36v battery won't charge

[Peter Brigg]

I have 2 electric bikes-both 36v Life po4 type so it is possible to troubleshoot by swapping components. They are used to get around a small farm-doing odd jobs. I carry pieces of wire attached to a box on the back of the bikes and unfortunately the end of a piece of wire found its way into the charging socket on the battery pack-a bit of a spark occurred and the centre pole of the 36v charging socket was destroyed. Naturally I thought that replacing the socket would return all to normal however now the pack won't charge.

I have tracked the charger voltage right into the pack with a multimeter-so there is no problem with connections-the pack just won't accept the charge (just over 40v)

Could the melting of the original socket caused an issue?
Peter

 

[fechter]

You may have damaged the BMS circuit inside the battery pack. You could try using your meter to measure the voltage present on the charger input jack to the battery. Compare this to the working one. If there is no voltage there, you probably need to disassemble the battery to check out the circuit board.

 

[redilast]

I think your only option is to open the pack and check the PCB board as likely something is fried. If you can see visable damage such as blown mosfets or caps you might be able to just swap those individual blown parts out. If not might need to look for an aftermarket BMS.

 

[Peter Brigg]

I have opened up the battery pack and looked at the BMS board-one mosfet is very blackened and burnt-the reverse side of the board is also blackened-will try to enclose a photo.

The existing mosfets are numbered P75NF75 and below that CZ01H-I have some spare mosfets here IRF740-are these a substitute?
Thanks Peter

 

[redilast]

I have opened up the battery pack and looked at the BMS board-one mosfet is very blackened and burnt-the reverse side of the board is also blackened-will try to enclose a photo.

The existing mosfets are numbered P75NF75 and below that CZ01H-I have some spare mosfets here IRF740-are these a substitute?
Thanks Peter

Can't see the blown mosefet in your picture, but I do see the board looks a bit burnt. I'm not surprised. How many mosfets are there total? It's possible the others are also internally damaged.

I'm no electrical engineer, but I believe it would be best to use mosfets that have the same or very similar specs to the original. I guess the most important thing would be the package size and the RDS(on) mili ohms resistance on the mosfet and matching that as closely as possible within reason to the other mosfets. Doing a quick search it looks like the P75NF75 has an RDS(on) of 11 mili ohms and an 80A max current. But the IRF740 is 550 mili ohms and 10A max current. So those are very different parts.

Looks like on ebay about 10 P75NF75 can be had for $5 shipped or less.

 

[Peter Brigg]

I have ordered 15 of the original mosfets on eBay-worth a try to replace some of them-one of then for sure needs replacing however others have also discoloured the board-the photo I have included already is of the back of the board showing where the overheated mosfet was.
I will now include a photo of the blown mosfet- photos taken with a mobile phone do the quality is not that good.

Does anyone have any hints re replacing the suspect mosfets? I can do basic soldering. The mosfet in the new picture has melted parts-plastic blown off one side and so on.

I have also looked at new BMS boards however replacing the mosfets seems the simplest option at this stage.
Thanks for your help.

Peter

 

[redilast]

As for removing the mosfet. Do you have a hot air rework station?

If not you could try a heatgun. But you may want to try to shield other components on the PCB from the heat of the heatgun. I've used kapton tape for this in the past. Heat up the part real good then lift it off with tweezers. But I've only done that with smaller IC's and resistors and stuff. Assuming the PCB has a decent thermally conductive layer it may absorb the heat pretty well. Might need to use a high value soldering iron (100W+) and touch several sides near the edge of the mosfet over and over again until you get the solder completely heated up underneath the mosfet so you can lift it off. Probably start though with removing the legs from the mosfet so the rest of the package is easier to remove.

 

[fechter]

I use a desoldering station, but if you don't have one, it may be easier to cut the legs off the bad FETs, leaving enough to grab onto with a needlenose or tweezers. Then heat the leg while applying some tension until it pulls out. You have to be careful not to lift the traces off the board when doing this. Be sure the solder is fully melted before pulling and don't pull too hard. It helps to heat the leg from the same side you are pulling from.

