Jorvick trike electrical problems (inrush?)

[Clivington]

Hello all! I'm new to this forum and totally new to electric bikes so please bear with me. I'm a retired electronics engineer who's into cycling. I offered to help a friend with her Jorvick trike which she uses because she's unable to walk without support. I don't know the model or year but I suspect it's maybe five+ years old. Mechanically it's pretty horrible and electrically it seems to use generic Chinese components. She tells me that it's had a new motor and I'm guessing that the display and controller will have been changed at the same time. The battery is original and everything went fine for a few months after the replacements. The fault now is that the thing doesn't turn on reliably. Sometimes it does, and it works every time, so it's difficult to demonstrate the fault. Other times it constantly fails. It doesn't give her any confidence to use the trike because if it fails (eg when going shopping alone) then she's stuck.

So... I inspected the battery connections and all electrical connections. They all seem sound. The battery voltage off load seems reasonable at around 41-42V after charging (it's a 36V battery). I tested the interlock switch on the battery off the trike, first with a 10k resistor and then with a filled mains kettle as a load (about 28 Ohms). The switch is fine. After putting the battery on the bike and disconnecting the power cable I noticed a crackle and flash when reconnecting it. That made me wonder if it was an inrush problem. I'm presuming the controller must have a large capacitor on the supply side. To test the inrush theory I made an inductor from a coiled 30m 3-core mains cable so effectively a coil of 90m wire between the battery and the controller. It might be the inductance or it might be the resistance but that worked! Every time I plug the battery in now with the coil connected the trike powers up. I reckon I have an inrush problem that is sometimes triggering the BMS in the battery to shut down. Is this likely to be because the cells are decaying? (The trike has never been heavily used.) Or could the BMS be tired? Or should there be an inrush limiter in the circuit? I haven't found any commercial inrush components for use with e-bikes so I have some 5 Ohm 20mm thermistors on order. They can each handle 7A continuous, so I reckon that three in parallel should be good enough for the likely motor loads, and I'm hoping the resistance (1.67 Ohms) will be enough to reduce the inrush at turn-on.
Do you think I'm on the right track? I'm trying to avoid my friend having to buy a new battery because of the cost - but maybe it will be inevitable if the existing one is gradually dying. I also don't want her to buy one if it's not that. I don't know anyone locally with a similar battery I could just try for a few minutes. I would welcome the thoughts of someone with experience of these kind of things. Ta.

 

[saneagle]

You're on the wrong track. The only time you get inrush is when you physically connect the battery. After that, the switching is done by the control panel unless you have a separate battery switch and are switching it off.

When the BMS shuts down due to over-current, it resets itself pretty quickly.

The LCD is gets its power directly from the battery, so if it doesn't switch on, theirs a connection issue somewhere.

There is the possibility that water in the LCD could cause that, so worth checking if it comes apart.

Another problem that happens from time to time is people pull cable ties too tight, which breaks the wires inside the insulation and makes them intermittent.

 

[Clivington]

Thanks Saneagle. Yes, you're right, the inrush only occurs when I connect the battery. However, the battery connector has only got two pins and they go directly into the control box down near the bottom bracket. There is no second battery or any provision for one. The LCD panel cannot get power without it coming from the control box. When there's a fault and I measure the voltage across those power wires coming from the battery (going into the control box) it is effectively zero, so the BMS, or something inside the battery box, is latching off. Once latched it doesn't recover unless I turn the battery off and physically disconnect it. There is also a little state-of-battery LED display panel at the top of the battery which shows red very briefly and then nothing confirming this latching. Once I pull the battery box out, that LED display panel shows full charge again.

Unplugging the LCD panel makes no difference: the voltage from the battery is sometimes there and sometimes isn't.

I think that, by not mentioning it, you've confirmed that there is generally no inrush limiter component, so there isn't one missing from this trike.

Maybe you can tell me: when a battery gets old and a number of cells are low voltage, is it a hard fault? Or is there a period when the BMS sometimes turns off and sometimes doesn't? Is there hysteresis in the voltage detection? I'm wondering if this is an early design of BMS and not very sophisticated.

When the battery is working and the trike is being ridden around then it behaves normally. It doesn't hesitate or act intermittently as far as I can tell. The LCD display and buttons don't behave oddly. It's when you stop and turn off that you don't know if it will power up again. I don't suspect any broken wires or water but, if it comes to it, I could open things up and probe around.

Cheers.

 

[saneagle]

OK, you need to explain what you mean by "works sometimes". I assumed you meant that once it's switched on and stays on, you can do a normal 30 mile ride.

