BMS Help Needed

[d8veh]

I'm trying to help a guy get his bike going again. The history was that it stopped working, and the problem was diagnosed as faulty BMS, so a new one was bought and then the charger went pop when it it was plugged in, so he got a new charger, and the new charger has never charged it. The green light is on all the time. I measured 60v on the charger output.
All the stuff is from Conhismotor: 48v 30AH LiFePO4 battery, 5 amp charger, 60 amp BMS
Here's the links:
http://www.conhismotor.com/ProductShow.asp?id=132
http://www.conhismotor.com/ProductShow.asp?id=154
http://www.conhismotor.com/ProductShow.asp?id=134

The battery has a 16 pin balance lead connector, although the description of the battery says 16S, so I'm guessing that it uses the battery 0v as the 17th lead.

The problem at the moment is that with the battery measuring 52v, I'm only getting 36v out of the BMS and the charger doesn't want to do anything, so my questions are: Did the BMS get damaged when the first charger went pop (maybe wrong polarity?). If so, would it be repairable? Could someone recommend a not to expensive 40+amps BMS, ideally with the 15 pin connector, or should I get yet another replacement from Conhismotor?

Any help appreciated.

 

[dnmun]

you will have to check the gate voltage on the charging mosfets. and check the gate voltage on the output mosfets too. that is the ping signalab BMS for high current discharge.

 

[d8veh]

Thanks for the quick reply.

I did a bit more investigation: I had checked all the cell voltages at the multi-pin connector. They were all the same at 3.3v, then I reailised that that only accounted for 15 cells, so I checked between ground and pin 1 and I only got 1.7v. I thought maybe that was the problem, so I put cell one on charge. It's now on 3.1v, but I've got nothing (a few mv) on any of the FET gates and still 37v on the output and charge socket.

 

[dnmun]

only the output mosfets would be turned off if the voltage on #1 was below 2V. as soon as it crossed above 2V then the gate should show 14+ volts. the charging mosfets gates will show 8-10V on that v2.5 signalab. the charging mosfets are not turned off when the voltage on one channel drops below 2V. the voltage on the charging mosfet only turns off when the voltage on any channel climbs to 3.95V or so.

 

[d8veh]

only the output mosfets would be turned off if the voltage on #1 was below 2V. as soon as it crossed above 2V then the gate should show 14+ volts. the charging mosfets gates will show 8-10V on that v2.5 signalab. the charging mosfets are not turned off when the voltage on one channel drops below 2V. the voltage on the charging mosfet only turns off when the voltage on any channel climbs to 3.95V or so.

Does this mean that it's scrap and I have to get a new one?

 

[dnmun]

no, it means it is working. if the mosfets were shorted the battery would charge but i need to know if you measured the correct spots.

put the black probe where the BMS says B-, red probe on the leftmost leg of the mosfet.

the charging mosfets are the two in the corner where the blue wire is attached. all the pothers are the output mosfets.

all the gates are tied together so you need measure only one of them to know if it is turned on.

it should be working if you can measure the voltages on the pins at the sense wire plug.

 

[d8veh]

I've got 8.5v on both sides of the charging FETs and 0V on all the others - both sides.

 

[dnmun]

what does sides mean?

there are two mosfets in the corner where the blue wire attaches. you can look underneath the pcb and see how there is a section of silver where the source legs are soldered in place. those are the two charging mosfet. that is the back side. on the front side are the two mosfets side by side. each mosfet has a gate drain source leg from left to right.

you can see that the two gate legs of the mosfets are tied together, so they will both have the same voltage. that is what you measured to be 8V?

does the source leg also measure 8V too?

the third mosfet over is one of the output mosfets, there is a gate drain and source leg.

the gate leg measures 0V? just that, you don't have to measure everything just the gate on each of those two.

 

[d8veh]

I should have said both legs, so yes. both gate and source legs on the charging ones are 8.5v. Both source and gate legs on the others are 0v. Thanks for your help.

 

[dogman dan]

Problem could be simipler than you think. Make sure you are getting a connection at the plug between the charger and the bms. It could just be a broken plug on the battery side, or a pin backed out of the housing.

Green light means charged, or it means disconnected.

 

[dnmun]

having the source leg and the gate leg at the same potential is new for me.

the output mosfets should have 14+V on the gate though. but that is a separate problem for now.

normal operation of the mosfet is when the gate is raised to the 8-10V ABOVE the source leg potential. in this case both the gate and source are at the same potential so the charging mosfet is turned off. the source should not be 8V. but there is a 1 Mohm resistor between the source and gate so that is how the source leg is pulled up.

normally the charger negative voltage applied through that silver trace to the source leg would make the source leg go to 0V.

does this makes sense?

since the source is floating high, then it may be a bad connection from the blue wire on the BMS to the negative spot on the charger. if you open the charger, while it is unplugged from the AC so not active electrically, and with the charger plugged into the BMS, find where the negative output wire comes off the pcb inside near the fan. the shunt loop wire is part of the negative so it can also be used, but you want to measure with the ohmmeter scale on your voltmeter the resistance between the blue wire on the BMS and the negative output of the charger, putting the probes on the negative of the charger on one end and on the source leg of the charging mosfet on the other end.

that has to be 0 ohms, no resistance. or a continuous connection. you are checking continuity of the wires from the charger to the BMS.

i am assuming that the reason the output mosfets are turned off is because you still have low voltage on one of the channels and the BMS is shut down because of that.

once the charger finally starts up it should allow the LVC shut off to clear. but you wanna charge the pack up once we solve this problem with the charger.

