TSDZ2 OSF for all displays, VLCD5-VLCD6-XH18, LCD3, 860C-850C-SW102.

[mbrusa]

Did I get it right?
On the 36V motors that suit you you have OSF, while on the 48V one you have stock firmware, and you would like to have the same performance leaving the stock firmware.
It is not possible.
The only way is to install OSF on the 48V motor too.
Is there a reason you haven't done this yet?

 

[dgc]

Before I even put the OSF on the 36v motors there was a noticeable difference between them and the 48v motor, so thought I would try and sort the 48v motor out first before updating the it's firmware as I will also have to change over the VLCD13 display to my spare VLCD5 with brake sensors as the VLCD13 has at a modified data cable which I have grafted brake sensors onto.
If I can get the spare VLCD5 to display TE and TE1 that would be a help in the first place or as previous read the values via the TSDZ2 Parameter Configurator or the STVP program.

 

[Elinx]

Before I even put the OSF on the 36v motors there was a noticeable difference between them and the 48v motor, .....

The sensitivity of all Tsdz2 torquesensors is different, so with stock firmware it is possible to feel that difference.
You can read TE and TE1 with Vlcd5, but not change it inside stock firmware.

Only with OSF it is possible to optimize the range of the torquesensor and the feeling between the different motors will be more equal. This is one of the advantages of OSF.
So the advice is to flash OSF on the 48V version too and calibrate the zero/max. torquesensor values.

 

[dgc]

Elinx, Many thanks, but as elsewhere I try to look on the VLCD5 at the TE and TE1 values but get to the JAP EUR mode selection and the display jumps out unless I have to press something else other than the "i" button.

Totally understand there are variations between the torque sensors as years back I use to build and calibrate high pressure strain gauge transducer where 0 - 12 mv represented 0 - 7000 bar but we offset the transducer output with a calibration resister at around 80% full load.

 

[NoWay66]

Display shows "ODO" > You access the "normal " hidden menu with pressing the Power and I-Button .
Display shows "TRIP" > You access the torque-menu with pressing the Power and I-Button.

 

[dgc]

Thanks guys, will have a play.

 

[dgc]

OK guys, with your help we have the results, which opens up more questions.

36v motor standard firmware
TE TE1
81 81 0 kgs
101 5 kgs

36v motor OSF firmware
TE TE1
156 0 0 kgs
39 5 kgs

48v motor standard firmware
TE TE1
81 81 0 kgs
100 5 kgs

On all three with a "calibrated boot" I can get a TE1 reading of around 150 so I will have to get more weights out, I think on the 36v OSF motor the TE reading of 156 is where the firmware doe's not reset on power up in the same way there is a 5 - 10 mph indicated on the display which clears almost immediately once you move the bike or when moved from TE to TE1.
So both the bikes with 36v motors are fine, the 48v results seems to suggest to me that the 48v motor is fine but is this down to the differences between the 36v and 48v motor setups. Looks like OSF is going on the 48V motor.

 

[Elinx]

....

On all three with a "calibrated boot" I can get a TE1 reading of around 150 so I will have to get more weights out, ....

imho you can't compare torque values of stock FW vs OSF
To know the sensitivity or full range of the torquesensor, measure the max. value with you full bodyweight on the pedal.
5kg is too few for a comparision, because the range isn't lineair.

 

[dgc]

Elinx: yes, will be having a go tomorrow with me as the load, was more interested today in getting at least a reading from the three torque sensors. Use to do 10 point calibrations on pressure transducer with a dead load tester this time I'm the dead load.

 

[Yanis]

Hi mbrusa ,
I have what seems to be a small bug on 20.1C.2-update-3 with a VLCD6 display :
With the 'walk assist debounce time' option on, when releasing the 'up' button, the motor remains active not just during the time set, but until a button is pressed again. So the bike keeps going by itself unless you touch the display again.
I kinda like it this way, but if I understood well, it's not how it's supposed to work
I didn't use this option before, so I don't know if this behaviour is present on previous versions.

 

[mbrusa]

Hi Yanis.
It has always been like this, no one has ever noticed because it is a rarely used feature.
Once time runs out, walk assist must end, so there's a problem to fix.
However, it is possible to interrupt the debounce time by changing the assist level, or with the brakes if the sensors are installed. This is an expected feature.

 

[dgc]

OK guys two, as previous but with one human 65 kgs dead weight on the pedals

36v motor standard firmware
TE TE1
81 81 0 kgs
101 5 kgs
150 65 kgs

36v motor OSF firmware
TE TE1
156 0 0 kgs
39 5 kgs
95 65 kgs

48v motor standard firmware
TE TE1
81 81 0 kgs
100 5 kgs
153 65 kgs

Both the 36v and 48v motors with standard firmware have a very similar TE1 output which leads me back to differences between the 36v and 48v motor setups.

 

[labzoli]

This is actually the case, there is a bug already from version 20.1C.2
I never noticed because I never tried the weight calibration again. Thanks for the report.
It will be fixed with the new version that I will release soon.

