Page 319 of 320

Re: New "TSDZ2 Torque Sensor Central Motor"

Posted: Oct 14 2020 7:00pm
by jbalat
Haven't had to repair the bosch motor yet knock.on wood and out of warranty but I know there is a bearing kit you can buy for it that I'm sure is way overpriced:) agree tsdz2 is easy to work on, I'm sure bosch will not be any different and there are videos out there to help if the need arises.

Re: New "TSDZ2 Torque Sensor Central Motor"

Posted: Oct 15 2020 7:52am
by knutselmaaster
Bosch doesn't want you to repair the motor. They want you to replace it entirely so that they can have your money.
The only part they provide is 1 bearing, but only in the complete motor cover so that the replacement costs for the bearing are almost $300.
And don't even think that you can use old motors for spare parts, they only send a new motor once they have received the old one that needs to be sent back.
And I can know, I'm an official Bosch service center.
When I told them that their policy is rediculous, they told me I should be happy because it made me earn money...

Re: New "TSDZ2 Torque Sensor Central Motor"

Posted: Oct 15 2020 12:53pm
by AZUR
casainho wrote:
Oct 10 2020 11:37am
AZUR wrote:
Oct 10 2020 10:50am
I would like to share with the Forum what I did on the TSDZ2 engine to improve heat dissipation.
Would be important if you could share also on the wiki, on the specific page for this topic. Maybe you could even create a new page with all this information, adding your name on the page and then link on the main page of this topic.
Hi Casainho

Ok, for sure, I will be very happy to contribute to the wiki.

Thanks

Re: New "TSDZ2 Torque Sensor Central Motor"

Posted: Oct 16 2020 3:06pm
by sysrq
bergerandfries wrote:
Oct 13 2020 10:05pm
LeftCoastNurd wrote:
Oct 13 2020 5:28pm
How was the max temp of 85C/185F determined? my motor reaches that rather quickly on a 27C/80F ambient kind of day like today wehn I'm climbing hills and such. I did some poking around, and it seems like a lot of DC perm magnet motors are rated to operate at temps nearly twice that hot, like 150C
cashino discovered it by personal experience. He has pictures of the fried motor in the wiki. 85C might be conservative, but I bet it would cook off at 95C, so not SUPER conservative. True that high quality motors might tolerate high temps, but the consumer grade motor in the TSDZ2 has proven that it will not tolerate 150C. Somewhere someone posted that the lower grade magnets are the issue...
That might be the reason why it seems to be getting weaker and weaker despite Panasonic thermal modification. Not sure if low grade Chinese CYCBT battery is to blame (watt meter shows weaker and weaker performance after each monthly 40 mile excursion).

Re: New "TSDZ2 Torque Sensor Central Motor"

Posted: Oct 16 2020 3:13pm
by LeftCoastNurd
so we need to find a higher temp rated motor in the same form factor, eh? I gather these are three phase brushless motors, with some sort of 3 phase speed encoder (guessing based on the 3 power wires and 3 signal wires I see in the pictures)

Re: New "TSDZ2 Torque Sensor Central Motor"

Posted: Oct 16 2020 9:21pm
by LeftCoastNurd
I'm looking at the doc page on the Assist level table, and noting the default 20 assist levels went up to 1.50, which if I'm reading correctly, represents 1.50 * human_Watts == amps, 1.5 * 100 watts is 150 amps?!? while I haven't yet calibrated my torque, its telling me that if I'm outputting 100 watts pedaling at cadences around 75 with light pressure, and 125-150 watts if I'm pedaling a bit more vigorously, and that's *WAY* over the max 18 amps (which I have recently detuned to 12A in a attempt to get the heat under better control). in fact, anything much over 0.100 is too much (0.1 * 150 watts human == 15 amps)

using 10 steps, my current Assist level table looks like...

step setting
1 0.020
2 0.030
3 0.045
4 0.068
5 0.101
6 0.152
7 0.228
8 0.342
9 0.513
10 0.769

and even with those values, anything much past '5' is too much power per my calculations?

am I totally misunderstanding this? sadly, it was 98F here today, and I just wasn't in the mood for a ride since my rides have to end with a mile long 5-8% grade with no shade.

