Torque sensors

[mwburns]

Hi
The w013 and e010 fault indicates torque sensor. I’ve been reverse engineering the torque sensor and main board in the e7000. I’m not getting far though

Is it still in 2026 accepted wisdom that these errors mean the motor is toast?

I have had success with one torque sensor failure and that was due to excess grease between the floating induction coil and its plastic cover.

Thanks

 

[cream]

Hi
The w013 and e010 fault indicates torque sensor. I’ve been reverse engineering the torque sensor and main board in the e7000. I’m not getting far though

Is it still in 2026 accepted wisdom that these errors mean the motor is toast?

I have had success with one torque sensor failure and that was due to excess grease between the floating induction coil and its plastic cover.

Thanks

Next to impossible to repair a faulty torque sensing element on a Shimano motor. That was a very complicated design in the first place. Imagine the sensors themselves have a built in microprocessor along with the sensing elment. They communicate digitally with the moving multilayered PCb which in turn relays the infomation to the stationary PCB through magnetic coupling (transformer) then to the mainboard microprocessor involved in motor control. I’ve tried myself to cure this fault but didn’t get anywhere, back then. I never got the exact part number for the sensing element, but I did find out it was manufactured by Alps, probably a custom part for Shimano. Even If I could get that part there are no guarantees that it will work, without calibration that only Shimano could do at the factory.

For me, having to repair Shimano motors brought good feelings and bad feelings. Their electronics are trully special. They are so complex/miniaturised yet the user experience is so simple, E8000/7000 screen had perfect ergonomics for emtb scenario. Cables, communication between nodes, whole different level than the rest (at that time). Even the power button had a mini pcb inside, with a microcontroller that could be updated OTA over bluetooth. All that to turn on the bike
Worst disappointment is grease choice which will breakdown chemically (under mechanical stress or temperature) causing irreparable damage to gear splines.

P.S.

When talking about torque sensor implementation, got to hand it to Bosch, as their implementation is way more simple and reliable.

 

[ncs]

Anything one man design another man can repair, main problem is available of details, datasheets, FW, programs, IC. When unit not detailed - it only means manufacture greedy.

 

[mwburns]

Next to impossible to repair a faulty torque sensing element on a Shimano motor. That was a very complicated design in the first place. Imagine the sensors themselves have a built in microprocessor along with the sensing elment. They communicate digitally with the moving multilayered PCb which in turn relays the infomation to the stationary PCB through magnetic coupling (transformer) then to the mainboard microprocessor involved in motor control. I’ve tried myself to cure this fault but didn’t get anywhere, back then. I never got the exact part number for the sensing element, but I did find out it was manufactured by Alps, probably a custom part for Shimano. Even If I could get that part there are no guarantees that it will work, without calibration that only Shimano could do at the factory.

For me, having to repair Shimano motors brought good feelings and bad feelings. Their electronics are trully special. They are so complex/miniaturised yet the user experience is so simple, E8000/7000 screen had perfect ergonomics for emtb scenario. Cables, communication between nodes, whole different level than the rest (at that time). Even the power button had a mini pcb inside, with a microcontroller that could be updated OTA over bluetooth. All that to turn on the bike
Worst disappointment is grease choice which will breakdown chemically (under mechanical stress or temperature) causing irreparable damage to gear splines.

P.S.

When talking about torque sensor implementation, got to hand it to Bosch, as their implementation is way more simple and reliable.

Thanks for the detail.

I have two motors here, both e7000.

One [1] had a E010 and W013, which i cured by cleaning excess grease off the wireless transmission. Except since i've put it together, I get zero assist - the BLDC does not turn. No errors, no assist. I'm thinking i've fried a mosfet. Has anyone seen this condition before? No errors, no codes, no motor-turny.

The other [2] E7000 I purchased with a W013 + E010. I momentarily managed to fix that by replacing the non-rotating wired PCB from the working E7000 [1]. It turned on and no error for a few minutes, and now I just get a e010. No W013. My thinking was that the pairing code/calibration data (if any exists - im doubtful) is in the fixed PCB whereas the floating PCB is likely version-paired to the bonded torque sensor. Which, incidentally, i've found the manufacturer for and it is not a microcontroller. All the microcontrol takes place on the floating board. It requires a stable voltage. The pairing data - or more likely version dats - is held in two x4 pin IC's.

Curious to know - did you try that sort of mix n match testing of the torque sensors?

 

[mwburns]

Update. I’ve managed to transplant a torque sensor into another motor so that is something. Basically used one half of the torque sensor circuit and it worked.

 

[cream]

Curious to know - did you try that sort of mix n match testing of the torque sensors?

Almost every time, the rotating torque sensor part gets damaged (not the PCB, but the sensors themselves epoxied to the metal piece. You can mix stationary and rotary parts, as long as they are not damaged, they should work. Internal calibration is on the sensor itself (as it also has a digital processor under that epoxy).

Sometimes the faults are intermittent, you think you clean them and they work but a couple of miles later, errors.

Of course, there are cases when one of the PCB is damaged but it's quite rare and most likely because of water ingress.

 

[mwburns]

Thanks. For some I reason I thought mix n match was a no go.

Pretty sure the torque sensor under the blob is a melexis but happy to be corrected

Now that I’ve one working motor I’ll dig into the torque sensor more. Shits and giggles etc