Torque sensor performance

[Ravi Kempaiah]

Hey JB riders,

I was wondering if you could provide some feedback on the torque sensor used in Juiced Bikes.

From your experience, how has been the torque sensor performance?

Has it been 100% reliable?
Did you have to reset it anytime?

How is the response time and sensitivity range at different cadence?

Thanks

 

[Captain Slow]

Torque sensor has worked great for me. Has been 100% reliable, no resets ever required. I actually wouldn't know how to reset it.

Feels completely natural, no hint of lag at all. Can't say I've paid attention to performance at different cadences, but I would say I've never noticed anything unusual and I have over 7,000 km's on the bike now.

 

[teskow]

My RC torque system has worked flawlessly. I have never done anything with it with in over 1000 miles. From time to time I will try the cadence setting but go right back to the torque sensing mode. There is on slight glitch though and that is when hitting a sharp bump while peddling it will surge. I understand that is a characteristic of the torque sensing system. I just stop peddling when I am about to encounter a sharp bump and that works to prevent the surge.
Riding in cadence mode seems so unnatural to me. I would never buy a bike with only a cadence system.
The clearance of the sensor is close tolerance so I blow it out occasionally with compressed air for my satisfaction . Probably not necessary though.

 

[smorgasbord]

Honestly, going from a cadence-sensing hub motor bike to a torque-sensing mid-drive motor bike wasn't as big a change as I was expecting - at least for on-road riding. The Bafang Ultra torque-sensing motor is way more powerful, but I have to admit that the R&M SuperDelite I test rode felt more natural than either of those other bikes.

I'm not 100% sure what's going on. The Bosch-powered R&M just made me feel stronger while on the other two bikes I can feel them helping me along more. Now, maybe that's because the R&M is under-powered and so one has to pedal harder in the first place, and that translates into a more natural feeling (less assistance is a more natural feeling kind of by definition, eh?). I didn't have a lot of time on the SuperDelite test ride, but I did get to take it up the 4th steepest hill in San Francisco. I have about 500 miles on the cadence-sensing RadCity, and only 150 miles on my new mid-drive bike.

I need more riding time, but where the torque sensing bikes should shine is off-road, where the terrain is not only more varied, it varies more quickly. For on-road with cadence-sensing, if the bike is giving you too much power, you can choose a higher gear to slow down your pedal cadence and reduce motor contribution - or conversely choose a lower gear to pedal faster and get more power out. And that's all within the same PAS level, which, of course, you can adjust as well. At some point I think these adjustments become instinctive and then there's really nothing wrong with a cadence sensing bike - at least on-road.


BTW, not to drift off-topic, but I do believe that the tuning of the Bafang Ultra motor could be improved, especially at low PAS levels. The thing is so damn powerful that even at PAS level 1 it's giving me too much assistance on level ground unless I want to be riding 25MPH or more. I did buy a programming cable and may tweak max amperage settings on the lower PAS levels. It would be nice to have two sets of settings: one for on-road and one for off-road.

 

[highpockets]

Feedback: Never thought about it until you asked. Must be working great!

 

[Johnny]

Torque sensor has worked great for me. Has been 100% reliable, no resets ever required. I actually wouldn't know how to reset it.

Feels completely natural, no hint of lag at all. Can't say I've paid attention to performance at different cadences, but I would say I've never noticed anything unusual and I have over 7,000 km's on the bike now.

Captain may I ask if it feels jittery,on/off when the chain bounces on uneven terrain? Another issue is how easy to torque the rear wheel back after fixing a flat tire? These sensors tend to cause problems if the wheel is not torques right.

 

[Ravi Kempaiah]

@highpockets , @Captain Slow , @teskow , @smorgasbord ,

thank you for the response, gentlemen!

Kudos to Tora/team in China to getting this sensor to work at a very reasonable price point.

 

[Captain Slow]

Captain may I ask if it feels jittery,on/off when the chain bounces on uneven terrain? Another issue is how easy to torque the rear wheel back after fixing a flat tire? These sensors tend to cause problems if the wheel is not torques right.

I haven't paid attention to that, but I would say I hadn't noticed a jittery on/off. I've had the chain drop many times, it's really annoying.