Re: New "TSDZ2 Torque Sensor Central Motor"

Posted: Oct 16 2020 9:31pm
by mctubster
LeftCoastNurd wrote:
Oct 16 2020 9:21pm
I'm looking at the doc page on the Assist level table, and noting the default 20 assist levels went up to 1.50, which if I'm reading correctly, represents 1.50 * human_Watts == amps, 1.5 * 100 watts is 150 amps?!? ... Snip

am I totally misunderstanding this? sadly, it was 98F here today, and I just wasn't in the mood for a ride since my rides have to end with a mile long 5-8% grade with no shade.
I have a question re this. I'm not really sure why the assist calculations became so complicated to the end user. Earlier versions of the firmware simply multiplied the users power or torque output. Very similar to other mid drive systems like Bosch and Brose.

Why was the change made to something that is not intuitive?

Re: New "TSDZ2 Torque Sensor Central Motor"

Posted: Oct 16 2020 10:01pm
by LeftCoastNurd
gain /should/ be on some sort of logarithmic scale, as you want small steps at the low end and bigger steps at the high end. too many naïve programmers implement linear gain steps (for instance 10%, 20%, 30%, 40%, 50%, 60%, 70%, 80%, 90%, 100%), if you do this with volume, brightness, or most any other power unit, you find the lower steps are too big, and the upper steps are too small.

Re: New "TSDZ2 Torque Sensor Central Motor"

Posted: Oct 17 2020 1:14pm
by Wapous
If the main gear's sprag clutch no longer drives the chainring and no "crunch crunch" is heard then it is sufficient to clean its mechanism. This mechanism must be clean and without lubricant.
After removing the protector, each of the components of the sprag mechanism can be dismantled, which includes 13 rollers and 13 springs. With good tweezers the work is easy. Each of the 26 components must be removed and cleaned as well as the protective cover and the housing. I used isopropyl alcohol as a cleanser.

Re: New "TSDZ2 Torque Sensor Central Motor"

Posted: Oct 17 2020 1:14pm
by Elinx
LeftCoastNurd wrote:
Oct 16 2020 3:13pm
so we need to find a higher temp rated motor in the same form factor, eh? ...
Like the one Vbruun found. A 220oC motor with hall sensors, but twice the price of the bare tsdz2 motor.

Re: New "TSDZ2 Torque Sensor Central Motor"

Posted: Oct 17 2020 2:39pm
by Swannking
Elinx wrote:
Oct 17 2020 1:14pm
LeftCoastNurd wrote:
Oct 16 2020 3:13pm
so we need to find a higher temp rated motor in the same form factor, eh? ...
Like the one Vbruun found. A 220oC motor with hall sensors, but twice the price of the bare tsdz2 motor.
Retro-fitting any 3rd party motor is going to require some serious skills and imagination.

Re: New "TSDZ2 Torque Sensor Central Motor"

Posted: Oct 17 2020 3:09pm
by LeftCoastNurd
Swannking wrote:
Oct 17 2020 2:39pm
Retro-fitting any 3rd party motor is going to require some serious skills and imagination.
I'd imagine you could scavenge the end plate off your baked TSDZ motor, and mount the replacement motor on the same plate... of course you'd need a compatible pinion gear and shaft fit onto the replacement motor... and possibly reprogram the firmware to understand any specific differences in the new motor (phase/windings count, hall sensor configuration, etc).

you'd want to replace all the lubricants in the TSDZ2 with ones better suited to higher temperatures,


Meanwhile I need to start testing mine at lower power limits to see where I can operate it without it overheating when I'm right in the middle of too much fun :D

Re: New "TSDZ2 Torque Sensor Central Motor"

Posted: Oct 17 2020 3:34pm
by leisesturm
LeftCoastNurd wrote:
Oct 16 2020 10:01pm
gain /should/ be on some sort of logarithmic scale, as you want small steps at the low end and bigger steps at the high end. too many naïve programmers implement linear gain steps (for instance 10%, 20%, 30%, 40%, 50%, 60%, 70%, 80%, 90%, 100%), if you do this with volume, brightness, or most any other power unit, you find the lower steps are too big, and the upper steps are too small.
I think this would be true mainly with firmware with a limited number of power levels. Say, 5 to 10. With 20 steps like the v 1.0 TSDZ2 OSF, the steps are so close together as is, that a strictly linear progression should not be all that bad. I could be wrong, but I don't think the 'non-intuitiveness' that a recent poster complained about has anything to do with the steps curve being made logarithmic. I don't think it is. But I don't really know all that much about it. Yet.