Once you get the legs out, the hole will be filled with solder. Then you can use Solderwick to clean the hole or do what I call "heat and beat", where you melt the solder in the hole then quickly smack the board against the bench to knock the solder out. Once the holes are clear, the new FETs can go in.

 

[Peter Brigg]

I have a week or two before the mosfets arrive in the post-thanks for those hints. The whole setup is getting older and more prone to failure.
The original plastic case to hold the cells and BMS board is brittle and doesn't close up properly-looks like it will become a plywood box if things get any worse-it is all about 5 years old and has had a lot of use.

Am I correct in thinking that the cells themselves can have a fairly long life? Each one is about 6 inches long -cylindrical and covered with a blue plastic sleeve.

Peter

 

[fechter]

Am I correct in thinking that the cells themselves can have a fairly long life? Each one is about 6 inches long -cylindrical and covered with a blue plastic sleeve.

Peter

Sound like Headway cells. They are supposed to have good calendar life. I haven't seen too many aging reports on these. I have some abused A123 LiFePO4 cells that are over 10 years old and seem to be still OK. One thing that will damage them is to let the cells get below 2v or so. If it goes to zero volts, it might not survive.

 

[Peter Brigg]

I measured the voltage of each cell-all around 2.8V-so that is OK.

I also removed the most damaged mosfet (going on appearance-the most charred burnt mosfet).
A photo of the board (after the mosfet was removed) is attached-there appears to be a thin piece of aluminium under this mosfet -partly burnt away by the heat.

Is this piece of aluminium just a type of heatsink or does it have the job of electrically joining all the bodies of the mosfet?

The board appears to have all it's tracks intact so perhaps just replacing one or two mosfets will fix the problem.

The removed mosfet was cracked and pieces of plastic broken away from it.

The photo is a bit blurred.
Peter

 

[redilast]

I measured the voltage of each cell-all around 2.8V-so that is OK.

I also removed the most damaged mosfet (going on appearance-the most charred burnt mosfet).
A photo of the board (after the mosfet was removed) is attached-there appears to be a thin piece of aluminium under this mosfet -partly burnt away by the heat.

Is this piece of aluminium just a type of heatsink or does it have the job of electrically joining all the bodies of the mosfet?

The board appears to have all it's tracks intact so perhaps just replacing one or two mosfets will fix the problem.

The removed mosfet was cracked and pieces of plastic broken away from it.

The photo is a bit blurred.
Peter

Thats one fried mosfet!

What piece of aluminum, I don't see one in the pictures? But if its aluminum its more than likely just to act as a heatsink for the mosfets as they can get quite hot under heavy load. Mosfets shouldn't have any aluminum electrical connections. The legs are were the electrical connections are. Also after you remove the bad mosfet I'd clean the area around it with some alcohol and a q-tip or something...

 

[fechter]

Can't tell from the picture, but in most cases, the aluminum is mainly a heat sink and there are coppper traces on the board that make the connections. The aluminum will be in parallel with the copper, so also acts as a conductor.

What I would worry about is the gate driver for the toasted FET. Many times when a FET blows, it shorts all 3 legs and the gate driver can be damaged. After removing the FET, you could try connecting the pack and measure the voltage across the two outside pins where the FET went. If the driver is healthy, you should see something between 8-15v or so. Outside this range may indicate other problems.

 

[Peter Brigg]

I should have said that the piece of thin foil like metal under the mosfet that has been partly burnt up-on closer inspection it was copper not aluminium. I sandpapered a small piece of it and it had that coppery look.

When the mosfets arrive I will try replacing them however it is looking like there is every chance that the board is beyond repair-most of the other components on the board are very small (about the size of a grain of rice) and would be difficult to replace. The mosfets are the largest components on the board by far- One copper track has burnt and lifted however I can bridge that.

Thanks for all the help.

Peter

 

[Peter Brigg]

The mosfets are here however when I soldered the burnt one in the middle leg wouldn't solder-nothing there to solder to as the entire copper part of the board was powdered under the solder exterior-burnt out.

Taken into account with another burnt mosfet and a track damaged I have decided to buy a new BMS board-certainly cheaper than a new battery pack and all the cells read approc 2.8v so are probably OK.

On eBay these BMS units look OK 221172139610 and 321048904700.

