New "TSDZ2 Torque Sensor Central Motor"

[Matze_Senpai]

has anyone tried this 1t2 cable before?

https://cycles.eco/collections/tsdz2-accessories-parts/products/1t2-cable-w-8-pin-male-for-use-with-tsdz2-w-8-pin-female-and-all-bafang-bbs-motors?variant=31596433997906

currently im running emebrusa firmware, i want to go from vlcd5 to vlcd6 but also keep the throttle function

would this work ?

 

[ilu]

has anyone tried this 1t2 cable before?

https://cycles.eco/collections/tsdz2-accessories-parts/products/1t2-cable-w-8-pin-male-for-use-with-tsdz2-w-8-pin-female-and-all-bafang-bbs-motors?variant=31596433997906

currently im running emebrusa firmware, i want to go from vlcd5 to vlcd6 but also keep the throttle function

would this work ?

No. That display connector is different (and 5 pins) than the regular TSDZ2 6-pin display connector. You would need to cut and solder similar 5-pin connector to your vlcd6 display. Also the throttle connector is different from TSDZ2, although it would work with another general cheap throttle.

 

[Matze_Senpai]

has anyone tried this 1t2 cable before?

https://cycles.eco/collections/tsdz2-accessories-parts/products/1t2-cable-w-8-pin-male-for-use-with-tsdz2-w-8-pin-female-and-all-bafang-bbs-motors?variant=31596433997906

currently im running emebrusa firmware, i want to go from vlcd5 to vlcd6 but also keep the throttle function

would this work ?

No. That display connector is different (and 5 pins) than the regular TSDZ2 6-pin display connector. You would need to cut and solder similar 5-pin connector to your vlcd6 display. Also the throttle connector is different from TSDZ2, although it would work with another general cheap throttle.

thanks didnt see that. but a sw102 mini would work i think? just checked one and its 5 pin. assuming it would, then would the OSF accept this modification?

 

[ilu]

thanks didnt see that. but a sw102 mini would work i think? just checked one and its 5 pin. assuming it would, then would the OSF accept this modification?

Yes but you need to open the display to update it's firmware and use different OSF version meant for that display.

https://github.com/OpenSourceEBike/TSDZ2_wiki/wiki/Flash-the-bootloader-and-firmware-on-SW102-using-SWD

There are also SW102 displays meant specifically for TSDZ2 and those would (probably) work with the same OSF as for vlcd5/6 but then they probably have the similar 6-pin connector also.

 

[seattlesockey]

I have 2 older setups one using the 860C display and one the old B&W LCD, both with TSDZ2 motors. 750W, with 48V batteries.

I want to upgrade the B&W LCD to a 860C display, on the older system and do the upgrades to the firmware,on both systems.

Can someone please point me to the last release for the firmware for display & motor, for these now vintage setups?

Look here:
https://github.com/emmebrusa/TSDZ2-Smart-EBike-860C/releases/tag/v20.1C.4-860C-update-2

 

[Matze_Senpai]

thanks didnt see that. but a sw102 mini would work i think? just checked one and its 5 pin. assuming it would, then would the OSF accept this modification?

Yes but you need to open the display to update it's firmware and use different OSF version meant for that display.

https://github.com/OpenSourceEBike/TSDZ2_wiki/wiki/Flash-the-bootloader-and-firmware-on-SW102-using-SWD

There are also SW102 displays meant specifically for TSDZ2 and those would (probably) work with the same OSF as for vlcd5/6 but then they probably have the similar 6-pin connector also.

thank you for your help sir, i think i found the solution to my request.
https://a.aliexpress.com/_mKhxQrY
this seems to be the cable i was looking for.

 

[casainho]

Do you mean re-use your python - or the osf c code?

I mean reusing the Pyhton code.

I assumed (maybe wrongly) that your python ebikeapp must be doing pretty much the same as ebikeapp() in the osf - is the functionality significantly different at the core? Obv it's in python not c - but the underlying algorithm must be broadly the same?

