, not pin headers) 85 usd.
85: 30 = $2.83
28:4 = $ 7.00
That is a lot extra for connector and pin headers.
Obviously I like to sponsor @stuart, that is not the question.
My idea for the mouse bites was to reduce the costs, not triple it!!!
If you need sponsoring for the great work at the DIYPCB please provide your PayPal.
Abusing my suggestion for mouse bites this way make me feel really bad for even giving this idea.
Really.
Its cheaper to buy 30 pre-assembled from JLCPCB then 3 sets of mouse bites (=12)
and you don’t even need to solder the attiny841.
Not that it is a big problem, but I did manage to kill 7 during my last hand solder set of 30 pcb…
You really should provide 6x attiny841 in the set for this price as there can easy be a few mishaps.
Perhaps I was misreading, and the kit is 16 cell modules for 27 dollar??
(in case people wondering, I have 80 cell setupS16, total 1016Ah, and like to provide each module with a cell PCB.
that will use the 5x controller boards I have 
)
I accidentally ordered a set from JLCPCB that did not have the attiny. I think someone claimed the qty between when I checked and when I pulled the trigger. I wasted a ton of time learning how to solder those mini pins. It was so difficult to find the #1 pin indicator. On some of the chips, the dent was invisible. I thought I correctly figured out what orientation the writing was, and I was able to see the writing if I held the chip at the right angle with the right light. I soldered 4 before realizing I should test it. I was wrong. Of the remaining 6, one crapped out. I bought the desolder stuff and was able to get the chips off. I resoldered one, but it didn’t work. The second one worked. The cats were hiding in terror. I vowed never to buy any board that requires surface mount soldering, even though I am now somewhat competent at it.
The time cost is ludicrous.
The experts have no clue how much they know, and how many ways a novice can screw things up.
I will run my 8s pack with just one module before I buy another module where I have to solder the attiny. I check lcsc.com every day to see if they have the attiny in stock.
My kill rate was 4/10.
1 Like
Novice to novice…
I suddenly feel expert by just frying 7 out of 30
Thanks @John_Taves for backing me up here.
It’s not that easy and really takes hours to do, or in my case, 3 weeks as the thermistor (R19) was out of stock and I needed to do that one by hand also.
Compared to a 0603 the attiny was easy.
Man…
I used to be proud to be able to solder the jumper connector!! (2.54 things)
This is a whole different ballpark that @stuart seems to find hard to understand that most people really don’t have the skills to solder attiny841.
@John_Taves, I think we should talk PM…
Quite possibly, until I started the diyBMS project I’d also never soldered SMD stuff. I’ve only recently got a hot air soldering station, before that I used solder paste and a frying pan!
Don’t try and solder the ATTINY841 with a soldering iron, you can do it but likely to bridge multiple pins. A little solder paste and an oven/hot air station or even a frying pan all work and the paste magically “sucks” to the pads and pins like magic.
2 Likes
Hi Frank, sorry if you feel this way, I was in discussion about stocking the parts with Glyn at Open Energy Monitor long before any mention of mouse bites cropped up, they seemed like an ideal format to put into the shop as most people buy the modules in multiples of 8.
Obviously the shop has overheads, there is also the issue of import duty and VAT which UK payers are subject to (for JLC orders).
You do realise that you can add multiple PCB orders to a single JLCPCB order? To avoid multiple shipping charges.
As you have seen JLCPCB rarely have ATTINY841 in stock, this is often cheaper to purchase from the main suppliers like Farnell/Digikey etc.
Pardon me for butting in, but we all had to learn at some time or other. It’s just that the experts have done their learning, but you’re doing it now.
1 Like