TSDZ8 OSF (open source firmware)

mstrens

100 W
Joined
Nov 8, 2024
Messages
126
Location
belgium
I propose to discuss in this thread all the questions/issues about the open source firmware for TSDZ8 motor.
Initially the first posts about this OSF firmware have been put in another thread ( e.g. new tsdz8 pswpower ) but I realized that it can be confusing to know if a post is related to the original firmware or to the OSF version.

So it seems me better to use a specific thread just for the OSF version.
 
Cool. What is the max amperage the osf will allow to the motor? I would like to use at least 23 amps
OSF can accept the max amperage that you want if you compile yourself with the source that I provide on github.
In the precompiled version, I expect to set the max Amp on 23A it it is confirmed that the original firmware support this Amp.
Still, this seems a quite high value because with a 48V battery that is fully loaded it corresponds to 54V X23A = 1242W.
I am not sure that TSDZ8 can support such a power continuously.
It is up to you to take the risk or not.
 
I did my first ride today with the TSDZ8 motor flashed with OSF. The assistance was a bit choppy uphill when you gave more power to the pedals, when the motor had to assist more. A bit the same feeling as with a worn axle on a TSDZ2 motor. It felt like every other pedal pushing assisted well and every other pedal pushing less well.
I don't know if it's a TSDZ8 motor feature or OSF related.
It also felt like the torque sensor was not working properly. In java configurator I have set the motor to start in torque mode.
 
if the bbs02 can handle 25amps, and the tsdz8 is just slightly better build then bbs02, then i think tsdz8 should also be able to handle 25amps
 
I did my first ride today with the TSDZ8 motor flashed with OSF. The assistance was a bit choppy uphill when you gave more power to the pedals, when the motor had to assist more. A bit the same feeling as with a worn axle on a TSDZ2 motor. It felt like every other pedal pushing assisted well and every other pedal pushing less well.
I don't know if it's a TSDZ8 motor feature or OSF related.
It also felt like the torque sensor was not working properly. In java configurator I have set the motor to start in torque mode.
I do not expect that it is related to the torque sensor but I am not sure.
Could you try OSF in cadence assist mode. In the cadence assist mode, torque sensor is not use at all. If the issue exists also in cadence assist mode, we can be sure that it is not related to the torque sensor.
 
I do not expect that it is related to the torque sensor but I am not sure.
Could you try OSF in cadence assist mode. In the cadence assist mode, torque sensor is not use at all. If the issue exists also in cadence assist mode, we can be sure that it is not related to the torque sensor.
Yes I can test cadence mode someday.
 
Today (still torque mode) TSDZ8 was working better I drived 45 km:s. In the beginning of trip was little bit fluctuating in assist. Maybe bearings and gear are still little bit stiff.

Brake censors are not working. It would be nice to cut power fast pushing brakes in case of emergency.
 
Today (still torque mode) TSDZ8 was working better I drived 45 km:s. In the beginning of trip was little bit fluctuating in assist. Maybe bearings and gear are still little bit stiff.

Brake censors are not working. It would be nice to cut power fast pushing brakes in case of emergency.
I expect that brake sensor code in the firmware is correct (at least it worked in the first versions and did not change).
Still to work, brake must be enabled in the javaconfigurator. Is it so in your config?
 
I expect that brake sensor code in the firmware is correct (at least it worked in the first versions and did not change).
Still to work, brake must be enabled in the javaconfigurator. Is it so in your config?
Brake sensors are enabled in java configurator.

Today I will ride in cadence mode.
 
Please note that a small bug has been found and fixed.
In case of reflash of a new HEX file generated by the javaconfigurator, some parameters from the new configuration were discarded.
To fix this, mbrusa made an update of the javaconfigurator for TSDZ8.
So if you already have or will update a TSDZ8 configuration, please, use the new version of javaconfigurator and reflash the HEX file it generate.
 
I set TSDZ8 to start on cadence mode. I started to pedal and bike launched like a rocket. Assist setting in java configurator was 90, 80. 150 and 200. Assist was very strong and changing assist levels did not change anything. Assist was same at all levels. I went back to garage and set motor to start on torque mode. Now motor did not assist at all. I ride short trip and sometimes motor assist little bit but never well, most of the time there was no assist at all. I had same assist settings as cadence mode. A day before motor was working on torque mode quite well.

So my conclusion is new motor do not have working torque sensor and maybe controller is broken too because display functions are working bad. Brake sensors are not working and I cannot get access to deeper parameters of VLCD5 display.
 
