Connecting a Torque Sensor to a Controller

[Beagle123]

Hi:

I own one of these charger bicycles:

http://abc.eznettools.net/D300013/X300109/ebike1.html

I'm converting it from 24v to 36v. I'm only salvaging the motor and torque sensor. Everything should work fine except the new 36v controller uses a standard 3-wire throttle, and the torque sensor that 's attached to the chain of the bike doesn't seem compatible.

The problem is that the twist throttle puts-out a voltage of 0.4v to 3.5v when attached to a 3.9v battery. The torque sensor puts-out a voltage of 0.4v to 0.7v with the same 3.9v battery. So if I just connected the controller to the torque sensor, it would only give a slight signal to the controller. I want the torque sensor to produce the full-throttle 3.5v when pressed hard.

Is there a way to make the sensor output more voltage? Is there a way to make the controller use the lesser voltage? Do I need a custom circuit? Suggestions?

Thanks,
Eric

 

[amberwolf]

how does that torque sesoro work? the THUN type ebikes.ca uses runs on 10v i think it is, and i forget the output data format.

 

[Beagle123]

This torque sensor works by having the tension in the chain press down on it making the little arm bend. When it is on the bike, the chain rests on the top of the little sprocket on the top of the sensor. When you put pressure on the pedal, the chain straightens, putting downward pressure on the sensor. This flexes the arm which changes the voltage.

It has three wires:

black (ground)
red (positive)
yellow (speed signal voltage)

The problem is that the range of the speed signal voltage is very limited. It ranges from 0.4v to 0.7v. I need it to range between 0.4 and 3.5v.

Eric

 

[d8veh]

It may have been working off battery voltage (24v), which would give a range of 2.4v to 4.3v. The problem you now have is that the bottom value has to be below 1v, which is about the start point for the new controller to give power from the throttle, otherwise your new controller won't initialise, and if it did, you'd have power all the time.

Your old controller will probably be OK at 36v, but if the torque sensor does work off the unregulated battery voltage, you'll have the same problem with the low output being too high. If you can check the supply voltage to the sensor with your old set-up, it might give you a clue to what's happening. If the supply to the sensor is regulated, you should be OK with your old controller and new battery.

 

[emiyata]

This torque sensor works by having the tension in the chain press down on it making the little arm bend. When it is on the bike, the chain rests on the top of the little sprocket on the top of the sensor. When you put pressure on the pedal, the chain straightens, putting downward pressure on the sensor. This flexes the arm which changes the voltage.

It has three wires:

black (ground)
red (positive)
yellow (speed signal voltage)

The problem is that the range of the speed signal voltage is very limited. It ranges from 0.4v to 0.7v. I need it to range between 0.4 and 3.5v.

Eric

Can you still buy this load cell? My guess is that the cell runs off electronics that amplifies the voltage difference. I would be very surprised if the electronics don't run off regulated voltage. Thanks for sharing the pictures.

 

[-dg]

I have this bike too, I replaced the SLA batteries with an 8 cell Ping, so 26V 20ah. I also broke the frame where the rear stays come into the seat post. It should be triangulated like a normal bike otherwise it flexes at the hard spot there and cracks the seat tube. I had it repaired and added additional stays, I'll post a photo some time.

As for the torque sensor, I believe it is a strain gauge and according to wikipedia these are normally read with a wheatstone bridge. Possibly the bridge is already part of the part that mounts on the chain stay. I think a custom circuit is going to needed with a micro is probably needed anyway to smooth out the torque variation as you pedal.

Is your original controller still good? It might be better to mod that. Also, brushed motors are often not happy at significantly higher than designed voltages, 36 may be pushing it.

What is your goal for all this anyway, to repair a broken bike, or to hot rod a working one, or what?

 

[Beagle123]

Thanks for your responses:

d8veh:

Your old controller will probably be OK at 36v, but if the torque sensor does work off the unregulated battery voltage, you'll have the same problem with the low output being too high.

I agree I think the minimum voltage would go too high. RIght now the minimum voltage is perfect. Both the throttle and the torque sensor have the same bottom voltage of 0.4v. I may try running a higher voltage through it just to see what happens.

emiyata:

I would be very surprised if the electronics don't run off regulated voltage.

I think you're right. The bike must have a custom circuit for the controller that knows how to handle a throttle range of 0.4v to 0.7v.

All these electronics are built into one big component that fits in the middle of the bike's frame. The batteries, controller and charger are all in the big triangular box. Click on the link at the top to see a picture of the bike.

dg:

I believe it is a strain gauge and according to wikipedia these are normally read with a wheatstone bridge. Possibly the bridge is already part of the part that mounts on the chain stay. I think a custom circuit is going to needed with a micro is probably needed anyway to smooth out the torque variation as you pedal.