 

[d8veh]

A little progress. Thanks both for your help. I found that the wire between the charge socket and the BMS was open circuit, so I put a new one in and now it's charging. Over-night, I charged Cell 1 up to 3.3v to match the others with my lipo charger. With the charger on I get 53v on the main BMS output wires, but as soon as I disconnect the charger, it drops to 37V again.

Another question if I may: I'm using a 2.5 amp 36v Ecitypower charger, which has three trim pots inside. The one in the corner adjusts the voltage, so I turned it up to 60v. Do I need to turn down the current with one of the others to stop it over-heating? It seems OK at the moment.

 

[dnmun]

if you open up the other Vpower plastic charger you will see that the ICL is blown up and you can fix that and use that charger.

you wanna charge that pack up so that all the cells are measuring 3.65V or more when you measure them while charging.

 

[d8veh]

I think I've sorted it now. In the right light I saw what I thought was a crack in one of the resistors, so I got out my meter and found the whole row were open. I've now got 57v on the output and it's still charging. Thanks very much for your help and time. Other than adding some hot-melt glue to stabilise them is there any better way of doing this:

 

[d8veh]

I just put my meter probes each side of the resistors and got zero. All the other resistors showed a value.

 

[dnmun]

post up a picture of the back side of that BMS. it is not wired up like the high current siganalab i have and you have no leds on each channel either. but it looks very close so take a picture of the backside.

i think there may be one more resistor on the back side, which would make 9 total so you could have used just one resistor instead, Req=330/9. you wanna use the same resistance as the other channels so it will balance properly.

do you have the original BMS yet? or did you throw it away?

we should fix that one and the other charger.

 

[dnmun]

if you measured 0V across the shunt resistors then the shunt transistor is most probably turned on already.

usually when the shunt resistors are open circuit (which is infinite resistance not 0 ohms which is a short) then those resistors are all burned up and discolored so much that you cannot read their values on the surface.

you need to get out the old BMS and we can fix it too. that one may be easier to fix.

 

[d8veh]

You're quite correct. It's not charging all cells. About half of them are still on 3.3v and the rest are all rising. I'll try and post a picture later.

I don't have the old BMS because it's not my stuff, but I think the owner might still have it; however, he doesn't live nearby, so not easy to get hold of it. The owner posted his woes on Pedelecs forum. He was getting desperate after it went pop every time he tried to connect something new. I felt sorry for him so offered to help. It's a lovely bike - one of those low-rider types. I think the root cause was that the chargers are always wired back to front with the 0v in the "L" and the 60v in the "N".

Do you think it's worth proceeding with this or should I just get a new BMS?

 

[dnmun]

since you said the channel one measured 0 ohms when you measured it i am thinking that the problem is that the shunt transistor is on all the time.

did you remove the old shunt resistors that were there before soldering on those extra ones? did you measure the resistance across them after you soldered in the new ones?

you can test the shunt transistor where it is. that is why i wanted to see the back side. the BMS i have like that has one more shunt resistor and the shunt transistor on the back side so i wanted to confirm.

if you look at the shunt transistor, which is a p channel mosfet, you can see where the source leg is tied to the trace coming off the sense wire plug. that is the high side of that channel. the single leg coming off the other side goes through that through hole to the shunt resistors on the front side when the shunt transistor is turned on.

the other leg next to the source leg is the gate. there is a trace running from the comparator up to the SOT-23 transistor under it to the gate leg.

if that channel is still below 3.6V then the gate is at the same voltage as the source leg and you should be able to measure 0V across the source-gate.

when the shunt transistor turns on the gate will be 3.6+V BELOW the source leg. red probe on source leg, black probe on gate, should measure 3.6V. if the gate is turned on then it would look like 0 ohms when you measured across the shunt resistors.

so the fix would be either to replace the shunt transistor or the 5 pin comparator that drives it. i have had to do both before in order to repair a BMS but it is not an easy task unless you have skilled steady hands and eyes and a good fine tipped soldering iron. you will need solder wick and flux helps and fine wire 60/40 solder.

but first lemme know if the old surface mount resistors are still there, a picture of the back side, and if the gate of that shunt transistor is turned on.

i have parts if you need them to replace them. but if you can get the other BMS then it may be easier to fix for the guy.

i know the charger is an easy fix. those Vpower chargers all blow the ICL.

 

[d8veh]

Here's a picture of the backside of the board. I can see that the resistor is blown. I'll do some more measurements tomorrow. It's late here now. The measurements I made on the resistors earlier were bullshit. I've got a new automatic multimeter that has a delay before it shows a reading; however, there was a clear difference between the resistors on that row and the rest. Now that I know to wait for the meter stabilise. I'm getting about 36 ohms on the good row and 4 ohms on the bad row. I guess I'll have to remove all the SMT resistors and replace them if I can. I've done it before, but my fingers are not so steady now and neither can I see so well.

 

[dnmun]

nono, don't remove the shunt resistors, you need to test the shunt transistor first. if it is turned on your measurements mean nothing. since you had low voltage on that channel i think that is the most likely suspect.