Many Thanks
Actually I'm working in the sw testing area

 

[Elinx]

...., as previous but with one human 65 kgs dead weight on the pedals
....

36V sfw 81-150 and 48V sfw 81-153 looks both normal and indeed very similar, so you would expect about the same sensitivity.

36V OSF 156-95 looks strange to me with a "negative" range.
Or are these values 0-95 the right one? (Personally I never have done this with a Vlcd5)
Is this measured too, with Level 1-E04 (level 1 > 6 pushes + 1push on light)?

correction:
As mbrusa did explain here, the reading TE/TE1 with OSF is not what I thought it was.

36V OSF 95-215 which is a excellent range (delta) of 156, that doesn't even need calibration.

 

[dgc]

Elinx, with the OSF I found that when you power up the drive there is always 4 - 10mph indicated momentarily on the VLCD5 and then it zeros, now similarly the initial TE reading is 156 but directly you move to TE1 the reading is 0 and as you put load on the pedal it climbs up to 95 and back down to 0 when the load is removed. Are TE and TE1 data registers where the firmware store static or dynamic value. I not into the Level 1-EO4 yet you will have to explain, a bit more to learn. If TE and TE1 are raw data registers then the values are as measured and would not change as they are values prior to the firmware processing them. Use to program process instrumentation, PLC's, HMI's and built pressure transducers.

 

[mbrusa]

With OSF, TE is the adc value without weight on the pedals, TE1 is the delta.
So the adc value with weight on the pedals is TE + TE1, in your case 156 + 95 = 251.
They are 10-bit values.
With stock firmware, I don't know the size of the values, they could probably be 9 bits, or 8 bits multiplied by a correction coefficient.
I had one motor with limited delta (80 10-bit) that had higher assist than another with delta that was double (160 10-bit).
The stock firmwares were identical, there were just different bytes in data memory, but I don't know what they are for exactly, there's no documentation about it.

 

[dgc]

mbrusa, will have to let that all sink into my very old brain cells, might even try switching the 36v OSF drive back to standard, but leave that for another day, main thing I am interested in at the moment is looking at the differences between the 36v and 48v drives ignoring the firmware.

 

[Coolby]

Hi All,

I have on my tsdz2 500w the osf firmware, but I have the stock 850c came with the motor, which according to my understanding cannot be flashed. Which is the good source for a 860c, which for sure can be flashed (and delivers to EU) and where is a step-by-step guide for flashing?

Thanks

 

[dgc]

Have been looking as well for a 860C, so far Części i akcesoria do rowerów elektrycznych | ebikestuff.eu but the price is a bit enthusiastic at 97 euros, may not be a good source. I am not worried about re-programing a straight 860C display if I can find one. Not in a rush as the VLCD5's currently fitted are doing the job, shame as I have a VLCD13 on my hard-tail which is a better looking display but not included in the configurator options.

Anyway asking a daft question if I enable "STREET mode" in the configurator options how do I switch in and out of it on a VLCD5 display, said it's a daft question, it's late and I am going to sleep.

 

[The_Ball]

I bought this 860C: https://www.aliexpress.com/item/1005006139674088.html
I made sure it had Bafang UART in the name.
It is hardware version 1.3 and I was able to flash it using the bootloader method. Works great.

 

[mbrusa]

Have been looking as well for a 860C, so far Części i akcesoria do rowerów elektrycznych | ebikestuff.eu but the price is a bit enthusiastic at 97 euros, may not be a good source. I am not worried about re-programing a straight 860C display if I can find one. Not in a rush as the VLCD5's currently fitted are doing the job, shame as I have a VLCD13 on my hard-tail which is a better looking display but not included in the configurator options.

Anyway asking a daft question if I enable "STREET mode" in the configurator options how do I switch in and out of it on a VLCD5 display, said it's a daft question, it's late and I am going to sleep.

VLCD13 should work by setting 850C in the configurator.

 

[dgc]

Gents, many thanks for your advice, I will look for a version 1.3 display and try the VLCD13 with the 850C setup.

 

[Yanis]

Hi mbrusa,

In case this wasn't reported yet, v20.1C.5-beta-1-860C works perfectly on SW102. Thank you so much for your work !

Is there a way to automatically set the assist level to 0 or 1 at the startup of the display ? For now the assist level stays the same as it was when display was turned off.

 

[mbrusa]

No, currently there is no provision for setting a default level at power on, the previous one remains.
I'll consider whether to do it.

 

[mbrusa]

In recent days, forced indoors by the rain, I have completed the changes to the configurator. In Windows everything is fine.
Trying on Linux (Ubuntu), a system I'm not familiar with, there is a problem.
With an old version Ubuntu 20.10-64 and OpenJDK 11.0.11 everything is ok.
With Ubuntu 22.04.3-64 and OpenJDK 11.0.22, the graphics are corrupt.
I attach the screenshots.
Does anyone who knows Linux have any idea what this might be about.