I find it a huge pain to put the rear wheel back on after doing maintenance. But I've never had a problem with the torque sensor after removing and putting the wheel back on. I hate putting the rear wheel back on with a hub motor. So much so that I went out and added a Creo to my stable.

 

[Ravi Kempaiah]

I hate putting the rear wheel back on with a hub motor.

If you had a thru-axle quick release rear hub bike with the same performance as Juiced and reliable torque sensor, would it change your opinion?

 

[Captain Slow]

If you had a thru-axle quick release rear hub bike with the same performance as Juiced and reliable torque sensor, would it change your opinion?

Absolutely! I live in Vancouver so I had been eyeing one of the Grin motors with a thru axle and torque arm. But I'd rather not build from a kit. I'd rather buy something factory.

One thing I dislike about my CCS is all the wires. My Creo doesn't have a single wire from the electric system. The CCS has wires all over. But hey the CCS is a lot cheaper, but to me having few or no wires is worth paying more money. Maybe not as much as the Creo, but there are other reasons I spent way more than I ever thought I would.

 

[pennybags]

Kudos to Tora/team in China to getting this sensor to work at a very reasonable price point.

They are actually using a part that’s very common. TDCM has a catalog of parts that everyone uses. As long as you have the frame mount, it’s ready to go.

Tora absolutely has brought down the price point for his bikes, but realistically it’s gone from ~ $20 to ~ $16 in bulk pricing.

 

[Explorer-1]

The software has as much to do with the torque sensor as the sensor itself (probably much more). I'm an outlier; I like my cadence sensor bike just as much as my RCS torque sensor and often I like it better. I need more miles an much longer trips on my RCS, but I certainly would not hesitate to buy a cadence only sensor bike if I liked it based on my previous experience. I have not yet tried turning off the torque sensor on the RCS, but I'll try that one day to see if I prefer it on the RCS. I can notice a vibration in the motor on the RCS at certain speeds/cadences/assist levels, it's small but I've felt more strange motor feedback on the RCS than my cadence bike which has never produced any type of unexpected feedback I could feel in the pedals. I wish I could tune the RCS for the right assist for each torque level and have the assist levels switch between different levels for the same cadence. It annoys me when often I increase the assist level and nothing happens unless change gears or cadence. The programming on this bike is not what I would do, but overall so far it's a nice bike with it's own unique quirks.

 

[Ravi Kempaiah]

They are actually using a part that’s very common. TDCM has a catalog of parts that everyone uses. As long as you have the frame mount, it’s ready to go.

Tora absolutely has brought down the price point for his bikes, but realistically it’s gone from ~ $20 to ~ $16 in bulk pricing.

TDCM doesn't make torque sensors. I think you mis-read my Stromer technical report.
None of the budget companies use Torque sensors i.e., Rad Power, Aventon, etc.

The torque sensor on the juiced bikes was developed by a different company and Tora worked diligently to bring in to the market. It takes a lot of trial and error to get it right and make it work reliably for thousands of bikes.

 

[linklemming]

I have a CCX as well as 2 Brose Motor bikes (1 speed/1 normal motor).

As far as low speed feel/response, the CCX isnt even close to the telepathic feel of the Brose. The CCX acts more like a crude on/off switch at low speeds, especially with the annoying brake motor cutouts with huge delays.

That being said, at the higher speeds the CCX was designed for, the torque sensor is just fine. Still not as telepathic as the Brose but it becomes a moot point.

 

[Bruce Arnold]

Torque sensor has worked great for me. Has been 100% reliable, no resets ever required. I actually wouldn't know how to reset it.

Feels completely natural, no hint of lag at all. Can't say I've paid attention to performance at different cadences, but I would say I've never noticed anything unusual and I have over 7,000 km's on the bike now.

Pretty much says what I was going to say, except I've got around 5500 kilometers.

 

[Ravi Kempaiah]

I have a CCX as well as 2 Brose Motor bikes (1 speed/1 normal motor).

As far as low speed feel/response, the CCX isnt even close to the telepathic feel of the Brose. The CCX acts more like a crude on/off switch at low speeds, especially with the annoying brake motor cutouts with huge delays.

