New "TSDZ2 Torque Sensor Central Motor"
- Thread starter larsottar
- Start date
izeman
1 GW
I wouldn't heat treat it. Also depends on the steel used. But some CroMo steel should be capable enough. Better let it bend than break...RicMcK said:
izeman
1 GW
romelec was so nice to make the CAD of the axle. Some questions still remain. Is there someone with a broken one and can make some more pictures? Ideally together with the attached torque sensor?
Or maybe even send the broken one to me? I'd pay for p&p, so we can maybe make a stronger axle on day.
Or maybe even send the broken one to me? I'd pay for p&p, so we can maybe make a stronger axle on day.
safeaschuck
100 µW
- Joined
- Apr 21, 2020
- Messages
- 7
Axle Thoughts and drivetrain design observations:
Experience: Apprenticed at a gearbox manufacture which supplied F1, Rallycross, offshore powerboats, Defence etc.
Placements of up to 3 months in milling, turning, gear-cutting, heat treat, quality control, sub-contract buying depts. Also spent a spell at another contract manufacturer to Tier 1 motorsport and watched over the chief engineeers shoulder while he did FEA and asked a load of stupid questions!
The TSDZ2 is a clever design, I'd posit that Bafang haven't attempted a bolt through torque sensor because it's hard to pull off reliably for a low cost.
If this Motor was made in Germany or Japan it would not be breaking. As someone mentioned on this forum it was originally designed in Germany and what we are buying is, in one sense or another a Chinese native knock off. Legit or not, no German owned company would tolerate these tolerances or material quality. The specs were probably marginal for heavy riders pounding across rough terrain BEFORE high quality steel was substituted for Chiniesium, vacuum quenched case hardening was supsituted for a blowtorch and a bucket of water and the tolerances were multiplied by 10.
I sympathise with everyones failures but in order to avoid unnecesary frustration we have to be honest with ourselves about what we are dicking around with here. It is NOWHERE NEAR on par with even the shimano mentioned earlier for toughness. Also the shimano problem looks unrelated to my casual glance.
A bottom bracket is not big enough to allow a sufficiently chunky axle, torque sensor housing AND a structural casting to pass through.
If the materials were (relatively) expensive, the machines that produced the items high quality, well maintained and expertly operated and the quality control sufficiently strict we would have something whose performance (and probably cost) might rival a Bosch or Brosh. But we 'aint got that. We have something that cost a couple of hundred bucks that we are asking to do the same as something that cost a couple of thousand. We are trying to make a silk purse out of a sows ear, and thanks to some solid work on this forum, mainly by casainho, we are succeeding.
I can tell by some of the comments that many people realise this already and others may need a little help leveling their expectations to match where others are at. Hopefully that covers it without upsetting anyone, if it did, I am very sorry. Barring some inevitable frustration everyone seems to be having a great time coming together around this little unit!
If this was not a conversion motor restricted by bottom bracket inner diameter their would be enough space to thicken everything up to cope with the kind of corner cutting any penny pinching necessary to deliver such a capable product at such an amazing price.
When applied to it's biggest potential market the price/quality ratio makes a lot more sense, and of course it's biggest market is millions of commuters who own fairly inexpensive bikes used primarily on roads who are just desperate to pedal a bit less hard.
I personally would not suggest reducing any diameters nor drilling and pinning anything. I read the theories. NO.
2 Observations about the the expense and hassle of replicating the spline.
1. It is not necessary to have a spline.
Consider making a spline-less axle to save on cost and instead welding it to the torque sensor housing to match the existing configuration.
It is necessary for the axle to be a snug but not tight fit in drive side of the torque sensor.
Snug - Because if it is loose it will wobble
Not tight - Because if it has to be pressed in it will not be able to twist the couple of degrees each way every time the crank rotates in order to register torque from the Right hand/Drive Side Crank. This is the whole purpose of the elaborate design and why we are suffering so many breakages. if you only want torque sensing on the left crank weld that sucker up :lol: :lol: :lol:
This type of fit, snug-not-tight is generally refereed to as a 'clearance' fit in machine shops and the tolerances are best investigated elsewhere, sorry, my time is limited atm,
It would be HIGHLY advisable to have dissimilar materials for the contact surfaces between the axle and torque sensor housing because two pieces of hardened steel, which will have a similar hardness to each other can 'Gall' together (sort of like binding, or involuntary friction welding) but honestly, it ain't happening. even taking 1/2mm of each part and using a 1mm bush will weaken it too much given the failures people are posting of BOTH parts, and also, 2 items machined is 2 sets of machining costs.
