New "TSDZ2 Torque Sensor Central Motor"

Hello...first time post; apologies in advance if this has been covered...my search lead some posts back in May that discussed torque sensor calibration and helped a little but what I am seeking to understand before I make any changes is the process for application of weights and how it applies to the calc sheet vs what you enter into the torque sensor calibration (I've got the 860C with firmware v1.1).

The calc sheet weight increments more or less match my current default firmware settings: 0-5-10-15-19-43-57 -110.

My current understanding is that while at the Technical configuration menu and observing the torque sensor ADC value in real time, I apply a 5kg weight to the pedal and then observe both the weight and ADC reading. In my case this read as 10kg and 182 ADC. Other readings were also off (10kg read as 16kg, 15 as 21 etc)

Am I supposed to change the weight reading in calc sheet from 5kg to 10kg and the ADC reading? Or do I leave the weight reading at 5 and just change the ADC reading? Same for question for actually inputting info into 860C.

Another way it could be done I guess is to apply weight to pedal until the 860C shows 5kg (10-15-19, etc) and then record the actual ADC reading. But given that it is supposed to respond to an actual weight this doesn't seem the correct approach.

It would be great if there was a video showing how to do this (did a search for that too on YouTube).

Anyway, sorry if I am asking such a basic question and your patience is appreciated.
 
CKBuijs said:
....

My current understanding is that while at the Technical configuration menu and observing the torque sensor ADC value in real time, I apply a 5kg weight to the pedal and then observe both the weight and ADC reading. In my case this read as 10kg and 182 ADC. ...

... do I leave the weight reading at 5 and just change the ADC reading? ...
....
If you haven't calibrate before, the adc readings are not correct for the given weight.
Every torque sensor give different adc values for a certain weight.
So note the adc value for a real weight of 5kg, 10kg etc.
After this insert these measured adc values for that weight in the technical configuration menu.
See also the wiki
 
Elinx said:
CKBuijs said:
....

My current understanding is that while at the Technical configuration menu and observing the torque sensor ADC value in real time, I apply a 5kg weight to the pedal and then observe both the weight and ADC reading. In my case this read as 10kg and 182 ADC. ...

... do I leave the weight reading at 5 and just change the ADC reading? ...
....
If you haven't calibrate before, the adc readings are not correct for the given weight.
Every torque sensor give different adc values for a certain weight.
So note the adc value for a real weight of 5kg, 10kg etc.
After this insert these measured adc values for that weight in the technical configuration menu.
See also the wiki

Thanks! I did read the WIKI but I am thick (slow to understand)...my confusion was because when I added a real weight of 5kg, 10kg etc. the sensor reading shows a different weight.

So I will take the ADC readings with real weights, change only the ADC readings and leave the weights on calc sheet and on the display as they are (except for my weight in position 8(0-5-10-15-19-43-57 -X)). This makes sense to me.
 
My controller board has burnt out. It was right at the start of a ride (20 seconds in) so it would not have been hot at the time it failed, but could be the result of prior heat effect. I've done extensive heat dissipation mods and I've been running 600-700w continuously on rides, however this setup has probably only run about 2-3 hours in its whole life.
20211005_152956.jpg

Unless I've just been unlucky and this was a manufacturing fault, I'm thinking that perhaps it's due to the higher case temperature that comes from the heat mods. When the motor was stock it would hit temperature limit and yet the case was not even noticeably warm. With all the mods to move heat into the case my motor temps don't go above 70degC but the case gets quite warm to the touch. And it is the internal case temperature that sets the environment the controller board has to operate in.

With a replacement board on the way, I'm pondering mounting it in a separate box not far from the motor and extending the wiring out of the motor to the new box. I expect there would not be any electrical issues with the motor phases and hall effect signals, since other types of controllers pass these through quite long cables. I'm uncertain about the torque sensor cabling and how sensitive it might be to noise induced by longer wiring, since I don't know the operating principle of the torque sensor. I would just use best practice and use shielded cable for the sensor signals to try and avoid coupling noise from the motor cabling.

Is anyone aware of anyone doing this before?

Also, when I sent the above photo to the vendor I bought from, he said that he'd only really seen this type of failure occur due to OSF problems, though mainly in the past with early versions. Is this true? Is there any recent history of software failures causing hardware failure?
Cheers, Mark.
 
Which OSF version have you used? There has been controllers burning with OSF for stock displays version 20.1.something. I don't remember which version specifically, but I think the problem has not been definetely fixed yet. Check this thread for more information.
 
ilu said:
Which OSF version have you used? There has been controllers burning with OSF for stock displays version 20.1.something. I don't remember which version specifically, but I think the problem has not been definetely fixed yet. Check this thread for more information.
Was using v1.1, being the version that eco-cycles were loading on their osf kit for quite a while. Their recommendation was to upgrade to v20.1 but I think I'll wait a bit before doing that.
 
