Ping output voltage issue - bad BMS again?

[mrzed]

All,

I've got a 48v/15Ah ping v2.5 - last year, the BMS got replaced as one of the charge LED's was staying lit, and a single cell group got drained below the remainder. The weak group remained above 2.8V, so the damage was likely minimal. The BMS was replaced, and other than that weak group, the battery was working until one morning when nothing happened.

Checking the output voltage on the power leads now shows between 13-14 volts, but checking the balance leads shows that 13 of the 16 are all 3.65 when fully charged, with the remaining 3 only slightly lower (3.5+).

I'm not sure where to check next - it seems unlikely the power leads for 60% of my cells spontaneously disconnected. The BMS seems to charge properly, can I somehow bypass it to get full voltage out of the battery directly? I'm expecting it is on it's way out, but I'd like to get more service if I can from it. I'm not much for electrical repairs, but I'm not too afraid to try as it will be junk if I can't figure out something. Probably not worth getting another BMS replacement at this point though . . .

Thanks

 

[dnmun]

put the pack on the charger and leave it charging until the cells all balance. measure the voltage of each cell on the BMS where the pins are soldered into the pcb. put the black voltmeter probe on the B- spot for the first cell, and measure each cell from 1-16, list each cell voltage here. measure to 1/100 of a volt when it gets to the point where the charger turns off. take a picture of where you are measuring so we know you are measuring in the right place.

i doubt if the led was lit at 2.6V, maybe 3.6V.

what do you use for a voltmeter? use a digital meter on the 20V scale to measure.

 

[chvidgov.bc.ca]

Seems identical to the issue I had...here...

http://endless-sphere.com/forums/viewtopic.php?f=14&t=44112

 

[mrzed]

I'm measuring with a digital multimeter on the 20V setting. Using the instructions that Mr Ping supplied when I replaced the original BMS - putting the red lead on the pin where the red wire fits on the 16 pin connector and the black lead on the adjacent pin, then moving the pair of leads along to measure each cell voltage.

I did leave it on the charger for about 30 hours now. I've done this previously if only to keep the battery topped up while I wasn't using it. The cell voltages were coming come close to balancing. I got between 3.58 and 3.64 on each.

Today it looks like the high voltage cutoff on one cell group seems iffy - I measured at 3.74 on the 13th group (counting from the first as where the red wire on the connector fits). The lower end is still the same - no lower than 3.59, with most between 3.61 and 3.64.

I looked at the other thread - and the description fits.

 

[dnmun]

so #3 is 3.74V and the rest are just above where the v2.5 balances the cells to 3.6V where the led turns off. there should be 15V on the gate of the output mosfets and it should conduct current. #3 will go to 3.60V when you take it off the charger and they should all stay around there but you are saying there is no current coming through?

 

[dogman dan]

Any toasty looking fets on the bms? Not trying to blame you or anything, but I'm curious what amps you typically draw.

Just hook up a new wire and plug to the main + and - before it enters the bms to bypass it. Then ride like we do with RC lipo. Either have a low voltage warning beeper, or cellog 8's, or something to monitor the pack. Conservative DOD is the best thing, in case one cell is lower capacity. Once you know which cell is the culprit with that, you could just put a voltmeter on just that cell for example. Real wattmeters are the best way, once you establish what you ought to have avaliable you can just use less as measured by a CA or other wattmeter. The downside is, you really do have to pay attention to your monitoring stuff, whatever you are using.

My ping is aging, will be three this spring and doesn't stay balanced like new anymore. I just spliced on two 8s jst plugs to the bms balance wires. Now I can quickly see what is going on with cells using a cellog 8.

Looking at that other thread, any melted looking solder on the board by the fets? My experience is the charge fet is the one that gets hottest. That's why I wondered what you draw. Anothe thing, does the bms get any ventilation while you ride at all?

 

[mrzed]

Dogman,

The controller is a 45A as it was upgraded on warantee due to stock by eBikes.ca and I was not going to say no - but I've limited it using the CA to 20A or less - so I doubt that is the concern. I also rarely go below 50% DOD, and never below 80. My commute got shorter just after I upgraded from my 36V NiCad - so the Ping has had an easy life as far as DOD is concerned.

On the previous BMS - I watched temps carefully and it got quite hot charging, but never more then barely warm on discharge - so I had minimal ventilation while riding and kept it wide open for charge. I suppose it is possible that the new BMS was less tolerant or hotter on discharge.

I'm leaning to your idea of splicing on another set of leads for discharge, though the fact the BMS is now allowing one cell to get over 3.7V leads me to believe I can't necessarily trust it for balance charging any more either . . .