My melted Nylon fix for a loose fit match up in this area is detailed about 3 pages back and is currently hanging on in after a couple or hours but I haven't had much opportunity to ride the bike lately. It fixes the play cheaply, it's permanence is very much unknown but It is such a thin sliver of material I really hope that it cannot 'flow' out the end as suggested. Who knows though.
This is another potential reason why they have made this a loose fit, if this area does gall up, the motor looses drive side torque sensing.
2. Make a piece that can be welded onto a broken axle.
Obviously use good steel heat treat as necessary and probably do not replicate the circlip groove. Use a star washer type retainer, use loctite, just don't cut a groove there, but re-use most of the existing axle, with spline.
Things to consider -
A. Cut axle down as much as necessary/possible to move the weld location away from the area of highest stress.
Area of highest stress will be on the outer edge of the most outboard journal that is effectively supporting the axle, I do not hold out much hope that this is the double (with modification) ball bearings under the lip seal unless they are a good fit on the shaft or have the shims in, however maybe the axle deflects (bends) enough that is contacting the bearing inner journal?
B. Make the replacement piece to be welded to the axle a larger diameter and find a thinner/wider bearing to carry it.
Right, that took way too long to write, I have to go. Good luck to anyone experimenting with this, If I get a chance to do more I will share my progress.
Experience: Apprenticed at a gearbox manufacture which supplied F1, Rallycross, offshore powerboats, Defence etc.
Placements of up to 3 months in milling, turning, gear-cutting, heat treat, quality control, sub-contract buying depts. Also spent a spell at another contract manufacturer to Tier 1 motorsport and watched over the chief engineeers shoulder while he did FEA and asked a load of stupid questions!
The TSDZ2 is a clever design, I'd posit that Bafang haven't attempted a bolt through torque sensor because it's hard to pull off reliably for a low cost.
If this Motor was made in Germany or Japan it would not be breaking. As someone mentioned on this forum it was originally designed in Germany and what we are buying is, in one sense or another a Chinese native knock off. Legit or not, no German owned company would tolerate these tolerances or material quality. The specs were probably marginal for heavy riders pounding across rough terrain BEFORE high quality steel was substituted for Chiniesium, vacuum quenched case hardening was supsituted for a blowtorch and a bucket of water and the tolerances were multiplied by 10.
I sympathise with everyones failures but in order to avoid unnecesary frustration we have to be honest with ourselves about what we are dicking around with here. It is NOWHERE NEAR on par with even the shimano mentioned earlier for toughness. Also the shimano problem looks unrelated to my casual glance.
A bottom bracket is not big enough to allow a sufficiently chunky axle, torque sensor housing AND a structural casting to pass through.
If the materials were (relatively) expensive, the machines that produced the items high quality, well maintained and expertly operated and the quality control sufficiently strict we would have something whose performance (and probably cost) might rival a Bosch or Brosh. But we 'aint got that. We have something that cost a couple of hundred bucks that we are asking to do the same as something that cost a couple of thousand. We are trying to make a silk purse out of a sows ear, and thanks to some solid work on this forum, mainly by casainho, we are succeeding.
I can tell by some of the comments that many people realise this already and others may need a little help leveling their expectations to match where others are at. Hopefully that covers it without upsetting anyone, if it did, I am very sorry. Barring some inevitable frustration everyone seems to be having a great time coming together around this little unit!
If this was not a conversion motor restricted by bottom bracket inner diameter their would be enough space to thicken everything up to cope with the kind of corner cutting any penny pinching necessary to deliver such a capable product at such an amazing price.
When applied to it's biggest potential market the price/quality ratio makes a lot more sense, and of course it's biggest market is millions of commuters who own fairly inexpensive bikes used primarily on roads who are just desperate to pedal a bit less hard.
I personally would not suggest reducing any diameters nor drilling and pinning anything. I read the theories. NO.
2 Observations about the the expense and hassle of replicating the spline.
1. It is not necessary to have a spline.
Consider making a spline-less axle to save on cost and instead welding it to the torque sensor housing to match the existing configuration.
