Recurring problem with 48V 12Ah Li-Ion batteries

[spikeanded]

Last year I bought a 1kW Magic Pie kit from the GM  UK distributor. I have since been through 3 batteries under warranty. I am just about to take delivery of the 4th (!) and am keen to avoid a recurrence. It appears that the internal bms shuts the battery down after a few weeks or months use, and without warning. I'm not criticising the UK dealer because they have been very helpful and have always replaced the battrey under warranty, but it is very costly an incovenient for all invoved. They don't seem to get much response from GM in resolving the issue properly.

The sympton is that, with no prior warning, after 8 - 20 weeks of use, the battery will be dead one morning, shut down by the bms. Battery voltage is 52V open circuit dropping to 0v when connected to the rest of the system. We have tried different motors and controllers, always with the same result.

The bike is used 5 days a week always for the same 12 mile hilly journey and charged daily. There is a 10 mile train journey invloved too. Occasionally the bike will topple over in the guard's van of the train, but never causing any external damage.

So I thought I'd just ask if you or any other GM/Magic Pie have experienced similar issues, and what the fix is?

It strikes me that the internal bms is being overly sensitive to the battery voltages, and a bit too keen to shut the battery down. I bet there's plenty of life left in those cells!

Maybe if GM could provide the circuit diagram and advise which components dictate the cut off voltage and what values they could be substituted for, then if this happens again (outside of warranty period of course) I could get it modified, and carry on using the battery. Also,is there the possibility of an internal "reset" (maybe by disconnecting and reconnecting the bms)?

I used to have a Heinzmann NiMh system and that had a mod you could do to make the bms less finicky, recommended by the UK importer, and that was to do with tolerating a less balanced cell pack, which naturally occurs with age with NiMh, so it reminded me a bit of this situation.

[Cornelius]

When I started using my bike in the spring last year (after having the battery stored inside the house for the winter), it worked ok for 2 weeks, then I had the same symptoms as you; measuring full voltage when disconnected and 0V when connected...

With me, it was the key switch that had dirty/bad contacts inside. After cleaning with contact cleaning spray and then some contact protect spray, I haven't had any problem. Others here also have had the same problem.

So I wouldn't be surprised at all if ths was the case for you too...

[spikeanded]

Hi, if only this was the case. Unfortunately the UK distributor checked the key and switch, and it checked out OK. So I guess it's looking like the bms then? Anyone have any further suggestions? Thanks Tony.

[Cornelius]

It's not enough to just check the key switch visually or with/without load; one have to open the end-cap and test if the battery can deliver power to a load before the switch. (Bypass the switch alltogether...) (My key switch managed to deliver just enough power to light the 'tank-meter', but not to run the motor; when twisting the throttle, the tank-meter went dead.)

(And it is not enough to squirt any cleaning agent into the keyhole; one have to apply it on the inside-end of the key switch...)

[spikeanded]

Hi, I'm not sure how thorough the tests were. I didn't want to go poking about myself because it would invalidate the warranty. It's the sort of thing I'd try if the battery was out of warranty though. In my case, the battery does not even light the twistgrip lights. There is a thread regarding the bms here: http://goldenmotor.com/SMF/index.php?topic=3250.0. I have been told by a 3rd party electruic bike dealer in the UK that the key operated battery switch is particularly bad quality (there'll be a voltage drop across it, therefore current drain through it even when new) , so I think there's still some mileage in that theory too. He was suggesting a battery mod, and an external 40 amp 12 mosfet Infinion controller, but again that would invalidate the warranty, but I'd be inclined to try it if out of warranty. He says "you also need a mod on the battery as the main power goes through the key switch. The switch is poorly made and wont handle the 20 amps continuous ". Having said all this, no-one seems to have a bad word to say about the motor itself... only the internal controller, battery and bms!!

[Cornelius]

You're right, I don't think the key-switch are the highest quality either. Not only the voltage drop, but when switching on, there's a split-second, huge current draw, charging up some capacitors inside the controller that's creating a spark across the contacts inside the switch. I think that is the reason for bad connections after a time...

Since I've got the 901-pro kit with external controller, I've modified my battery; connecting both main + and - directly to the connector, and then added a thinner + wire via the key-switch that goes to the Control-On/Off pin on the controller.

That way, the key switch only handles a few mA, and I get to utilitize the Anti-Theft function in the controller. (in my opinion, this is the way the key-switch should have been wired as standard, since it is the only sensible way to use the anti-theft alarm, and save the switch... )

Edit:
I just read the thread you linked to, and he got full power by 'bypassing the bms' as he said; he didn't say how he did it, but i'll bet he'd get the same result bypassing the key switch... (Ok, i'll stop nagging about the switch now, but it's so easy to check, and the most probable fault, so... )

[GM Canada]

Since I've got the 901-pro kit with external controller, I've modified my battery; connecting both main + and - directly to the connector, and then added a thinner + wire via the key-switch that goes to the Control-On/Off pin on the controller.

That way, the key switch only handles a few mA, and I get to utilitize the Anti-Theft function in the controller. (in my opinion, this is the way the key-switch should have been wired as standard, since it is the only sensible way to use the anti-theft alarm, and save the switch... )

This is very intersting. If what you say here actually works for you then you are probably the first person I can recall that actually has the alarm function working. Have you posted this somewhere else in the forum for others to try.