Yes, the EBike Pyhton code is doing the same as I did previously on the C code. But I also added a ramp up / down of the motor current.
And I also looked at the code on the VESC for the EBike application (even the one for the Bafang M600) and the concepts are equal, so, there is nothing new from what you already know.

 

[ilu]

thank you for your help sir, i think i found the solution to my request.
https://a.aliexpress.com/_mKhxQrY
this seems to be the cable i was looking for.

I'm glad you found the cable, as I didn't know if they are available at all. SW102 display is a big improvement as you can adjust motor setting directly from it (instead of with a computer). But if you don't need to change them that often then the extra work might not be worth it.

 

[casainho]

The EBike modular DIY electronics (https://opensourceebike.github.io/), is featured as the Project of the Week on the Python on Microcontrollers Newsletter.

This newsletter is written by the Adafruit Industries, an open-source hardware company based in New York City, that is popular among makers.

 

[dameri]

I am replacing the main shaft's roller bearings.
I did try to only desolder the black wire. Then by rotating the circuit towards the interior of the main shaft the red wire still welded sneaks into the slot. You can replace the roller bearing more easily.

 

[dsjustice]

.....
this element is filling the gap between motor and cover

This is the link to the webshop that will sell them, when available.

If you're still refining the design: I wonder if it's possible to leave a channel on the inside for the temperature sensor and its wires? It'd be really nice to keep that right up against the motor housing.

 

[pxl666]

.....
this element is filling the gap between motor and cover

This is the link to the webshop that will sell them, when available.

If you're still refining the design: I wonder if it's possible to leave a channel on the inside for the temperature sensor and its wires? It'd be really nice to keep that right up against the motor housing.

There is still plenty of room between controller and motor

 

[Inertiatic]

Hi all

I've rebuilt my TSDZ2 with a repaired torque sensor (the clutch bearing seal had worked lose and chewed through the wires).

Since then its been noticeable underpowered compared to previously. About 8mph down on average cruising speed.

Using a VLCD5 with stock firmware and is a 750W unit

Getting into the settings seems very hit & miss (my display shows info slightly different to the commonly available manual andf seems to react a bit differently), but some combination of button mashing gets the torque sensor re & re1 value. The resting figure is a bit over 100 (mph?) and raises to ~160 stamping on the pedal.

What is a typical resting figure? I'm wondering if its feeling underpowered as I'm effectively only using some of the torque sensor range as its starting high?

Is it adjustable without going OSF?

Cheers

 

[pxl666]

Hi all

I've rebuilt my TSDZ2 with a repaired torque sensor (the clutch bearing seal had worked lose and chewed through the wires).

Since then its been noticeable underpowered compared to previously. About 8mph down on average cruising speed.

Using a VLCD5 with stock firmware and is a 750W unit

Getting into the settings seems very hit & miss (my display shows info slightly different to the commonly available manual andf seems to react a bit differently), but some combination of button mashing gets the torque sensor re & re1 value. The resting figure is a bit over 100 (mph?) and raises to ~160 stamping on the pedal.

What is a typical resting figure? I'm wondering if its feeling underpowered as I'm effectively only using some of the torque sensor range as its starting high?

Is it adjustable without going OSF?

Cheers

you might try hardware calibration of tqsensor . however its hard to do without osf . do you somehow read tqsensor without osf ?

 

[Elinx]

.....
Is it adjustable without going OSF?...

imho not by the settings of standard FW, but with hardware calibration.
In that case you unscrew and move the sensor a tiny bit to get a lower zero offset.
Without OSF it is only possible with a Vlcd5 display, because you can read the torque values (te1) with the hidden settings.

 

[Inertiatic]

Doing a bit more digging it seems that the lowest figure should be ~300 at 0 input.

If that's correct, then mine appears to be reading very low...which also makes sense why it would be underpowered.

I've ordered the bits to make up a data link and been playing with the OSF software...

 

[Elinx]

.... the lowest figure should be ~300 at 0 input.
...
I've ordered the bits to make up a data link and been playing with the OSF software...