It is necessary for the axle to be a snug but not tight fit in drive side of the torque sensor.
Snug - Because if it is loose it will wobble
Not tight - Because if it has to be pressed in it will not be able to twist the couple of degrees each way every time the crank rotates in order to register torque from the Right hand/Drive Side Crank. This is the whole purpose of the elaborate design and why we are suffering so many breakages. if you only want torque sensing on the left crank weld that sucker up :lol: :lol: :lol:
This type of fit, snug-not-tight is generally refereed to as a 'clearance' fit in machine shops and the tolerances are best investigated elsewhere, sorry, my time is limited atm,
It would be HIGHLY advisable to have dissimilar materials for the contact surfaces between the axle and torque sensor housing because two pieces of hardened steel, which will have a similar hardness to each other can 'Gall' together (sort of like binding, or involuntary friction welding) but honestly, it ain't happening. even taking 1/2mm of each part and using a 1mm bush will weaken it too much given the failures people are posting of BOTH parts, and also, 2 items machined is 2 sets of machining costs.
My melted Nylon fix for a loose fit match up in this area is detailed about 3 pages back and is currently hanging on in after a couple or hours but I haven't had much opportunity to ride the bike lately. It fixes the play cheaply, it's permanence is very much unknown but It is such a thin sliver of material I really hope that it cannot 'flow' out the end as suggested. Who knows though.
This is another potential reason why they have made this a loose fit, if this area does gall up, the motor looses drive side torque sensing.
2. Make a piece that can be welded onto a broken axle.
Obviously use good steel heat treat as necessary and probably do not replicate the circlip groove. Use a star washer type retainer, use loctite, just don't cut a groove there, but re-use most of the existing axle, with spline.
Things to consider -
A. Cut axle down as much as necessary/possible to move the weld location away from the area of highest stress.
Area of highest stress will be on the outer edge of the most outboard journal that is effectively supporting the axle, I do not hold out much hope that this is the double (with modification) ball bearings under the lip seal unless they are a good fit on the shaft or have the shims in, however maybe the axle deflects (bends) enough that is contacting the bearing inner journal?
B. Make the replacement piece to be welded to the axle a larger diameter and find a thinner/wider bearing to carry it.
Right, that took way too long to write, I have to go. Good luck to anyone experimenting with this, If I get a chance to do more I will share my progress.
knutselmaaster
1 kW
safeaschuck that is very accurate indeed.
The only part I disagree is that the Chinese aren't capable to do better.
I wouldn't be surprised if most parts from Bosch, Shimano, etc, are being made there too.
In my humble opinion it only is cost related.
Over there, they make what you pay for.
Making this motor 50% more expensive (to manufacture) by using better materials (bearings, axle, gears, housing) and lower tolerances would result in a motor that is a hell of a lot more reliable and resistant.
I think they believe that a higher price would make their product unattractive but I'm sure that I would be able to sell 10x as much of them in my shop (I only sell them for use cases where I can guarantee reliability, and that is not very often with the actual version of the motor).
Here in the west, people don't mind if the motor on their bike costs $800, if it is reliable.
They pay that for a replacement Bosch motor too when it fails after 10k kilometres (because Bosch won't let us repair their motors).
Making this motor the same price, but with the possibility to maintain and repair it would still leave an attractive alternative for the established big name motors because:
- Lower long term cost (because it can be repaired)
- Possibility to install it on an existing bike
- Open source
- Tweakable
TS just doesn't seem to understand...
The only part I disagree is that the Chinese aren't capable to do better.
I wouldn't be surprised if most parts from Bosch, Shimano, etc, are being made there too.
In my humble opinion it only is cost related.
Over there, they make what you pay for.
Making this motor 50% more expensive (to manufacture) by using better materials (bearings, axle, gears, housing) and lower tolerances would result in a motor that is a hell of a lot more reliable and resistant.
I think they believe that a higher price would make their product unattractive but I'm sure that I would be able to sell 10x as much of them in my shop (I only sell them for use cases where I can guarantee reliability, and that is not very often with the actual version of the motor).
Here in the west, people don't mind if the motor on their bike costs $800, if it is reliable.
They pay that for a replacement Bosch motor too when it fails after 10k kilometres (because Bosch won't let us repair their motors).
