2023 Trek Fx+ and Dual Sport+

[Gee_Whiz]

Taking in the account the short distance of your commute, the FX+ 2 is certainly a good option!

He meant Vado SL felt like a normal, not heavy bike. Easier to pedal. Actually what happens with Vado SL past 25 km/h is your thighs feel heavier but pedalling is so easy I often overlook going past the speed limit! I only cannot pedal faster than 27 km/h.


Impossible!

Stefan nailed it. FX+2 you can pedal off motor with a bit of effort, although it rides well in PAS0 or Off.. it doesn't feel like a regular bike imo (or it feels like a heavy MTB). It can be done however and its kind of fun/challenging and its not tough to do so. The Vado SL feels like a normal bike and even a light normal bike somehow and pedals FAST off motor.. slight inclines, uphill etc. I've ridden the FX+2 uphill with no motor and it climbs well due to gearing but just takes a bit more effort due to frame weight. But yes the FX+2 is a great buy and even the DualSport if you dont need the rear rack and extra weight.

 

[Richard Stallard]

There is a very recent review of the Fx+ Stagger (mid step) version on EBR. Unfortunately, the reviewer doesn’t comment on the smoothness of the motor boost in terms of effectiveness of torque sensor.

 

[tegnamo]

I test rode an FX+ (or maybe Dual Sport+?) at a Trek store and the delay between pedaling and getting assist made me think the torque sensor wasn't even working. It was pretty unpleasant after having ridden mid-motors.

 

[Richard Stallard]

There is an accompanying video review as well and the reviewer made a similar comment.

 

[Gee_Whiz]

I test rode an FX+ (or maybe Dual Sport+?) at a Trek store and the delay between pedaling and getting assist made me think the torque sensor wasn't even working. It was pretty unpleasant after having ridden mid-motors.

There isnt a torque sensor unfortunately, and this should really be addressed with Trek, as its literally false advertising

 

[Richard Stallard]

What is the source of your information? Trek spare parts listing indicates a torque sensor in the bottom bracket.

 

[Gee_Whiz]

What is the source of your information? Trek spare parts listing indicates a torque sensor in the bottom bracket.

Pedaling the bike you can discern its a cadence sensor installed vs the listed torque sensor. There is no natural motion to the motor engaging; you actually noted this first in your initial review

 

[Richard Stallard]

I found that simply turning the pedals was not sufficient to engage the motor. Bikes with cadence sensor only typically apply motor power whenever pedal rotation is detected even if no force (torque) is applied. I had to apply some torque but just a small amount of torque would engage virtually full power at the selected boost level.

My conclusion is that there is a torque sensor but the motor control algorithm does not make optimum use of it.

 

[Gee_Whiz]

I found that simply turning the pedals was not sufficient to engage the motor. Bikes with cadence sensor only typically apply motor power whenever pedal rotation is detected even if no force (torque) is applied. I had to apply some torque but just a small amount of torque would engage virtually full power at the selected boost level.

My conclusion is that there is a torque sensor but the motor control algorithm does not make optimum use of it.

In my experience the sensor on this bike, operates identically to other cadence sensing bikes ive owned and tried

 

[Richard Stallard]

In my experience the sensor on this bike, operates identically to other cadence sensing bikes ive owned and tried

That could well be the case as most people would naturally pedal and apply some torque, whereas I deliberately pedalled with zero torque to test the effect.

 

[Stefan Mikes]

I think the issue here is not "torque sensor or not" but the motor/system does not use the torque sensor to proportionally increase the assistance but simply provides the full power available for a given assistance level as soon as torque on the crank is detected.

 

[Richard Stallard]

That is my conclusion. It could also be that the control system is “proportional” but the “gain” associated with the torque sensor input is way too high so that a small torque value results in close to full motor response.

The Hyena web site indicates that their torque sensors do produce an analogue voltage signal proportional to applied torque, but that is of no benefit if the control system doesn’t respond properly. See https://www.hyena-ebike.com/sensor/

Now that others are reporting similar behaviour of the Hyena drive, we at least have confirmation that my wife’s bike is not faulty and is operating “as designed”, even though it seems the design itself is not very good. Had we not had the opportunity to compare the Fx+ with quality mid-drive bikes (Vado SL and Trek Verve+ 2) we probably would not have been so critical.

My wife actually loves the bike even with its quirks due to the fact that it is fairly light and rides like a non-ebike. If it was for me, I would have returned it under the Trek 30 day satisfaction guarantee.

 

[Gee_Whiz]

I think the issue here is not "torque sensor or not" but the motor/system does not use the torque sensor to proportionally increase the assistance but simply provides the full power available for given assistance level as soon as torque on the crank is detected.

Its moreso imo.. that the Vado SL for example.. you pedal a half-turn or full, and you immediately feel a semblance of boost 100% of the time that you pedal whether its 1% or 100% power.. there is no point in which you feel the motor isnt working if youre in PAS 1+.

