Gazelle Orange Innergy - intermittent loss of power assist function

[Miktavus]

I have a Gazelle Orange Innegry Electric Bicycle which is inconsistent in its reliability. It's 10 years old, has had moderate use but is in great condition and worth keeping running. It has a Bronze 36v/252Wh battery. Here's the sequence of events leading to loss of the power assist function:

1. Battery fully charged, indicating 100% (5 bars) on controller and all three assist functions work
2. After ~15 to 20 miles battery shows 80% - 100% but power assist stops working.
3. Took battery out and checked voltage with multimeter – 36v reading, as would be expected with a battery showing almost charged on the controller
4. Put battery back in bike and plugged in charger, controller shows 100%. Pressed on button on battery and controller shows power assist working!.
5. Battery fully charged, indicating 100% (5 bars) on controller and all three assist functions work (back to stage 1).

The charger works, there seems to be nothing wrong with the motor or the battery, despite its age, so the controller would seem to be the source of the fault, though its behaviour seems illogical.
Would a replacement generic controller be the answer to this annoying problem? If so, any suggestions? I'm willing to spend some money on it because the bike is in excellent condition and when it works it's great.

Thanks for any suggestions.

Mike

 

[vfr400]

Age isn't necessarily a problem. It's how much you use it that counts.

Anything can happen occasionally, but if the problem is consistent, always happening at around 20 miles, I'd say that it's nearly certain that the battery is knackered.

The meters often calculate on the basis of a long moving average, but an old battery becomes saggy, so it goes down each time you draw a high current from it, and it bounces back when you don't. In that case, the meter could still be showing a high average percentage. The only way to be sure is to put a real-time voltmeter on it to see what happens when it cuts out. If there's a simple jack or other charge socket on the battery, you can connect the meter there, otherwise you have to splice into a connector somewhere.

 

[Miktavus]

Thanks, testing the voltage on load makes sense. I should be able to make up some leads to fix the meter across the charger socket and monitor the voltage as I ride. If that proves the battery is at fault I'll look into replacing the cells.

 

[vfr400]

Thanks, testing the voltage on load makes sense. I should be able to make up some leads to fix the meter across the charger socket and monitor the voltage as I ride. If that proves the battery is at fault I'll look into replacing the cells.

The system will shut down if the voltage drops to about 31V. If that's what you see, there's no option but to get a better battery.

 

[Nealh]

At approx. 36v reading a 36v battery is at about 40% charge, one would think your issue may be more related to battery cell aging being some 10 years old.
Heavy voltage sag or voltage collapse is a sign of aging.

 

[Miktavus]

The system will shut down if the voltage drops to about 31V. If that's what you see, there's no option but to get a better battery.

Thanks, it's useful to have that voltage as a reference point. I've rigged up the voltmeter on the bike and will go for a monitoring ride tomorrow, I'm pretty certain that's what I'll find.

 

[Miktavus]

Further research and observation has pretty conclusively established that the battery is the problem, with its range due to age being about 20 miles maximum. Recelling it is complicated, not guaranteed and not for the fainthearted, and replacement Gazelle batteries are phenomenally expensive. As I'm not that bothered about the battery being a direct replacement in the same place, this dealer https://batteryempire.co.uk/146-batteries-for-ebikes#/ seems to offer quite a few cost effective options, and it's in the UK. I'd have no problems with making connections and devising suitable mountings. Is this the way forward?

 

[Nealh]

If any generic battery can be used then you can use what you like of the same voltage.

 

[vfr400]

Further research and observation has pretty conclusively established that the battery is the problem, with its range due to age being about 20 miles maximum. Recelling it is complicated, not guaranteed and not for the fainthearted, and replacement Gazelle batteries are phenomenally expensive. As I'm not that bothered about the battery being a direct replacement in the same place, this dealer https://batteryempire.co.uk/146-batteries-for-ebikes#/ seems to offer quite a few cost effective options, and it's in the UK. I'd have no problems with making connections and devising suitable mountings. Is this the way forward?

Before buying any battery, you need to confirm that the present one is only connected by only two wires, otherwise you're going to be disappointed.

 

[Miktavus]

There are three outlet pins, the middle one -ve, the others +36V and +9V. What is the function of the lower voltage? Presumably no supply to this would lead to the (unspecified) disappointment. Thanks.

 

[vfr400]

There are three outlet pins, the middle one -ve, the others +36V and +9V. What is the function of the lower voltage? Presumably no supply to this would lead to the (unspecified) disappointment. Thanks.

B ad luck. That wire could be for remote switching or anything else. You need to follow it and see where it goes.

 

[Miktavus]

B ad luck. That wire could be for remote switching or anything else. You need to follow it and see where it goes.

I'll do that. If it has that function it's going to be a low current drain, and I'm thinking that it could be supplied by an independent power source such as a PP3 battery.

