It needs quite a firm push on the pedals to bring the power in. That means I am doing a lot of the work, which is not my idea of an electric bike. ChatGPT says this is a common issue with these bikes. Is there anything I can do to reduce this? Change a spring in the torque sensor in the motor perhaps?

Possibly change to a cadence sensor ? But then I suspect you would need to change the controller and battery. Sell it and buy something more in keeping with your requirements ? What sort of mileage and terrain and how heavy are you ?

 

There's plenty of cheaper ebikes that have cadence sensors where the power comes in when you turn the pedals , irrespective of how hard you pedal

Edited October 13, 2025Oct 13 by Peter.Bridge

I've just found a review of a slightly earlier model which says it has a torque sensor attached to the chainring. Having checked mine, there is a wire in that area. Assuming it's not a cadence sensor on mine, I might be able to fiddle with something on that.

Are there multiple levels of assistance? Does the effect affect them all?

https://www.facebook.com/share/1Exergp9hX/

 

https://www.facebook.com/share/1GsaiW6HZi/

Are there multiple levels of assistance? Does the effect affect them all?

 

There is No/Eco/Tour/ Climb/Sport. It's hard in all modes, but least hard in Climb.

 

I've fiddled with pressing the pedals with the brake on. I can see that a disk attached to the pedal and the crank rotates a small number of degrees independently of the chainring. So I'm guessing there are springs between the two inside the chainring as I've had one like that before. I couldn't find an image, but it had the same sort of arrangement as in this car clutch plate.

 

 

Hopefully I'll be able to remove or replace some of them.

Edited October 13, 2025Oct 13 by ChrisOfBristol

Assuming that you have the old type chainwheel because you didn't say how old it its, do what I describe below. This one has the actual sensor screwed to the back of the chainwheel and it has a wire coming out of it. If your wire doesn't come from there, you have the new type for which there is no remedy.

 

Remove and dissassemble the chainwheel. Inside are 5 springs, just like that picture of the clutch. You can remove two to operate the motor with less effort. One end of the spring has a spiggot that engages with the plate, the other is flat and held in with friction. Lever out the free end to get them out.

 

I stripped one down and showed it in this thread:

https://www.pedelecs.co.uk/forum/threads/carrera-crossfire-vulcan-torque-sensor.30490/

Edited October 13, 2025Oct 13 by saneagle

Assuming that you have the old type

I don't know its age or exact model number and would be interested to know if anyone does. It must be the old type, as it has the type of torque sensor you describe. It is complicated and seems rather over-engineered, it had 5 strong springs and I removed 3. That was quite a struggle and I had to bend back some of the sensor tabs which had got in the way of my pliers. It's a pity they chose to include 5 springs as it's not possible to remove any and leave a symmetrical pattern.

I've just done a short test ride and the pedal force seems to be just about right. I'll take it on a longer ride tomorrow to confirm it.

Edited October 14, 2025Oct 14 by ChrisOfBristol

I think it's a Carrera Crossfire-E (Generation 1) 2017-2020. The display is marked HESC SR Suntour.

I've just done a short test ride and the pedal force seems to be just about right. I'll take it on a longer ride tomorrow to confirm it.

 

fantastic !

I don't know its age or exact model number and would be interested to know if anyone does. It must be the old type, as it has the type of torque sensor you describe. It is complicated and seems rather over-engineered, it had 5 strong springs and I removed 3. That was quite a struggle and I had to bend back some of the sensor tabs which had got in the way of my pliers. It's a pity they chose to include 5 springs as it's not possible to remove any and leave a symmetrical pattern.

I've just done a short test ride and the pedal force seems to be just about right. I'll take it on a longer ride tomorrow to confirm it.

There are two 'variables' in the response of a torque sensor, like a straight line graph plot: the gradient, which indicates the multiplication factor applied by the controller algorithm, and the y-axis intercept, which is the start-up torque the rider has to put in.

 

Your original description suggested a problem with the latter. The fix addresses the former directly, and probably the latter indirectly.