I'm curious about the source fot that value?
Flashing OSF is imho the easiest way for software calibration, but realize if the range of the torque sensor is very-very low there is a chance you got a sort of on/off switch after calibration and hardware calibration is still needed for a better experience.

 

[Inertiatic]

It was hacked back together so that may be the case.

To be honest, as long as I can get the power back a more cadencey feel isn't the end of the world as it's used as a pure commuter and I haven't the time nor patience to strip it down and tinker at the moment

 

[Inertiatic]

The value came from a link on here (OSF Calibration). I will find it later

 

[Elinx]

The value came from a link on here (OSF Calibration). I will find it later

Mentioning a value for torque doesn't mean that every value is the same with all firmware, that is why I ask for the source.
For stock FW, the Vlcd5-TE1 reading is not the same as OSF reading and older OSF readings where 8bit and now they are 10bit.

 

[Yanis]

So the B version of the TSDZ2 has been out for some time now, and I've read conflicting statements about the effectiveness of the double clutch system. Some people say the unpowered pedalling resistance is much lower, and for others it's the same or worst.. Could someone who had both versions comment about this ?

Also, the blue gear (which is white now) is supposed to be stronger now because there's no more one-way bearing. I can believe that, but I wonder about the material. Is the white nylon different, and more solid than the blue one, or did they just change color in order to be able to differenciate them at a glance ?

 

[pxl666]

since in b version you're not pulling big cog ,the drag must be lower . my colleague said it's a big difference as he needed to pedal with motor off for longer distance . the other problem is that new cog is incompatible with old spiders and there is a problem with adapting 32/34t chainrings. or maybe something changed since...

 

[Blacklite]

Doing a bit more digging it seems that the lowest figure should be ~300 at 0 input.

If that's correct, then mine appears to be reading very low...which also makes sense why it would be underpowered.

I've ordered the bits to make up a data link and been playing with the OSF software...

I’m not sure that value is correct. I think the stock firmware does what most of the OSF’s do if you haven’t chosen to do a software calibration, which is to read the unweighted value at startup and use that as a zero. Because of safety there has to be a bit of leeway, it might use 10-15 under the read value on startup. If you’ve only got a total range of 60 then there is not much to work with if the multiplier of that value is hard coded. I’d suggest MBrusa’s firmware… can fix the problem either with software calibration, which basically remaps that range of values to a larger one, or just by adjusting the actual power multiplier for each level. If the hardware range is small then obviously there is a low resolution. I’ve found mine ok with a range of about 85 from resting to standing on the pedals which has been satisfactory.

There is also the question of linearity of the sensor signal compared to input torque. Mbrusa firmware has some further calibration that can be done to try and more straight-line the curve, so to speak.

 

[Elinx]

....
There is also the question of linearity of the sensor signal compared to input torque. Mbrusa firmware has some further calibration that can be done to try and more straight-line the curve, so to speak.

The curve of the torque sensor will not be corrected.
"Further calibration" is done if you want a more exact torque display reading at the lower (linear) side of that curve.
Mbrusa has already calculated an estimated value inside OSF, depended of the measured torque range, so enabling this estimated value could be enough for most people.

 

[hayden222]

has anyone used these new(?) heat control/sink options? they are really neat!

Would love to know how many watts they can comfortably push with all the mods.

and the website/ parts used in the video:

motor core surrround
Thermal conductive insert for TSDZ2

external heatsinks
Body radiator for TSDZ2 to improve heat transfer

motor end heatplate
Heatsink ring to the motor tsdz2

I did a similar thing a couple of weeks back to mine but used a Permatex Steel Weld putty, the tape used on motor to prevent putty sticking to it is Tesa Harness Tape 51036 (they say its ok for up to 150 deg. Celcius...). The Steel Weld putty transfers heat super quick when i trialed a piece, using a heatgun on it, but prob not as good as a solid piece of aluminum...


they also have lm35 sensors (however tmp35gt9z is avail at many places) and the usual style heatsink that fits behind motor