Gary

[Bikemad]

The alarm function did work with some of the earlier external controllers, but I don't think it works with later versions of the cruise controller.

Alan
 

[Cornelius]

Gary, yes I did: http://goldenmotor.com/SMF/index.php?topic=1943.msg10052#msg10052

Alan: I haven't seen the later external controllers with the wiring harness (mine are from 2009), but if you have one, and there still are that thin red wire from B+ from the battery to the pin marked S1 on the 12pin plug on the controller; try disconnecting that thin wire, and press the cruise controller button for a couple of seconds  within 10 sec after disconnecting that wire... (With full power from the battery, ofcourse. )

[MonkeyMagic]

You can tell if the PCB has the high voltage transformer, it puts out around 15VAC

All internal controllers do not have this hence no alarm function. I believe Who42 posted a recent picture of the external controller, and I remember it had the high voltage tranformer in there (red square thing top centre of PCB)

Karmabike uses it still, but he complains (amongst everything else!!) that over time, the high voltage transformer overheated and died. I think perhaps he liked to hold down his horn button, so as an alarm only, the feature should last a long time.

The external controller Chris my mate bought end of last year, does have the alarm enabled so if they have changed it, it must have been recently unless he was just lucky !

[GM Canada]

Gary, yes I did: http://goldenmotor.com/SMF/index.php?topic=1943.msg10052#msg10052

Alan: I haven't seen the later external controllers with the wiring harness (mine are from 2009), but if you have one, and there still are that thin red wire from B+ from the battery to the pin marked S1 on the 12pin plug on the controller; try disconnecting that thin wire, and press the cruise controller button for a couple of seconds  within 10 sec after disconnecting that wire... (With full power from the battery, ofcourse. )

I'm definitely going to give this a shot. I just came accross a 48 v controller from before they were USB programmable. I briefly bench tested it and it still works. I think I'll swap it for the external on my current ride to see if I can get the alarm to work  Maybe I should test this with both a new controller and the old one I just mentioned to see if one, both or none work.

Gary

[GM Canada]

no-one seems to have a bad word to say about the motor itself... only the internal controller, battery and bms!!

I'm not sure I can agree with that. If I go by what I replace by warranty I would say of course the internal controller is the most likely item to fail. But things have steadily been getting better with newer versions of the internal MPII controller developed more often than you would think. The failures still may happen but nowhere near as often as before. Any long time forum member could probably think back and agree posts on "the five beeps of death" are a lot less frequent. I have had a few batteries fail from weak cells, but no BMS yet from any of my customers in the last 2 + years.

Gary

[Cornelius]

Gary; do test.

Triggering the alarm are also a good way to locate weak fets in the controller; take a look at this post: http://goldenmotor.com/SMF/index.php?topic=2080.0

I'm still using that 'smaller' fet mentioned in the above thread, and I have triggered the alarm several times, and it's still going strong.

Anyway; connecting the wires so the alarm functions will at least save the key-switch for much wear, and eliminates one point of voltage drop and possible trouble in the long run.

[spikeanded]

Hi, well my new replacement battery arrived 2 days ago - and guess what. It has a faulty switch!

I verified last night and this morning that the bike is useable as it is in the short term. It's a pain, because if I go  over a bump, the battery cuts out and I have to jiggle the key to get it working again.

I am concerned about invalidating the warranty, but I have suggested to the UK distributor thatthe switch is electrically bypased internally so it is always "on". (as has already been suggested). I will then have to cut the battery power by external means (removing the battery connector and replacing to cut/restore power). But to avoid wear to the battery connector terminals, I have suggested splicing in a similar extra pair of 50 amp high voltage Anderson connectors (with waterproof boot) in the battery cable as a way of immobilising the bike.

A rugged external connector like the Anderson is up to the job (marine quality, so it should be), and I'll put up with the inconvenience of having to make or break a physical connection every time I use the bike. I haven't found a suitable high current inline switch, but obviously that would be even better.

This way, the recurring problem will likely be resolved. If it isn't and we have another failure it's likely to be the bms. I think the least likely option is the cells.

Thinking back, my Heinzmann had a very rugged permanent battery connection (looked a bit like a cigar lighter), and the key operated into the controller (not the high voltage/ high current battery path) to immobilise the bike. What a shame the Golden motor system isn't like this.

What do you all think?

I'd be interested to hear what Golden motor think, as it seemsthey genuinely have a quality control issue with the switch?


Thanks


Tony

[arclight]

Hi guy"s I had some switch issues with the two new batteries I got with the Pie 2 48v's, and I solved my issue's, with both of the 48"s. key switch's cutting out,over minor bumps and having to hold the key up to make it run!
Lock the battery in the rack, grab it by the handle and give it a quick jerk backwards, don't break the handle! seems to 'snap' into the correct position! don't know why but it works for me, could it  be a ground issue?  could be preasure on the Anderson connector, because of the style of post rack I have it on(broken 3) changing to conventional rear axle suport rack today!  hope this is usefull. s