That being said, at the higher speeds the CCX was designed for, the torque sensor is just fine. Still not as telepathic as the Brose but it becomes a moot point.

I have to agree.

I think the holy grail of superb E-biking experience is a motor that is rated for 500-700 watts continuous power output (Brose, Shimano, Bosch have 350 watts nominal, ~600 watts peak) BUT has Brose-like sensitivity and Yamaha-like reliability..

If such a motor can be made, that would be ideal for a vast majority of use cases and be an absolute delight to ride. Current offerings from Bosch or Brose are good but they can't hold speed like Juiced CCX but a slightly beefier motor would be able to do that nonchalantly.

 

[smorgasbord]

I think the holy grail of superb E-biking experience is a motor that is rated for 500-700 watts continuous power output (Brose, Shimano, Bosch have 350 watts nominal, ~600 watts peak) BUT has Brose-like sensitivity and Yamaha-like reliability.

I think it's hard to have a do-it-all motor/setting.

If you look at Bosch's Gen4 Performance Line offerings - they have a Speed and a CX version. They both max out at 75Nm of torque, but that torque occurs at different PAS and speed levels. Here's a comparison:

Note that the Speed version only gets to 75Nm in TURBO PAS while the CX version gets there in both eMTB and TURBO. What is also happening is that the SPEED version only gets to 75Nm when going very fast, whereas the CX version gets there at lower speeds (well, it has to since its speed tops out at 20mph). At less than 20mph speeds, the SPEED version is not putting out 75Nm of torque. What's clear (to me anyway) is that the CX version is designed to get you more Nm at lower speeds, because that's what you need to climb off-road, whereas the Speed version uses the extra Nm to help you get up to and maintain speed. Same torque ratings, but for different use cases and so at different motor profiles.

Now the big question is whether there are in fact mechanical differences in these two motors, or whether it's all programming. At any rate, those that think with the Gen 4 that they don't need the CX because the Speed version has "the same torque" are, in fact, missing when that torque comes into play. Putting a Speed motor on your off-road bike because you also want to go fast on-road means you'll be sacrificing climbing abilities.

I believe this puts the Bafang mid-drive and, in fact, any hub-drive motor capability into perspective. At what speed do you want the high torque?

This is interesting to me, because in the Bafang Ultra programming thread, I'm trying to decipher how those settings work. People have been at it for years now, and we're all still guessing at many of the parameters. I was surprised to see that the different torque settings are not tied to PAS, but are tied to crank rotation speed - and that's regardless of the PAS setting! That itself may be a limiting factor for how natural we can make the Bafang Ultra feel. The PAS settings are just for limiting maximum current - you can't change the torque multiplier for different PAS settings.

Sorry if this has drifted off topic, but I do think it's relevant to the general topic of "torque sensor performance."

 

[Ravi Kempaiah]

They both max out at 75Nm of torque, but that torque occurs at different PAS and speed levels.

Thanks for your comments.
I concur with you that CX motor is designed for eMTB and it makes sense that higher torque is available at low speeds.
From that graph it is really hard to comment on PAS and speed levels. We can infer something valuable if we have a torque-speed-cadence curves.
The only reliable data I have found so far is when Grin tech does testing or some sort of Dyno testing.

Re: Bafang ultra, while it is a powerful motor, it is not as refined as Brose or Yamaha. The inherent limitation comes from the electronics/sensor setup and I would love it if an ultra motor can be made as smooth as Bosch or Brose. I will look fwd to experiencing that.

 

[Johnny]

The table only shows the maximum torque on different support levels and without the cadence information it is pretty much useless.

The graph that shows you what a motor is capable of is the power (or torque ) vs cadence graph and major brands don't give that information. In mountain biking riders do start or stop many times hence power available at low cadences matter so maybe CX can supply more torque (hence power ) at low cadences.

The explanation about "more torque at higher speeds" is not meaningful at all. It is just a simple formula anyways cadence x torque = power and speed has no place in this formula.

 

[vincent]

Well let’s hope that excess controller is going to do a lot to smooth out the ultra motor!

hopefully we get some feedback on that soon