Re: New "TSDZ2 Torque Sensor Central Motor"

Posted: Oct 17 2020 3:51pm
by leisesturm
So, I have a question about calibrating the torque sensor of the TSDZ2. Mine just arrived Thursday, but my battery is not here yet. In the meanwhile I have been reading the OSF Wiki and also this thread. My dealer says calibrating the torque sensor is not essential but is recommended. The process, in the wiki could be better explained, I think. Obtaining a progression of weights is not at all straightforward. How to immobilize the bicycle and where could be better illustrated, I think. Also the units involved.

Some speak of using "gym weights"! I am a gym goer. A 10lb weight is already getting too large to fit on a pedal. Do you really need a continuous range between 0 and your bodyweight? I weigh 195lbs. Is that really an amount of weight that I could apply to the pedals? Some mention using luggage scales. I like this idea. How is it implemented? Again, what kind of range is ideal. I was in the ACE hardware and they had torque wrenches and one of them went from 10 to 150 ft/lbs. if I am not mistaken. I think the more common unit is nm and I don't know if I could just take the cranks off and put a torque wrench on the BB spindle and calibrate it that way.

I don't think I am the only one who could benefit from much more detailed (and maybe illustrated) descriptions of torque sensor calibration. Thanks in advance for any thoughts, ideas or how to's.
-H

Re: New "TSDZ2 Torque Sensor Central Motor"

Posted: Oct 17 2020 4:13pm
by casainho
leisesturm wrote:
Oct 17 2020 3:51pm
I don't think I am the only one who could benefit from much more detailed (and maybe illustrated) descriptions of torque sensor calibration. Thanks in advance for any thoughts, ideas or how to's.
-H
After doing it it will be clear for you. Looking forward to see your contribution to the wiki.

Re: New "TSDZ2 Torque Sensor Central Motor"

Posted: Oct 17 2020 4:36pm
by LeftCoastNurd
Newton-meters is the SI/metric unit of force, while ft-lbs (properly lbf-ft) is the 'imperial' unit. 1 lbf-ft is about 1.35 N-m.

re: using dumbell/exercise weights, I think the thing to do would be to hang them off the pedal rather than try and put them on top of the pedal. My general plan that I'm still working through in my head is to get a couple 50s and a couple 10s, and do

1) my full weight (220 lbs)
2) 100 lbs (50+50)
3) 70 lbs (50+10+10)
4) 50 lbs (50)
5) 20 lbs (10+10)
6) 10 lbs (10)

and hope that's enough steps to interpolate the curve. in a pinch probably could do 50+50+10+10 for 120 lbs. to get all this under the pedal, I'll probably have to put the bike up on blocks (or I have a pair of truck ramps that would raise the wheels about a foot), and somehow brace the frame so it can't fall over while I'm wrestling with the weights, and lock the back wheel so the weight won't push the bike forward or spin the wheel. haven't really figured out what I'd use as a hook to stack the weights on, maybe a steel cable with a loop to go around the pedal, and a disk or cross bolt at the bottom to support the weights.

Re: New "TSDZ2 Torque Sensor Central Motor"

Posted: Oct 17 2020 11:15pm
by leisesturm
Sounds dangerous. There are no 50lb. weights btw. There are only 5,10,25 & 45lb weights. A 25lb weight is no joke and a 45lb plate is an absolute beast. That's why I think some kind of large torque wrench might be more workable. I didn't think asking here was a bad thing. What is the point of us all re-inventing the wheel? Aren't we supposed to help each other? I don't mind refining an idea but if I had any clue about all of this I wouldn't be asking. Obviously, this last isn't aimed at you. Only have one thread left before the wife gets mad.

Re: New "TSDZ2 Torque Sensor Central Motor"

Posted: Oct 18 2020 3:06am
by mallesepp
Wapous wrote:
Oct 17 2020 1:14pm
If the main gear's sprag clutch no longer drives the chainring and no "crunch crunch" is heard then it is sufficient to clean its mechanism. This mechanism must be clean and without lubricant.
After removing the protector, each of the components of the sprag mechanism can be dismantled, which includes 13 rollers and 13 springs. With good tweezers the work is easy. Each of the 26 components must be removed and cleaned as well as the protective cover and the housing. I used isopropyl alcohol as a cleanser.
Not sure if this is worth all of the work. It is certainly also wear and tear on the metal wedges in the game that the clutch no longer works!