 

If if rides as you want now, all good, but bear in mind the two things mentioned above if not. You should find that the lower assist levels now behave like the higher ones did before, and the higher ones are even stronger, at least until max motor torque is being requested.

 

A side effect will be reduced range.

Inside are 5 springs

Now only 2! I've done a long ride on it and it's perfect. It just needs a touch on the pedals to bring the power in. Thanks for your help.

the actual sensor screwed to the back of the chainwheel and it has a wire coming out of it[/url]

(Image from the thread you linked to)

The saga continues! Further riding has revealed that the sensor is badly designed. I went out on a rainy day and the bike stopped with error 21 Torque Sensor electrical problem - I was impressed with the straightforward and accurate fault finding. I took it off and could see that the cable outer sleeve was not secured to the housing. It just has a cable tie to stop it being pulled out, but that doesn't stop it waggling around. That means that the weak point is the 4 thin inner leads.

The next day after straightening out the inner leads and perhaps the circuit had dried out, it was working again. I shall probably cut the wire off and solder it to the circuit, then strip off the remains of the potting compound, epoxy the cable into the well in the corner, then fill the case with silicone sealant (non-acidic advised by ChatGPT).

Other have criticised the two tabs on the casing which locate either side of a metal tab on the disk on the back of the chainring. I agree, they are inadequate because they don't just have to locate the sensor, they also have to stop that disk rotating. Also, because they are only a loose fit around the tab, the sensor waggles about, stressing the wires.

I assumed the "ATS Torque Sensor" in your thread was an upgraded one, but the link to it is broken and I haven't found any trace of it so far. ChatGPT gives me the impression there was no compatible upgraded version, although there was a six-pin version (mine has 4). It probably wouldn't help anyway as the problems are mechanical and it looks the same.

Edited October 22, 2025Oct 22 by ChrisOfBristol

[ATTACH alt=Zback.jpg]64563[/ATTACH]

(Image from the thread you linked to)

The saga continues! Further riding has revealed that the sensor is badly designed. I went out on a rainy day and the bike stopped with error 21 Torque Sensor electrical problem - I was impressed with the straightforward and accurate fault finding. I took it off and could see that the cable outer sleeve was not secured to the housing. It just has a cable tie to stop it being pulled out, but that doesn't stop it waggling around. That means that the weak point is the 4 thin inner leads.

The next day after straightening out the inner leads and perhaps the circuit had dried out, it was working again. I shall probably cut the wire off and solder it to the circuit, then strip off the remains of the potting compound, epoxy the cable into the well in the corner, then fill the case with silicone sealant (non-acidic advised by ChatGPT).

Other have criticised the two tabs on the casing which locate either side of a metal tab on the disk on the back of the chainring. I agree, they are inadequate because they don't just have to locate the sensor, they also have to stop that disk rotating. Also, because they are only a loose fit around the tab, the sensor waggles about, stressing the wires.

I assumed the "ATS Torque Sensor" in your thread was an upgraded one, but the link to it is broken and I haven't found any trace of it so far. ChatGPT gives me the impression there was no compatible upgraded version, although there was a six-pin version (mine has 4). It probably wouldn't help anyway as the problems are mechanical and it looks the same.

It's a bad design, and it gives all sorts of problems. The most common one is that the inner part of the chainwheel sticks to the outer due to crud or corrosion, then they go round together when you pedal, it breaks off the plastic lugs from the sensor, then, as it goes round further, it pulls the wires out. Also, the optical sensors inside get dirty after the seal between the two parts perishes, so you have to dismantle the sensor to clean it.

It's a bad design

Agreed. 2 of the 3 other e-bikes I've owned had better methods, even my first one, a Chinese one with a lead-acid battery that weighed a ton, partly due to the batteries and partly due to the bike being designed like a battleship. It had a Hall-effect sensor which screwed into the motor hub, although that obviously only sensed cadence not torque.

The one that wasn't had a sensor inside the bottom bracket which failed.

I'll also use a cable tie to stop the sensor wobbling about.