 

[dnmun]

i don't follow the logic of that. nothing wrong with 3.7V.

the cell reaches the 3.7V level because it is full but the other cells have not charged up so there is no way to keep it from going to 3.9V unless you cut the charging current to 60mA. the v2.5 signalab has a shunt current of about 60mA and if the pack is balanced then all of them will divert that amount of current during the CV stage.

he was asking about the heat because the mosfets can get so hot on the output that the solder melts which i mentioned could be the cause of the short between the gate and source in the other thread under your other identity.

 

[dogman dan]

Now that I put some 8s balance plugs on my ping, I see I get 3.8v on the highest cells just about the time the charger shuts off. I couldn't say if it was always this way, or just happening latetly though. Once the charger shuts off, you can see the bms pulling those high cells down.

Shoulda put balance plugs on that thing long ago, if only for the fun of watching what the bms does even if i don't understand it.

Sounds like you weren't running that bms up at it's upper limits with a CA limiting to 20 amps. I was wondering if you had it water tighted enough to melt solder. Mine gets some air, but it also, through the shrink, touches the metal toolboxes I carry a ping in. So you can feel some of the heat, as it somewhat sinks into the whole metal toolbox.

 

[mrzed]

dnmun,

Thanks for the clarification on the high voltage cutoff - I was only concerned as I had monitored individual cell voltages before and never seen such a high measure. Perhaps now that the pack is less balanced, I'm seeing the high voltages climb more as a result.

My "other identity" is in fact Chris - who helped me with a previous issue by generously letting me test my stuff at his house. He has a small fleet of ebikes and lives across town from me. I suspect the similarity of the issue is more due to us both running the same version of Signalab BMS rather than that I have been carefully impersonating two Victoria-based nerds this entire time. I do appreciate your general patience with electrical newbs though, so if you occasionally get testy with us, I'm fine with that.

I'm not electronics savvy - but I'm assuming from context in the other thread that the output mosfets are the row of 5 big black ones with three legs - the middle one of which is cut off, and that the shunt is the set of 5 stiff wires that is attached to that big gob of solder you mention on the back of the pcb and joins to the left or gate leg of the first 4 mosfets. Measuring with the black probe of the voltmeter on the shunt or the B- spot, and the red voltmeter probe on the gate leg I get 0V on all but the last mosfet- with the last being the one that is not connected to the shunt and measures 5.89V.

Visual inspection of the PCB does not show any obvious evidence of melted solder as I might expect to find if it go that hot. I double-checked and the pack voltage is now sitting at 57V, and the high cell has dropped to 3.65 after the day of discharge balancing, so along with your comments on the high voltage measure being normal, I'm starting to consider using the BMS as a charging aid and bypassing it for discharge. Open to any suggestions on troubleshooting to revive it though.

 

[dnmun]

that is low for the charging mosfet gate too. you can use the BMS to charge and then discharge without using the BMS but you need to keep track of where the low voltage is so you don't overdischarge the cells.

3.7V is not high, the HVC is 3.9V for the v2.5 signalab and the cell can climb up there and not be hurt by that kinda voltage until you go over about 4-4.2V. but i have had cells go to 4.6 briefly and still survived. not good though. 4.2 did not seem to hurt the one that i did that to. but 3.9V is the upper limit. i do worry about the cells i abused.

i wonder why the mosfets driver is not turning on the gate since you seem to be ok with the cell voltages. when stuff like this happens to me it is because i shorted the BMS on the battery somewhere. but then the mosfets were shorted on and would not turn off. your charging fet should have a higher voltage on the gate too. stumped.

 

[Jason27]

This is getting a little ridiculous with all these bms failures.

 

[dnmun]

it's only ridiculous to you because you don't understand how a BMS works. if you had a clue you would know what we were talking about. kinda why you decided to skip algebra, because it was too hard.

 

[Jason27]

Lol ..we got a comedian here.

 

[mrzed]

OK, I'm about to bypass the BMS for discharge - so I thought I'd check in hoping someone can confirm how to wire it.

-P is where I am currently connecting the black wire for the main discharge connector. It is here I am getting ~14V. I am going to instead connect this wire to the wire currently leading to -B, where the BMS apprears to get it's discharge connection from, and routs it through that row of MOSFETS. Measuring to -B I get pack voltage of 57V after charging. My question is, do I leave -B connected on the BMS and splice in before it gets to the BMS, or do I disconnect that power from the BMS entirely? It looks like the discharge on the BMS is mostly independant fromt the charge, but I'm not sure if it will still need that connection.

 

[dnmun]

you can just solder a wire to the P- spot and over to the B- spot on the BMS. that shorts out the mosfets so the BMS is on all the time.