Re: New "TSDZ2 Torque Sensor Central Motor"

Posted: Oct 18 2020 6:06am
by Wapous
mallesepp wrote:
Oct 18 2020 3:06am
Wapous wrote:
Oct 17 2020 1:14pm
If the main gear's sprag clutch no longer drives the chainring and no "crunch crunch" is heard then it is sufficient to clean its mechanism. This mechanism must be clean and without lubricant.
After removing the protector, each of the components of the sprag mechanism can be dismantled, which includes 13 rollers and 13 springs. With good tweezers the work is easy. Each of the 26 components must be removed and cleaned as well as the protective cover and the housing. I used isopropyl alcohol as a cleanser.
Not sure if this is worth all of the work. It is certainly also wear and tear on the metal wedges in the game that the clutch no longer works!
Yes It was worth all the work. And the main reason is that I can keep using my bike (my only bike) rather than waiting for a replacement part for weeks.
I mentioned in my post that if the sprag clutch makes noise like a CRUNCH then it's a WEAR problem. In my case there was NO noise because the problem was the dirt of the mechanism which prevents the rollers (claws) from moving freely.
You are not asking how long it took me but you are saying that it is too long. Well if you had asked, I would have told you about an hour and a quarter including the time to take the part apart.
And if I took the time to post this on this forum it's because the results are there. I tried and it worked. Period.

Re: New "TSDZ2 Torque Sensor Central Motor"

Posted: Oct 18 2020 6:07am
by ilu
leisesturm wrote:
Oct 17 2020 11:15pm
Sounds dangerous. There are no 50lb. weights btw. There are only 5,10,25 & 45lb weights. A 25lb weight is no joke and a 45lb plate is an absolute beast. That's why I think some kind of large torque wrench might be more workable. I didn't think asking here was a bad thing. What is the point of us all re-inventing the wheel? Aren't we supposed to help each other? I don't mind refining an idea but if I had any clue about all of this I wouldn't be asking. Obviously, this last isn't aimed at you. Only have one thread left before the wife gets mad.
I think using a luggage scale is safer and easier, I used just a super cheap mechanical spring scale from a hardware store. I hanged the scale from a chin-up bar above the bike, with a rope that I could also pull down. The hook of the scale goes to pedal, and keeping the crank horizontal with my foot I was pulling down on rope attached to the scale. Fancier scale without a lenghtening spring would be easier to use as you don't need to change the position of the scale when keeping the pedal levelled. Torque wrench sounds like the easiest possible solution, as long as you find a fitting piece to the axle. I used weights between 5-30kg and my own weight 85kg and got a good curve. Between 30 and 85kg the difference between sensor readings are already minimal.

Re: New "TSDZ2 Torque Sensor Central Motor"

Posted: Oct 18 2020 4:04pm
by LeftCoastNurd
ilu wrote:
Oct 18 2020 6:07am
I think using a luggage scale is safer and easier, I used just a super cheap mechanical spring scale from a hardware store. I hanged the scale from a chin-up bar above the bike, with a rope that I could also pull down. The hook of the scale goes to pedal, and keeping the crank horizontal with my foot I was pulling down on rope attached to the scale. Fancier scale without a lenghtening spring would be easier to use as you don't need to change the position of the scale when keeping the pedal levelled. Torque wrench sounds like the easiest possible solution, as long as you find a fitting piece to the axle. I used weights between 5-30kg and my own weight 85kg and got a good curve. Between 30 and 85kg the difference between sensor readings are already minimal.
i'm not sure I'm picturing what you're doing with the luggage scale? I do have one that goes up to 50 lbs, which could be used for the low end readings, but pulling down on it with a consistent force, I'm not sure I can do that without its readings bouncing all over.

I'd have to think about the relationship of rotational torque in lb-ft to linear lbs before I'd mess with a torque wrench. I think you'd need to take the crank arm length into account...