I will research this. Does anyone think this sensor should be wired differently than the throttle?

Is your original controller still good? It might be better to mod that. Also, brushed motors are often not happy at significantly higher than designed voltages, 36 may be pushing it.

The original controller is still good, but its a part of the huge component mounted in the center of the frame (that I eliminated.
I wasn't aware there was a problem going from 24-36 volts. I'm sure I'll fry this little motor. It's pretty small, and it has no vents on it. I'll be running about 500 watts through it and cross my fingers

What is your goal for all this anyway, to repair a broken bike, or to hot rod a working one, or what?

Both. It was a weak bike, then the buttons broke and the batteries went dead. I didn't see any reason to repair it, so I'm trying to make a stronger bike out of it. Also, this bike is hard wired to not to exceed 18 mph. It sucks. As soon as to get up to 18mph, the motor stops. So, I junked all the electronics.

So, the consensus is that I need to make a circuit to amplify the voltage right? I'm coming to that conclusion myself. I'm researching opamps. I think there is a pretty easy solution there. If anyone knows about electronics, it would be very helpful. I'm going to take a stab at it now.

Thanks!!!
Eric

 

[-dg]

Is your original controller still good? It might be better to mod that. Also, brushed motors are often not happy at significantly higher than designed voltages, 36 may be pushing it.

The original controller is still good, but its a part of the huge component mounted in the center of the frame (that I eliminated.
I wasn't aware there was a problem going from 24-36 volts. I'm sure I'll fry this little motor. It's pretty small, and it has no vents on it. I'll be running about 500 watts through it and cross my fingers

What is your goal for all this anyway, to repair a broken bike, or to hot rod a working one, or what?

Both. It was a weak bike, then the buttons broke and the batteries went dead. I didn't see any reason to repair it, so I'm trying to make a stronger bike out of it. Also, this bike is hard wired to not to exceed 18 mph. It sucks. As soon as to get up to 18mph, the motor stops. So, I junked all the electronics.

So, the consensus is that I need to make a circuit to amplify the voltage right? I'm coming to that conclusion myself. I'm researching opamps.

The emiyata quotes are actually -dg.

Anyway, the 18mph thing is easy to fix. There is a pickup, like a bike computer speed sensor, on the left chainstay that reads a spoke magnet on the rear wheel. To unlimit the speed, get a strong small magnet and mount it to the inside of the left crank arm and then move the pickup to read the magnet on the crank instead of the wheel. This fools it that you are going slower than you are. My magnet is just magneticaly stuck to the end of the pedal axle, simple as can be.

The problem with the op-amp instead of a micro is that pedaling torque is not smooth. When the cranks are horizontal there is a lot of torque, when they are vertical there is almost zero torque. So just amplifying the sensor is going the make the power vary from max to zero twice for each revolution of the pedals. This does not sound like a fun ride, and it also does not feel like the original bike so I think the controller was filtering and smoothing the signal somehow.

These bikes are frustrating because in some ways they are really crappy bikes and in some ways they are really excellent engineering. The torque sensor and control are one of the excellent parts as is the drive mechanism (see the GNG mid-drive thread for something similar done worse). The actual bike is not anywhere near as good as a typical steel mountain bike from that period, and there are some complete bloopers. But then it was designed by aerospace engineers, not bike people.

The batteries suck, but SLA was all there was then and brushless motors were not really available either so given the time period they got the power vs range balance pretty good.

 

[-dg]

Looking at your posted photos of the sensor, you have 5v, ground, signal coming into the first blob and then 4 wires going to the actual strain gauge element. So it looks like the detection/amplification is already done on the device so whatever you need to read it only needs to deal with the scale it presents. This makes it simple. An 8 pin AVR and some caps is probably all you need. That rest is a simple matter of programming.

The reason you can't increase the voltage on brushed motors as I understand it is that the brushes fail quickly. The windings should not be a problem. If I were doing this, I'd be looking for a brushless motor that would fit in the same mounting and use a brushless controller then you could run the voltage up. But, with the brushes, it might be best to keep it to around 30 volts.

Btw, the 20ah Ping pack is much stronger than the SLA original batteries, so you may not really need much more voltage to get more power. I'd try to keep the original controller, replace the buttons (which should be easy), and use 10ah of 8s lipo for around 30 volts.

The big thing in the middle is actually one of the better parts of the bike given that they had to fit two hulking heavy SLA batteries and not totally screw the handling. If you are using RC lipo there is more flexibility.