I set TSDZ8 to start on cadence mode. I started to pedal and bike launched like a rocket. Assist setting in java configurator was 90, 80. 150 and 200. Assist was very strong and changing assist levels did not change anything. Assist was same at all levels. I went back to garage and set motor to start on torque mode. Now motor did not assist at all. I ride short trip and sometimes motor assist little bit but never well, most of the time there was no assist at all. I had same assist settings as cadence mode. A day before motor was working on torque mode quite well.

So my conclusion is new motor do not have working torque sensor and maybe controller is broken too because display functions are working bad. Brake sensors are not working and I cannot get access to deeper parameters of VLCD5 display.
Sorry to hear this.
It is strange that you do not get access to deeper parameters of vlcd5 display. Can you send me the last .ini and Hex files that were used/generated by the javaconfigurator.
Did you tried afterward to upload original firmware to see if controller is broken or not?
 
I'm little bit busy now. I ordered new parts to motor. I don't want to even try original firmware because I have heard of users that it assist too much on first two levels. I try send .ini later.
 
I'm little bit busy now. I ordered new parts to motor. I don't want to even try original firmware because I have heard of users that it assist too much on first two levels. I try send .ini later.
Flashing the original firmware (or one of 2 the versions that has been delivered - see e.g. in facebook file section or ask me the link) was just to see if the controller is still OK.

Ebikestuff is also testing OSF and got an issue that destroyed the controller.
The problem is that we do not know why.
 
Flashing the original firmware (or one of 2 the versions that has been delivered - see e.g. in facebook file section or ask me the link) was just to see if the controller is still OK.

Ebikestuff is also testing OSF and got an issue that destroyed the controller.
The problem is that we do not know why.
I have original version on computer which was on motor when I get it. I wonder if it has 5 assist levels because it came with 860C display and when I tested it on the table I saw on display 5 levels. I am not in facebook. Ebikestuff has some version to TSDZ8 on their website.
 
I have original version on computer which was on motor when I get it. I wonder if it has 5 assist levels because it came with 860C display and when I tested it on the table I saw on display 5 levels. I am not in facebook. Ebikestuff has some version to TSDZ8 on their website.
ebikestuff versions are the same those on facebook
 
I set TSDZ8 to start on cadence mode. I started to pedal and bike launched like a rocket. Assist setting in java configurator was 90, 80. 150 and 200. Assist was very strong and changing assist levels did not change anything. Assist was same at all levels. I went back to garage and set motor to start on torque mode. Now motor did not assist at all. I ride short trip and sometimes motor assist little bit but never well, most of the time there was no assist at all. I had same assist settings as cadence mode. A day before motor was working on torque mode quite well.

So my conclusion is new motor do not have working torque sensor and maybe controller is broken too because display functions are working bad. Brake sensors are not working and I cannot get access to deeper parameters of VLCD5 display.
@dameri
This PM I made a test and it seems that I also destroyed my controller. I was using uc_probe_monitoring when it happens.
It could be that it was was also the case when ebikestuff destroyed his controller.

In fact uc_probe_monitoring can change some data in the microprocessor. If it is used with an elf. file that does not match the hex file that has been flashed it could update data at unexpeted address and so generate unpredictable results. If you never used uc_probe_monitoring than it could that there is another issue in your case.

Please let me know if you ever use uc_probe_monitoring?
 
@dameri
This PM I made a test and it seems that I also destroyed my controller. I was using uc_probe_monitoring when it happens.
It could be that it was was also the case when ebikestuff destroyed his controller.

In fact uc_probe_monitoring can change some data in the microprocessor. If it is used with an elf. file that does not match the hex file that has been flashed it could update data at unexpeted address and so generate unpredictable results. If you never used uc_probe_monitoring than it could that there is another issue in your case.

Please let me know if you ever use uc_probe_monitoring?
I have not even heard of uc_probe_monitoring. I'm a user not developer. I have only flashed controller with J-link.
However I'm interested in what really happens.

Maybe my motor controller was already damaged when I got the motor. Bad manufacturing.
 
I have already said I don't want use original firmware because users have said it assist too much on first two levels.
Yes but it is only to make a test to see if your controller is still OK or not.
You can flash another firmware afterwards.

As my controller is dead and I do not have a spare part, I can't continue currently my test/development.

I asked about uc-probe because a bad use of it (like I probaly did) could perhaps explain why my controller died.

If you controller also died, then there must be another reason that we do not know and this is a big issue.

I do not think you controller was defect at the origin because the first tests you did give similar results as the tests done by ebikestuff.
 
Back
Top