Re: New "TSDZ2 Torque Sensor Central Motor"

Posted: Oct 18 2020 4:19pm
by mctubster
LeftCoastNurd wrote:
Oct 18 2020 4:04pm

i'm not sure I'm picturing what you're doing with the luggage scale? I do have one that goes up to 50 lbs, which could be used for the low end readings, but pulling down on it with a consistent force, I'm not sure I can do that without its readings bouncing all over.
You pull up in it. Have the crank arm horizontal facing rearwards. I lean on the seat to counteract and help the rear wheel with traction. A second person to hold the brake works also. I measured up to 50KG like this and used my body weight for the final reading

Re: New "TSDZ2 Torque Sensor Central Motor"

Posted: Oct 18 2020 5:06pm
by leisesturm
LeftCoastNurd wrote:
Oct 18 2020 4:04pm
ilu wrote:
Oct 18 2020 6:07am
I think using a luggage scale is safer and easier, I used just a super cheap mechanical spring scale from a hardware store. I hanged the scale from a chin-up bar above the bike, with a rope that I could also pull down. The hook of the scale goes to pedal, and keeping the crank horizontal with my foot I was pulling down on rope attached to the scale. Fancier scale without a lenghtening spring would be easier to use as you don't need to change the position of the scale when keeping the pedal levelled. Torque wrench sounds like the easiest possible solution, as long as you find a fitting piece to the axle. I used weights between 5-30kg and my own weight 85kg and got a good curve. Between 30 and 85kg the difference between sensor readings are already minimal.
i'm not sure I'm picturing what you're doing with the luggage scale? I do have one that goes up to 50 lbs, which could be used for the low end readings, but pulling down on it with a consistent force, I'm not sure I can do that without its readings bouncing all over.

I'd have to think about the relationship of rotational torque in lb-ft to linear lbs before I'd mess with a torque wrench. I think you'd need to take the crank arm length into account...
I've looked at luggage scales on Amazon https://www.amazon.com/Luggage-Scales/b ... 2477391011 and they all read to 110lbs. They are cheap (to me) as well. I've seen one that reads much higher that I am working out the details of how it might be made to work: https://www.amazon.com/Smart-Weigh-Ship ... 487b18b923. I don't know but I think a more or less consistent force could be applied long enough to get a reading from both the scale and the display. Also, if a torque wrench is NOT used in conjunction with readings using the crank-arms, the crankarms length does not have to be taken into account. Make all the measurements with the torque wrench, including the final one, which I have not decided must equal ones bodyweight, and you have a proper curve of relationships between 'real' weights (torques) and the interpolated ones of the ADC.

Re: New "TSDZ2 Torque Sensor Central Motor"

Posted: Oct 19 2020 4:27am
by Jaybee258
bergerandfries wrote:
Oct 13 2020 10:05pm
LeftCoastNurd wrote:
Oct 13 2020 5:28pm
How was the max temp of 85C/185F determined? my motor reaches that rather quickly on a 27C/80F ambient kind of day like today wehn I'm climbing hills and such. I did some poking around, and it seems like a lot of DC perm magnet motors are rated to operate at temps nearly twice that hot, like 150C
cashino discovered it by personal experience. He has pictures of the fried motor in the wiki. 85C might be conservative, but I bet it would cook off at 95C, so not SUPER conservative. True that high quality motors might tolerate high temps, but the consumer grade motor in the TSDZ2 has proven that it will not tolerate 150C. Somewhere someone posted that the lower grade magnets are the issue...
The 85°C will be specific to the position of the temperature sensor within the motor. The inner parts of the motor will be much hotter than this, but it would be difficult to measure these temperatures. So, it has been found that keeping the temperature of that sensor location below 85°C will stop the hottest parts of the motor from overheating.

Re: New "TSDZ2 Torque Sensor Central Motor"

Posted: Oct 19 2020 4:54am
by Mr.Flibble
jbalat wrote:
Oct 14 2020 7:00pm
Haven't had to repair the bosch motor yet knock.on wood and out of warranty but I know there is a bearing kit you can buy for it that I'm sure is way overpriced:) agree tsdz2 is easy to work on, I'm sure bosch will not be any different and there are videos out there to help if the need arises.
There is at least one UK company that fixes Bosch motors, and sells rebuild kits.
It looks quite easy to work on - as long as you have a bearing press.