One comes with clear wiring instructions.

Any comments -any other way to go?

Thanks
Peter

 

[Peter Brigg]

Something more noticed-the 2 eBay items are from the same supplier.

My original BMs has 6 mosfets.
From the photo the eBay BMS seems to have 8 mosfets -perhaps they are not all mosfets-they say it is 36v so no doubt it is.

For the money I think it is the way to go
peter

 

[fechter]

The mosfets are here however when I soldered the burnt one in the middle leg wouldn't solder-nothing there to solder to as the entire copper part of the board was powdered under the solder exterior-burnt out.

Taken into account with another burnt mosfet and a track damaged I have decided to buy a new BMS board-certainly cheaper than a new battery pack and all the cells read approc 2.8v so are probably OK.

On eBay these BMS units look OK 221172139610 and 321048904700.

One comes with clear wiring instructions.

Any comments -any other way to go?

Thanks
Peter

Those look OK. Probably the easiest route at this point.

On your fried board, the middle leg on a FET is also attached to the heat sink tab, so many times there is a connection via the tab and you don't need the middle leg. Alternately, just place a jumper wire from the leg with the bad trace to one of the parallel FETs.

 

[Peter Brigg]

The BMS board arrived from China-about $60 Australian and after reading the instruction a few times I wired it in and it works perfectly. Really pleasing.
Having said that I wired it up incorrectly at first and thought it was cooked-not so-a really robust bit of electronics.

It has given the battery pack a whole new lease of life for a modest outlay.

The next issue is the plastic housing enclosing it all. It has become brittle and the assembling screws no longer hold the lid to the base. Knowing how it all works will make it easier to put in a new box.

Thanks for the help-a good outcome

Regards
Peter

 

[cycleops612]

A moral here in the lifepo4 debate.

How much simpler it is to work with say8 x 20amp pouches in series for 20ah 24v, than god knows how many 2500ma 18650s in series and parallel.

Worse, hunting down a failed cell.

 

[Peter Brigg]

You will recall that I mentioned earlier that I have two 36v elec bikes. I bought one BMS recently and installed it-now the second bike appears to need a new BMS also. They are getting a bit old now(like me!) so it is not unexpected.

On the first occasion a stray piece of wire entered the charging port and melted the port and ruined the BMS.

On this occasion there was no dramatic end to the BMs-it just doesn't work.

Here are the tests I have done.

On the battery outlet I placed a 240v light bulb (approx 100w) on the + and - terminals-the bulb didn't light even tho with a multimeter you get a reading of approx 38v.
I then put the light bulb over the + and - terminals of the battery pack before they entered the BMS and the bulb glowed a dull red for some time-about 3 minutes-I could have gone longer however this test indicates to me that the battery pack had power and the BMS is not passing it on. The voltage over the terminals before they enter the BMS is approx 41v

So the BMS is faulty?

Is may reasoning correct here?

Thanks
Peter

 

[Peter Brigg]

Something more on this 36v electric bike battery-I have 2 of these bikes both with 36v LiFePo4 batteries-they are fairly problem free however things happen.

Recently the battery cuts out after using the bike for a short time-eg a 1 km ride. I changed the battery over to the other bike and this confirms the battery is the bother.
After taking the cover off I put a 2amp load on each cell and after a minute or so I measured each cell- the readings are-

First 6 cells 3.4v next 3 cells 3,6v and the final cell 3.9v (all readings taken under a 2A load)
It seems to me that the BMS is faulty-am I correct here? If a cell was dodgy it would read low?

Also looking up the Chinese parts supplier I see some BMS units are sold with Bluetooth. What is the idea here? How is Bluetooth useful with a BMS?

Regards
peter

 

[amberwolf]

How is Bluetooth useful with a BMS?

There's some threads and posts about them around ES; maybe they give enough info about the features?
https://endless-sphere.com/forums/viewtopic.php?f=14&t=88676&hilit=Bluetooth+BMS

https://endless-sphere.com/forums/search.php?keywords=+Bluetooth++BMS&terms=all&author=&sc=1&sf=all&sk=t&sd=d&sr=posts&st=0&ch=300&t=0&submit=Search