DIY Electric Car Forums banner

Detachable BMS; should it work?

7K views 13 replies 7 participants last post by  peskanov 
#1 ·
Hi,
Working in our conversion (a Mohave Quad) some time ago, we disliked the idea of having a BMS, with some many cables and conections, attached to the battery box.
Having worked a bit with RC LiPo packs, I thought we could emulate their system. That is, having the BMS outside the vehicle, along with the charger.
That way you avoid having a fragile circuit jumping in the Quad, but still get full cell balancing when charging.
Here there are some pictures of the components we are using to build our system:
Charger on the left, BMS on the right.

Another view,

Big, industrial plug,

The cells,


Now that we are building the pack, I am starting to have some doubts. So these are my questions to the experts here:

- Small wires from the cells to the BMS will be 50-70cm long. Could this length cause a voltage fall, fooling the BMS?

-Can frequent plug/unplug system affect the BMS (i.e. sparks from capacitors)?

-Is there any other problem I should know/prevent for this kind of system?

The components we have:
Charger: http://www.evassemble.com/index.php?main_page=product_info&cPath=3&products_id=173
Cells: 24 Sinopoly LiFePo, 60AH. Bought from Mike Schooling on this forum.
BMS: http://www.evassemble.com/index.php?main_page=product_info&cPath=6&products_id=46
 
See less See more
4
#2 · (Edited)
we disliked the idea of having a BMS, with some many cables and conections, attached to the battery box.
I think you identified the issue correctly. But it seems to me that then you went about the solution in the diametrically opposite direction, actually exacerbating the problem.

Clearly, you've made a choice, and you're going to live with it.

Still, since you asked the question, let me suggest that a distributed BMS would have solved your concerns quite neatly.
 
#3 ·
Now that we are building the pack, I am starting to have some doubts. So these are my questions to the experts here:

- Small wires from the cells to the BMS will be 50-70cm long. Could this length cause a voltage fall, fooling the BMS?

-Can frequent plug/unplug system affect the BMS (i.e. sparks from capacitors)?

-Is there any other problem I should know/prevent for this kind of system
Small wires: If you are shunt balancing over these then the wires need to be heavy enough to handle the current. If the BMS measures the voltage with the balancing current turned off the length won't matter because there is no voltage drop. What does matter is that you will want a fuse on every wire connected to the battery at the battery end of the wire just in case.

Plug/unplug: Probably not sparks from capacitors as there shouldn't be any caps on the voltage sense lines. But static electricity could cause issues. If a mechanism is used to alleviate this your problem will be wearing out the connectors from the daily use.

Other problems: Lots of potential issues. The one thing that comes to mind is if you are going to public charge you wont have your BMS because it is back in your garage.

Having installed your BMS this way only covers issues during charging. The bigger issue is bricking cells when you over discharge. Charge is actually a non issue once you have balanced the pack and selected an appropriate CV point for your charger. Discharge is more difficult because you can't just pick a voltage and declare you will never go below that. Load changes this voltage. Your pack can easily see voltages below 2.0vpc under heavy loads and this is not harmful to the pack. This is why bottom balance as a protection mechanism works so well. You bottom balance and then it is very hard to brick a cell when you drive the car to the point where it wont move.

As an engineer I love data. I would like to see the voltage and temperature of every cell in the pack. I have talked myself out of a BMS because the data is useless for the normal operation of a car. It is useful only when things go wrong. A half pack voltage comparison will tell you when something is going wrong with a cell during charge and discharge. A bottom balance will protect you from yourself when driving the car. Undercharging the pack will protect you while charging. The additional wires and circuits is a reduction is reliability. My conclusion is that a BMS is not necessary with LiFePo4 chemistry cells and in fact may cause more problems than it cures. Ask yourself how many extra cells can you buy for the cost of the BMS.
 
#4 ·
Elithion, dougingraham, thanks for your expertise!
I wanted to try one charge before coming back to answer, and as usual putting all in place took much longer than anticipated.

I forgot to mention in my original post that we wanted strong regenerative braking and a cheap BMS system. That was one on the reasons for this experiment, as cheap BMS are no able to handle the higher charging currents (>50A).

Now, about our first charges. When we put the system to charge, we tested the voltage of several cells and compared to voltage on BMS plugs, to detect any voltage fall.
We have 24 cells. Cell 1 plugs on the negative wire, cell 24 in the positive.
What we found was important voltage falls in the extremes on the pack, but none in the intermediate cells.
Examples:
Cell 1 -> 3.44 on bms, 3.30 on cell
Cell 2 -> 3.33 on bms, 3.31 on cell
Cell 15 -> 3.31 on bms, 3.31 on cell
Cell 24-> 3.40 on bms, 3.31 on cell

Does anybody knows what could be going on here? Are our wires too thin/large?

Elithion,
you are right, but we have not found a BMS like that as cheap as we needed. We are using one of cheapest ones in the market, I think. Only e-bike and RC ones are cheaper!

dougingraham,
as you predicted, we have had no problems on pluging/unpluging the BMS.
We are thinking now about splitting charger & BMS, so at least we can load the charger on the Quad and be able to perform opportunity charging (without balancing).
Thanks for the fuses suggestion, we will implement it.
About undervoltage control, the controller should protect us (it has a setting for that) and we also have a cheap cell state display system. When a cell goes under ~3V it starts beeping.