Making this motor the same price, but with the possibility to maintain and repair it would still leave an attractive alternative for the established big name motors because:
- Lower long term cost (because it can be repaired)
- Possibility to install it on an existing bike
- Open source
- Tweakable
TS just doesn't seem to understand...
barrettrussell
10 mW
- Joined
- Nov 14, 2018
- Messages
- 24
Thanks!vlvuk said:
knutselmaaster
1 kW
Indeed, I only use them for that. Not too heavy users, not too hard usage.
Not 1 problem occurred yet.
Not 1 problem occurred yet.
Retrorockit
1 kW
Unfortunately the light weight,compact size, and smooth PAS does make them attractive to off road users. I certainly wouldn't want to try to pick my way through a rock garden with my BBSHD.
Wouldn't want to carry it either!
Maybe there is a market for an ISIS splined MTB axle.
Wouldn't want to carry it either!
Maybe there is a market for an ISIS splined MTB axle.
HerskerHans
1 W
- Joined
- Apr 12, 2020
- Messages
- 56
[youtube]https://youtu.be/jmra6UoH830[/youtube]
My tsdz2 breaks off
[youtube]https://youtu.be/rdSjb88jAXs[/youtube]
I break my tsdz2 pedal axle for the second time. Dont do jumps with this motor, dont ride trails. Keep it on asphalt or gravel.
My tsdz2 breaks off
[youtube]https://youtu.be/rdSjb88jAXs[/youtube]
I break my tsdz2 pedal axle for the second time. Dont do jumps with this motor, dont ride trails. Keep it on asphalt or gravel.
romelec
100 W
HerskerHans do you have pictures ?
Where did it break and did you install the second bearing on the left side ?
Where did it break and did you install the second bearing on the left side ?
HerskerHans said:
So basically this motor is intended for utilitarian/functional use? I tried to stand on a single pedal (85kg) and nothing has happened yet. The chainring is moved 2mm outboard due to second chainring which allows the crank to be tightened to 13Nm torque only + threadlock so there might be even greater bending load on the axle.
barrettrussell
10 mW
- Joined
- Nov 14, 2018
- Messages
- 24
This may be an odd request. Is there a way I can get a photo of the soldering, wire layout for the endbell pcb? Mine has some bridging and I am just wondering if it is intentional or possibly caused by heat.
Thanks
Barrett
Thanks
Barrett
Retrorockit
1 kW
HerskerHans said:
I notice that frame has a curved downtube that puts the motor in a lower position than is possible on other bikes. Not sure if it would have made any difference or not. On my straight tube bike it's still the lowest part.
Retrorockit
1 kW
[/quote]
So basically this motor is intended for utilitarian/functional use? I tried to stand on a single pedal and nothing has happened yet. The chainring is moved 2mm outboard due to second chainring which allows the crank to be tightened to 13Nm torque only + threadlock so there might be even greater bending load on the axle.
[/quote]
Fatigue related failure happens over time. At some point, probably landing from a jump, or even off of a curb. A small crack will occur at the sharp corner of the snap ring groove ( the stress riser). From then on all the force will be focused on the weakened area and the crack will grow over time until the part fails. The shaft can be 3 or 4 times as strong as it needs to be to support your weight, but over time it can fail due to fatigue anyway. There are many things engineers and manufacturers can do to prevent this crack form forming. But most of them are beyond the average DIYer.
The true test would be a bunny hop, or drop in where the crank is twisted the maximum amount from both ends. Something that almost never happens on a commuter bike.
So basically this motor is intended for utilitarian/functional use? I tried to stand on a single pedal and nothing has happened yet. The chainring is moved 2mm outboard due to second chainring which allows the crank to be tightened to 13Nm torque only + threadlock so there might be even greater bending load on the axle.
[/quote]
Fatigue related failure happens over time. At some point, probably landing from a jump, or even off of a curb. A small crack will occur at the sharp corner of the snap ring groove ( the stress riser). From then on all the force will be focused on the weakened area and the crack will grow over time until the part fails. The shaft can be 3 or 4 times as strong as it needs to be to support your weight, but over time it can fail due to fatigue anyway. There are many things engineers and manufacturers can do to prevent this crack form forming. But most of them are beyond the average DIYer.
The true test would be a bunny hop, or drop in where the crank is twisted the maximum amount from both ends. Something that almost never happens on a commuter bike.