 

[Clivington]

I think 30 miles is a reasonable assumption. Once switched on I have just propped the front wheel up and rigged the throttle, and had it run continuously for more than 12hours but admittedly with little load. The battery still had plenty of life left and recharged OK. I wouldn't want to ride it 30miles myself - I'm not used to riding trikes and this thing is clunky. But I've managed to ride it round the block a few times and it displays the speed and changes assistance level etc just fine. It's just that it doesn't reliably power on, but otherwise it seems fine.

 

[saneagle]

Is there any way you can test it with an alternative load, like a heater coil, dump load, 3 headlight bulbs or anything like that?

One other thing to try : Disconnect the motor and see if it's still the same.

 

[Clivington]

Well, like I said at the beginning, I tested it with a kettle full of water. I measured the kettle as 28 Ohms when cold. I thought that was a fairly significant load - I didn't leave the water long enough to boil but it did get too hot to touch (without the water the element would probably have overheated.) I just wired the battery to an old mains socket and plugged the kettle in. I could probably use a double socket and plug a heater in as well as the kettle. Are you thinking that a heavy load might cause the BMS to trip or maybe would reveal tired cells?

 

[Sturmey]

Well, like I said at the beginning, I tested it with a kettle full of water. I measured the kettle as 28 Ohms when cold. I thought that was a fairly significant load - I didn't leave the water long enough to boil but it did get too hot to touch (without the water the element would probably have overheated.) I just wired the battery to an old mains socket and plugged the kettle in. I could probably use a double socket and plug a heater in as well as the kettle. Are you thinking that a heavy load might cause the BMS to trip or maybe would reveal tired cells?

I =E/R = 36/28 = 1.3 amps. Your load is less than 10% of normal maximum. You need more kettles . Incidentally, I have used an electric shower element at 5.5 ohms that draws about 7 amps (36v battery) to load test.

 

[saneagle]

Well, like I said at the beginning, I tested it with a kettle full of water. I measured the kettle as 28 Ohms when cold. I thought that was a fairly significant load - I didn't leave the water long enough to boil but it did get too hot to touch (without the water the element would probably have overheated.) I just wired the battery to an old mains socket and plugged the kettle in. I could probably use a double socket and plug a heater in as well as the kettle. Are you thinking that a heavy load might cause the BMS to trip or maybe would reveal tired cells?

Did it switch on every time like that? That's what I'm trying to establish because if it did, there's probably nothing wrong with the battery, though a test at higher current would help.

 

[Clivington]

Did it switch on every time like that? That's what I'm trying to establish because if it did, there's probably nothing wrong with the battery, though a test at higher current would help.

Yes, it switched on reliably with that load. I'll see if I can repeat the test with a heavier load in case it misbehaves.

 

[Clivington]

I =E/R = 36/28 = 1.3 amps. Your load is less than 10% of normal maximum. You need more kettles . Incidentally, I have used an electric shower element at 5.5 ohms that draws about 7 amps (36v battery) to load test.

Thanks Sturmey - the shower element is a great idea!

 

[guerney]

I =E/R = 36/28 = 1.3 amps. Your load is less than 10% of normal maximum. You need more kettles . Incidentally, I have used an electric shower element at 5.5 ohms that draws about 7 amps (36v battery) to load test.

That's a great idea! I happen to have a 36Ω electric cooker element: 36/36 = 1A. If I had 20 of them, I could load test my 22A BMS at 20A + 2A battery soldered lights = 22A, to see what happens. Would the BMS shut down, reset/restart for some time interval continuously, or will transistors or whatever melt?

A test for a cold winter's day, if I can find a bunch of old cooker/shower/boiler/kettle elements.

 

[Clivington]

I =E/R = 36/28 = 1.3 amps. Your load is less than 10% of normal maximum. You need more kettles . Incidentally, I have used an electric shower element at 5.5 ohms that draws about 7 amps (36v battery) to load test.

A friend knows a plumber and I will shortly have a shower heater element . Did you run it continuously and, if so, how did you stop it overheating? (V^2/R = 236W at 36V or 290W at 40V) Or was it just for a few seconds while you ran some tests, so it didn't overheat?

 

[Clivington]

That's a great idea! I happen to have a 36Ω electric cooker element: 36/36 = 1A. If I had 20 of them, I could load test my 22A BMS at 20A + 2A battery soldered lights = 22A, to see what happens. Would the BMS shut down, reset/restart for some time interval continuously, or will transistors or whatever melt?

A test for a cold winter's day, if I can find a bunch of old cooker/shower/boiler/kettle elements.

36V at 22A is 790W. You can get that from a butchered 3kW convector heater. Just remember to use rivets rather than solder

 

[guerney]

36V at 22A is 790W. You can get that from a butchered 3kW convector heater. Just remember to use rivets rather than solder

It just so happens, I bought one of those totally ineffectual tiny plug-in 500W "ceramic" heaters at the British Heart Foundation charity shop sale going for £2, for this very purpose. Price slashed from £5. It has a very fiddly to use remote control, which it only responds to if you have better sniper skills than that dude who tried to shoot Trump's nose off. I'll keep an eye open for additional resistors...