The whole thing about bottom balancing is new to me. I have read some references before but I still have to really learn about it.
 
#7 ·
Elithion,
well, as we cannot afford a good BMS, controller protection is better than nothing. We also have alarms in the cell monitors. Poor, but again better than nothing.

DIYguy,
we have 3 cell monitors, 8 cells each. We can (visually) compare the voltage of the 3 packs, and also inspect each cell to find any fault.
 
#8 ·
If you are getting higher voltages on the extreme ends of the pack, if it is truly the cell voltage, it could be a difference in temperature. But if it is just your readings that are off, I suspect that the BMS draws its operating power from these ends of the pack, and the measurement wire is shared with the power. It's easy to get 100mV or so of voltage drop which would skew these readings. If you can, run separate signal and ground leads, or use the heaviest wire you can for the power connections.

Do you have technical details of your BMS?
 
#10 ·
Hi Luis,

Great to see your project coming on :D

In my initial testing of some BMS systems I have found that voltages read differently on the cell terminals and on the end of the BMS wires.

The best method I found was to have all wires the same length no matter where the cell is in the pack - essentially find the longest lenth you will need and cut equal lengths for all your cells. This means that voltage drop is as even as you can make it. Also consistently use good connectors with good crimps, same type of wire etc..

I do think that the distributed version of miniBMS could do what you need it to.

Cheers,

Mike
 
#11 ·
PStechPaul,
cells are cool, no heat at all. We are charging at modest rates, 6 amps for 60ah cells.
We are thinking about upgrading the wires. Right now we use thick wires for + & -, and thin wires for each individual cell conected to the BMS. I don't have the size number here, unfortunately. Lenght is about 2 feet (maybe a bit more) from cell to BMS conection.

I don't have much info on the BMS, we bought it at evassemble. There a few specs there:
http://www.evassemble.com/index.php?main_page=product_info&cPath=6&products_id=46

major,
yes, this one or a clone. We bought it from Ali-express.
Connection is very unreliable, we are going to solder a new one on it soon.

skooler,
hi Mike, yep the project is moving fast this summer, thanks.
The Sinopolys are behaving well (granted, we are not pushing them hard yet) and the old motor seems to cope well with the load, but the little Kelly KIM we bought is probably too weak for it.
Well, that's life. As I told you, our budget was small :)
I should post a video soon, on AEVA.

What you say about the wiring sounds right, so I guess we will have to do some re-wiring there. However, there is still a lot of work to do on the rest of the vehicle so I guess we will check the BMS later. I should have calculated the voltage fall looking at amsp anf the wire sizes before connecting all the mess!
 
#12 · (Edited)
Hi there; after all those months, the BMS problem has evolved and we finally removed it from the charger. I am posting our observations here in case anybody founds them useful.

As the bad balancing persisted, we tried to experiment adding more wire to the suspicious cells, to see if it was a problem of voltage falling due to thin wires...it did not change results at all. Same unbalancing.

A few month laters, the bad balancing mismatch grew, letting half of the battery empty. It was time to try something new, so we went for bottom balancing.
We removed and discarded the BMS, discharged all cell to 2.75V, and charged until the first one hit 3.5V. It worked ok, all cells were pretty close to 3.4V, so we have built a 3.5V detector for 3 cells and set a relay for charge end.

We have been running the package BMS-less for a few weeks now, and all cells seem good. The pack charges correctly, end of charge is not premature, all cells being close to full charge.
The pack also discharges nicely, cells are always 0.00 to 0.01V from each other.
This pack of Sinopolys has already a lot of work hours on it and is behaving wonderfully without the balancer, so it's pretty clear the BMS was ******* it!

I have read the threads about "bottom balancing yes or no", here and in the endless sphere forums. Arguments from both side are pretty convincing, but in absence of cheap and reliable BMS, I am joining the bottom balance side, it works perfectly!

In defence of the BMS, though, I must say we found a possible cause of troubles. When we installed the BMS and the charger in the same box we observed the voltmeter and ammeter behaved poorly, showing unstable readings. After some weeks both of them died. We were puzzled, and I suspected of PWM. Maybe the meters were not ready for a PWM source...
Recently we discovered the cause: we installed the charger without its aluminium box, and therefore let it's induction field hit the nearby circuits. It seems to be a strong induction field, as it caused havoc with the meters.
The BMS was also close, so maybe it was also affected by this problem.
 
#14 ·
Hi Mike,
we are usually very careful about not discharging the battery. In the last year the lowest we went was 3.1V. But it's good to know the cells can survive a full discharge!

Yes, I want to shoot a video; the Quad is finished and has a nice Curtis 1236 now (bought used). I also have access to a cool gimbal camera perfect for that kind of movies.
Eventually I will do it, but now I am too busy working in my car conversion. It seems I got the EV fever and it will not le me go... ;)
 
This is an older thread, you may not receive a response, and could be reviving an old thread. Please consider creating a new thread.
Top