HerskerHans
1 W
- Joined
- Apr 12, 2020
- Messages
- 56
Retrorockit said:
Yeah, with some awareness it's fine that it Hangs low. I also noticed that my 750w tsdz2 48v 624wh is slow and weak compared to a 250w brose or 250 shimano steps 8000 motor in boost. That surprises me a bit, i tested my friends bikes. My other tsdz2 motor had the same power, felt the same, the one that broke. I think the 750watt rating is very optimistic. My 500 watt bafang m600 feels twice as strong.
Retrorockit
1 kW
My 1500W BBSHD says 750W on the decal. It's just matter of tuning. So it's hard to say what's really going on sometimes. There's also the matter of TSDZ2 backing off power above 80rpm and other nonsense. Caisanho's tuning may find the missing power. But more power can lead to cooling and rubber gear issues.
I'm trying to figure out what load is breaking the shaft. At first I thought it was bending from the riders weight at the bottom of the stroke. But MTB riding adds the possibility that it's torque from landing with the cranks flat. Something I do on the street occasionally. A new offroad shaft may need to be designed as a torsion bar.
I'm trying to figure out what load is breaking the shaft. At first I thought it was bending from the riders weight at the bottom of the stroke. But MTB riding adds the possibility that it's torque from landing with the cranks flat. Something I do on the street occasionally. A new offroad shaft may need to be designed as a torsion bar.
izeman
1 GW
Don't mix up nominal and peak wattage. That's all about legal stuff. EVERY pedelec motor available today BOOSTS around 600-900W. Still they are 250W nominal.HerskerHans said:
The 750W of the TSDZ2 are PEAK, and with this tiny cheap motor construction this is really PEAK for a SHORT time.
Bosch, Yamaha etc can do that for a very long time.
Has anyone tried oil-cooling the motor and controller yet? Looking at pictures of the guts, I wonder if you could seal around the motor and fill the cavity with light mineral oil? The case would have to seal stongly - it has an o-ring but I wonder how much it flexes during use? You'd also need a filler/sump and breather holes - plus the downside would be how messy it would get if you needed to take it apart! 
Retrorockit
1 kW
There are people doing that with overvolted hub mtors. Either ATF, or metallic colloidal suspensions (ferro fluids) that cling to the magnets. But they say the 10x higher RPM of the mid motors makes it less practical. There is a lot of improvement to be had already by filling the air gap between the cover and stator with metal plates and TIM , or even just TIM pads. One difference is that the magnets on a hub motor are on the outside and hold the ferro fluid in a useful position.
Liquid cooling isn't "better" than air cooling. The heat has to be transfered to the air at some point. If you have air cooling with fins and air flow it's just as good until you need more surface area, then you get into pumps and radiators and fans. Heat is energy. It's coming from your battery.Overcooling has a price also.
Liquid cooling isn't "better" than air cooling. The heat has to be transfered to the air at some point. If you have air cooling with fins and air flow it's just as good until you need more surface area, then you get into pumps and radiators and fans. Heat is energy. It's coming from your battery.Overcooling has a price also.
Retrorockit
1 kW
ATF is thinner, and has Anti foaming properties. If you want to do this you can clean the grease out of the roller/sprag clutches also. ATF is the perfect lube for those. ATF has become the normal lube for a lot of manual transmissions, and even the final drive on fwd cars. Adding fins to the outside of the cover will be useful just as with air cooling. The drag of the rotor in the ATF will tend to heat up the fluid. Leaks will probably happen. But I'm all in favor of someone else trying it.
Interesting! I had missed the experiment by NexusG, but I wasn't suggesting filling both sides of the casing, just the motor-side which houses the controller as well. This side looks pretty much sealed, but the other side has the rotating shaft coming out of it. Improving conduction/convection of heat away from the shell of the motor and power electronics by using a higher-density medium should improve short-term overheating but wouldn't be as useful in the longer-term. Unless you add a radiator etc. which as you say comes at a price.
:lol: Me too! Maybe when the next motor fails?!?
Retrorockit said:
:lol: Me too! Maybe when the next motor fails?!?
Retrorockit
1 kW
Somewhere there is a shaft from the motor into the gearcase. It probably won't be sealed because there was just air in both sides. You might ask NexusG what happened there. Who knows may be you'll end up with an automatic drip chain oiler.
Similar threads