On these bikes, there is a delay when you pedal, and you get 0% power until a partial rotation of the pedal, similar to a cadence sensing bike. There seems to be a sensor somewhere near the pedal or bottom bracket, that tells the bike when to apply power based on whatever your Hyena (or default) settings are.

So its moreso when the power is being applied (should be a constant 1-100% in PAS 1-3, up to 20mph) vs how much power is being applied overall.

 

[Captain Slow]

Richard thanks for posting your experience with this bike. I know there are lots of debates about hub vs. mid drive. The only thing I really care about with hub drives Is how easy/hard is it to remove the rear wheel. If you get a flat when you’re out on the road, are you going to be able to repair it easily. It sounds like it shouldn’t be a problem with this bike.

 

[Richard Stallard]

Removing the rear wheel is certainly a little more complicated than usual, but not too bad once you know what to do. The rear hub has hex head bolts either side, with bushes to match the frame dropouts which prevent the axle from rotating under motor torque. These bolts require a long 5 mm hex key to get sufficient leverage. Prior to wheel removal, a plug and socket must be disconnected just in front of the hub motor and a bolt removed which clamps the motor cable to the frame. There was also a cable tie on the motor side of the plug and socket, which I had to cut off. I didn't replace the cable tie as the cable was fixed well enough to not get caught in any moving parts.

When re-installing the rear wheel, make sure the plug and socket are fully engaged. It took me a couple of goes to get it right.

The front hub is described as QR and appears to be a standard QR type but, instead of a QR skewer, it has long bolt with a 5 mm hex socket to fit the thru hole forks, which are non-standard.

It would pay to practice removing the wheels before riding far from home base.

The diagram below of the rear hub is from the Trek Maintenance Manual - See https://retailerassetsprd.blob.core...=UogJIteiFltPX66np2M0a3esSu1uZzABYHTFInUlT/o=

 

[hud95404]

There is a very recent review of the Fx+ Stagger (mid step) version on EBR. Unfortunately, the reviewer doesn’t comment on the smoothness of the motor boost in terms of effectiveness of torque sensor.

I've just seen it, thanks for the info!

 

[Johnny]

I found that simply turning the pedals was not sufficient to engage the motor. Bikes with cadence sensor only typically apply motor power whenever pedal rotation is detected even if no force (torque) is applied. I had to apply some torque but just a small amount of torque would engage virtually full power at the selected boost level.

My conclusion is that there is a torque sensor but the motor control algorithm does not make optimum use of it.

Imo, the problem you describe is a result of torque sensor not being at the motor axle. If you measure torque at the axle the controller can immediately modulate motor torque accordingly.
In this system, gear ratio is not known, that is the controller does not know which cog your derailleur is at, hence can not compute the translated torque just from the proportional signal from the bottom bracket torque sensor. In this case the only way to measure the torque applied to the motor axle is to measure rider power first which also requires cadence measurement.

Hence the controller is waiting for a cadence measurement to measure power and compute the translated torque at the motor axle and you observe a latency very similar to cadence only systems when you start from a standstill.

This is not needed when the torque sensor is right next to the hub like a tmm4, or in mid drive systems where the sensor is at the bottom bracket which is directly connected to the motor axle.



I like the bike overall though, it looks like a good bike for riding on the pavement.

 

[fooferdoggie]

Imo, the problem you describe is a result of torque sensor not being at the motor axle. If you measure torque at the axle the controller can immediately modulate motor torque accordingly.
In this system, gear ratio is not known, that is the controller does not know which cog your derailleur is at, hence can not compute the translated torque just from the proportional signal from the bottom bracket torque sensor. In this case the only way to measure the torque applied to the motor axle is to measure rider power first which also requires cadence measurement.

Hence the controller is waiting for a cadence measurement to measure power and compute the translated torque at the motor axle and you observe a latency very similar to cadence only systems when you start from a standstill.

I was wondering when I heard about the delay on some bikes.

 

[Richard Stallard]

The main problem isn’t the delay. It is that the motor power is constant and not in proportion to rider power. The bottom bracket torque sensor also measures cadence so it should perform better than it does.

Ignoring start-up, once the bike is moving the controller should be able to compare road speed and cadence to deduce the gear ratio in order to determine the required motor power.

Has anyone ridden the Electra bikes with Hyena drive? I wonder if they are better or the same?

 

[Johnny]

The main problem isn’t the delay. It is that the motor power is constant and not in proportion to rider power.

How did you come to this conclusion? What is your method to measure the amount of power motor supplied vs rider input?

The bottom bracket torque sensor also measures cadence so it should perform better than it does.

Those cadence measurements has a significant delay that is the problem. These bottom bracket sensors are not the best for power measurement. Good power meters cost a lot and even they may have delays...

Ignoring start-up, once the bike is moving the controller should be able to compare road speed and cadence to deduce the gear ratio

Not that simple. Since there is freewheeling involved that ratio will not always be equal to the gear ratio. Riders coast, may start/stop pedaling or change gears and in each of these cases you will need to look into a significant window to re-estimate that ratio. This once again means a significant delay...