 

[Miktavus]

Tracing the wiring was not conclusive but the 9v supply certainly ends up at the hub, where I understand the controller is housed (not behind the headlight as with some earlier Innergy models). A posting elsewhere says of this supply "The third white wire from the battery is a linbus data connection between the battery BMS, the motor and the display.
The normal voltage on it is maximum 12V. The devices pull this voltage to gnd to place data on the bus. (It's a digital serial connection)"

That would imply that it doesn't require a big current, so would the combination of a new rechargeable 36V battery and a separate 9V battery be a solution?

 

[vfr400]

12v is the gate voltage for mosfets. As I said before, I can only guess that it's being used to switch something. I would have guessed that it's switching the battery on, but if you can measure 36v on the other terminals when the battery is disonnected, that seems less likely unless there is some more complicated logic in the switching.

I would also guess that the 12v line is held at 12v by a pull-up resistor rather than a direct supply. If you substituted a 12v source that could provide current to that, you'd blow whatever in the controller that would normally short it to ground. You can test it with a 12v LED, 1k resitor or something loke that with the other side connected to ground. If the voltage colapses, it's coming from a pul-up resistor.

Without a schematic of the bothe the controller and the battery's BMS, it would be difficult to figure out what is switching what.

In the diagram below, the mosfet would be switched on when the switch is open because the leg gets 12v from the pull-up resistor as no current flows through the resistor. When the switch closes, the leg is shorted to ground and The mosfet switchs off. in that state, 1.2 miliamps flow through the resistor from the 12v source. If you had a direct 12v source there from another battery, all you'd do is short it to ground.



On the other hand, the 12v could be a simple power supply that powers the controller's pcb, but in that case, I would have expected 5v rather than 12v.

 

[vfr400]

Looking for a picture of the controller, I came across this article, which details everything:

Renew a Faulty E-bike

I got an old but still working Gazelle Orange Plus Innergy two years ago from second hand. It worked well seamlessly at that time, but I…

otapi.medium.com

That links to this one from our adweb dutch friend, Albert van Dalen , who has been an absolute goldmine of info about how ebikes work.

Gazelle rebuild

Here, I share technical solutions that arised during the development of my solar-bike.

avdweb.nl

 

[obcd]

The later innergy motors with 6 pins connector don't have hall sensors. Some bikes use a direct hub motor and others use an internal gearing (XT and X2 motors) and have a buildin freewheel. The motors with the internal gearing have a very low impedance winding, and most generic controllers don't like it. The direct hub motors have a more normal winding. The third wire between the battery and the controller in the hub is used to communicate with the battery bms. The original controller asks the battery serial number, it's voltage, current, remaining capacity. If the battery isn't responding, it will display an error A4 and D1 on your display. As the controller is busy to long with the battery, the display show a communication timeout as well. Also note that most sensorless controllers expect your motor to rotate so they can determine the rotor position based upon the voltage generated in the windings. The Xt and X2 motors won't rotate when you start riding as you have the freewheel. This could also make it hard for a generic sensorless controller to drive such a motor. The 2 thin wires coming out of the motor windings are used to measure the motor temperature. No idea if it's an ntc or diode or something else.

 

[Miktavus]

The later innergy motors with 6 pins connector don't have hall sensors. Some bikes use a direct hub motor and others use an internal gearing (XT and X2 motors) and have a buildin freewheel. The motors with the internal gearing have a very low impedance winding, and most generic controllers don't like it. The direct hub motors have a more normal winding. The third wire between the battery and the controller in the hub is used to communicate with the battery bms. The original controller asks the battery serial number, it's voltage, current, remaining capacity. If the battery isn't responding, it will display an error A4 and D1 on your display. As the controller is busy to long with the battery, the display show a communication timeout as well. Also note that most sensorless controllers expect your motor to rotate so they can determine the rotor position based upon the voltage generated in the windings. The Xt and X2 motors won't rotate when you start riding as you have the freewheel. This could also make it hard for a generic sensorless controller to drive such a motor. The 2 thin wires coming out of the motor windings are used to measure the motor temperature. No idea if it's an ntc or diode or something else.

Thanks, that's extremely useful information. My motor has no internal gearing, so from what you say it's got a better chance of working with a generic controller, and it does work, though not as smoothly as the original Gazelle set up. The information on the third wire out of the battery explains why I had communication problems when trying an alternative power source. Also the two thin wires in the motor are of no consequence in the current set up so I'll leave them. Overall, with a bit of playing around with the programming l've probably achieved as much as I can, which is a usable power assisted bike.

 

[obcd]

Some of the gazelle innergy models had 2 sensors. One measuring pedal rotation (PAS) and another measuring pedalling force (TMM) Those gave motor assistance proportional to the own force on the pedals. With only a pas sensor, your motor starts to assist as soon as it detects pedal rotation and it assists up to the assist level choosen on it's display. This can make the assistance appear a bit more rough.
If your original controller pcb is still in one piece, I might take it over from you. I have a motor with a controller that has water damage. I might be able to fix it with a controller pcb.