I'm using V20.1.1C.3 with a 860C display and so far I have not experienced any problems.
Is there only a controller problem when using this version with the VLDC5 display?
 
bpratl said:
...
Is there only a controller problem when using this version with the VLDC5 display?
Yes, only the OSF version for stock display have sometimes a problem with the latest versions.

For some unclear reason there is a chance that one of the Fets wil burn if suddenly the power shuts down when the motor is running. This could happen when you switch off power, blow a fuse or accelerate with a low Voltage battery.
Almost all the OSF code is the same, so problaby it could be inside the old conversioncode for stock display.
 
Elinx said:
bpratl said:
...
Is there only a controller problem when using this version with the VLDC5 display?
Yes, only the OSF version for stock display have sometimes a problem with the latest versions.

For some unclear reason there is a chance that one of the Fets wil burn if suddenly the power shuts down when the motor is running. This could happen when you switch off power, blow a fuse or acceleratie with a low Voltage battery.
Almost all the OSF code is the same, so problaby it could be inside the old conversioncode for stock display.
Thanks, that's good to know because I would hate to change versions. I'm very pleased with V20.1.1C.3
 
Hi all.

What is the current best, safest .hex file to flash to my 2019 36v 350w motor?

Thanks
 
Wapous said:
A short video about the torque sensor concept.

https://youtu.be/mK3fbxvDF3I
:bigthumb:
Thank you for that great illustration video.
Stancecoke had already explained the working of the torquesensor on the German forum, but this video is a great addition to that.
 
Wineglass said:
20211005_152956.jpg

Unless I've just been unlucky and this was a manufacturing fault, I'm thinking that perhaps it's due to the higher case temperature that comes from the heat mods. When the motor was stock it would hit temperature limit and yet the case was not even noticeably warm. With all the mods to move heat into the case my motor temps don't go above 70degC but the case gets quite warm to the touch. And it is the internal case temperature that sets the environment the controller board has to operate in.

Cheers, Mark.

The thermal mods are not there to protect the circuit board primarily. They are to prevent damage to the permanent magnets buried inside the motor with no air flow there. On hubmotors the magnets are usually on the outside part which is moving in the air stream. On the Bafang BBSxx motors the motor itself is exposed to the air. The covers only go over the end. The TSDZ2 cover actually insulates the motor. The cooling mods make the cover part of the cooling and more like the 1000w-1500W BBSxx series. The cover not only has to get hot, but it has to get hotter than the surrounding air before it has any cooling effect at all on the motor itself. The temperature of the electronics probably doesn't change very much. But if anything they would be lower due to the heat form the motor leaving the case, and not being trapped inside with the electronics.
 
Wapous said:
A short video about the torque sensor concept.

https://youtu.be/mK3fbxvDF3I
Great work!! I just added that link to the How TSDZ2 torque sensor works page: https://github.com/OpenSourceEBike/TSDZ2_wiki/wiki/How-TSDZ2-torque-sensor-works
 
Because 850C / 860C display are expensive, suddenly fail and were done to not be repaired, I now have our simple, DIY, easy to repair display, based on the SW102 display. It works very well but still needs some more firmware development, mainly to refine some things.

Here is a video of the display working on my EBike with TSDZ2 motor:
[youtube]viimR74NPZk[/youtube]

I did the 3D printing enclosure in a way it fix to the handlebar in a similar way of the Garmin GPS displays, using the elastic, so should be easy to fix on any handlebar position.
On the bottom there are 4 small screws on the corners, to easily open in the case of needed repair:




It also works with our mobile app, where we can make the configurations, change the assist level and see motor data:
android1-small.jpg
android2-small.jpg
android-config1.jpg


rcx194 said:
Hi all. First time I've tried using my TSDZ2 bike since Covid homeworking started 19 months ago for me. The bike wouldn't power up. The 860C display back light comes on but display stays black. I'm able to flash the display and upgraded the firmware but still same. I guess it just died.

Later I will also use this same display for my Xiaomi scooter:
 
Wapous said:
TSDZ2 engine is no longer listed on Tongsheng website "www.tsbicycle.net".
And most of the sellers have stopped selling it.
It is rumored that Bafang is suing Tongsheng.
Will replacement parts be available in the long term?
There is a discussion about this on facebook
https://m.facebook.com/groups/734929600180813/
I was thinking of buying two TSDZ2 engines this spring for friends but I will follow the development of this story.
Thanks for the news!! this is an important warning. I never expected this end :-(
 
Back
Top