 

[Sturmey]

A friend knows a plumber and I will shortly have a shower heater element . Did you run it continuously and, if so, how did you stop it overheating? (V^2/R = 236W at 36V or 290W at 40V) Or was it just for a few seconds while you ran some tests, so it didn't overheat?

The shower element is rated a 9 Kw (230v) normally in water but it easily works in air for a few hours at 36v at 250 w or so.. I just leave it down on the fireplace as it does get hot enough to say melt plastic.

 

[Clivington]

Update following some load tests:
Following Sturmey's suggestion, I obtained this 8kW shower element to use as an electrical load. It's actually quite useful because it's two non-identical elements and can be configured in series, parallel, or you can use a single element. It came with microswitches attached - handy for turning things on or off.

Anyway... using a single element gave a load of 12.8 Ohms when hot. Battery no-load voltage was 41.8V. Under load, the battery voltage fell immediately to 40.2V. After 50 minutes it dropped to 36.98V at which point the BMS kicked in and switched off the battery. I removed the load and then reconnected it - the voltage recovered a bit but then started falling and after a few seconds the BMS switched off again at the same level. Reminder: this is a "10Ah" 36V battery and the test load current was round 3A. Not an entirely conclusive test IMO - to determine the true capacity of the battery I should probably use only 0.5A - but I think it indicates the battery is weak.

I can also say that the BMS consistently didn't trip when first applying a 6A load (the two elements in parallel). That level of inrush isn't high enough.

With the battery on the trike, the LCD display wouldn't power up for more than 75% of the times I tried switching on the battery.

It didn't matter whether the motor was connected or not, so the motor isn't causing the inrush - it's in the controller and probably the input capacitor.

I made an inrush limiter from three 5 Ohm thermistors in parallel - with that in series with the battery +ve, the trike powers up every time.

I think I'm going to have to tell my friend that she's going to have to bite the bullet and fork out for a new battery, or have the old one re-celled, but I would welcome thoughts.

 

[saneagle]

I don't get it. If you leave the battery connected to the controller, there is no inrush when you switch on because the capacitors are on the unwatched dive of the controller.

 

[Sturmey]

Anyway... using a single element gave a load of 12.8 Ohms when hot. Battery no-load voltage was 41.8V. Under load, the battery voltage fell immediately to 40.2V. After 50 minutes it dropped to 36.98V at which point the BMS kicked in and switched off the battery. I removed the load and then reconnected it - the voltage recovered a bit but then started falling and after a few seconds the BMS switched off again at the same level. Reminder: this is a "10Ah" 36V battery and the test load current was round 3A. Not an entirely conclusive test IMO - to determine the true capacity of the battery I should probably use only 0.5A - but I think it indicates the battery is weak.

I can also say that the BMS consistently didn't trip when first applying a 6A load (the two elements in parallel). That level of inrush isn't high enough.

My guess is that you have one parallel group of cells that has lost capacity ( or possibly very much out of balance). Its coming up ok on charge but it discharges early and this low voltage of one cell group will cause the BMS to do its job and disconnect the load.
If it were my battery, I would open it up to check all the cell group voltages with a multimeter and visually check the nickle strips and for cell leakage (sweet smell) on the suspected low cell group.

 

[Ghost1951]

My guess is that you have one parallel group of cells that has lost capacity ( or possibly very much out of balance). Its coming up ok on charge but it discharges early and this low voltage of one cell group will cause the BMS to do its job and disconnect the load.
If it were my battery, I would open it up to check all the cell group voltages with a multimeter and visually check the nickle strips and for cell leakage (sweet smell) on the suspected low cell group.

The BMS cutting off power when the pack was at 36volts is a pretty good giveaway that one cell group is much lower than the other cell groups. The pack seems to be unbalanced. Sturmey's suggestion of opening the pack and checking the voltage of all of the cell groups is a good one.

To a lesser extent the exact same thing happened to my battery last week when I was doing a full range riding check in a three and half year old battery. At 43 miles, the LCD was showing 33 to 32 volts under a 250 watt load when the battery suddenly shut off and the assistance stopped.

When I examined the battery at home, one group - a group I have had trouble with before, was under 3 volts - around 2.9 volts as I remember.

The good news is that it may well be possible to balance the pack manually by charging that one group to the level of the others with a 4.2 volt lithium charger. Take it to the same level as the other cells in the battery and then charge the whole pack in the normal way.

An alternative is to use the load to drain the other cell groups to the same level as the weak group and then charge the pack normally.