Wisper Faulty Pedelec sensor

[z0mb13e]

I think it is about time I fixed the pedelec sensor on my Wisper.

Does anyone know exactly what sensor is embedded in the crank housing? I figured it is a hall effect transistor as it has 3 leads, but I want to take it off the bike and test it on the bench so it would be good to know the transistor specs (i.e. is it NPN or are all hall effect transistors NPN? Anyone know the model number so I can look for a datasheet?) before hooking it up to test it...

Edit: When I say "what sensor is embedded in the crank housing" I mean embedded in the sensor housing that is attached to the mounting ring that is held on by the threaded ring on the bottom bracket...

 

[z0mb13e]

Ok a quick google answered the "are they only NPN" question. The answer is no. There are PNP hall effect devices too...

 

[Deleted member 4366]

if you have a disc with a ring of magnets in, then the sensor is a hall sensor that pulses the output on the signal wire. There's different ways that they pulse depending on which hall sensor and which magnet arrangement is used.
if yours doesn't work, check these things:
The magnet ring should be very close to the sensor.
The magnet ring should be the right way up. Did you or anybody else ever take it off?
All the connector pins need to make contact with the controller.
All the wires in the connectors need to be in the right sequence so that the colours match from side to side.
I've seen one where the wire had been pulled and was broken inside the insulation.

You can use any 5v supply to test the sensor, but it pulses quite quickly, which makes it difficult to check. you need to turn the crank very slowly to see the signal wire go alternately high and low

 

[z0mb13e]

if you have a disc with a ring of magnets in, then the sensor is a hall sensor that pulses the output on the signal wire. There's different ways that they pulse depending on which hall sensor and which magnet arrangement is used.
if yours doesn't work, check these things:
The magnet ring should be very close to the sensor.
The magnet ring should be the right way up. Did you or anybody else ever take it off?
All the connector pins need to make contact with the controller.
All the wires in the connectors need to be in the right sequence so that the colours match from side to side.
I've seen one where the wire had been pulled and was broken inside the insulation.

You can use any 5v supply to test the sensor, but it pulses quite quickly, which makes it difficult to check. you need to turn the crank very slowly to see the signal wire go alternately high and low

Hi d8veh,

I should have said, I tried all the obvious things, the disk is the right way round, the wires all look ok, and I have replugged the connectors etc. I want to determine if the sensor is working before turning my attention to the controller.

So if I can hook the sensor up to 5v, what about current limiting or is the sensor an integrated device (i.e. a hall effect transistor with an integrated circuit)?

I want to take the sensor off the bike to test so I can hook it up to an oscilloscope so signal speed isn't an issue.

 

[neptune]

@d8veh . From the information in your last post you obviously have a lot of knowledge about these things . I worked on washing machines for many years and so I know a bit about fault finding and multimeters , but I am new to electric bikes . So , faced with the above problem you have three wires . The problem is that you have to work out which wire does what ,unless you have a circuit diagram , which is unlikely . So , would it work if you did the following . Unplug the three wires . Test the three wires coming from the controller with a voltmeter [and the bike switched on ] and find which 2 wires have a voltage of 5 volts between them . The remaining wire is then the signal wire . Find a way to disconnect the signal wire only , leaving the supply wires connected to the sensor . Connect an analogue volt meter between the negative supply wire and the signal output of the sensor. When the pedals are turned , analogue meter will swing between zero and 5 volts .Is this a valid test or am I talking rubbish ?
On my 2003 model Powabyke , the pedal sensor is hidden inside the bottom bracket , and the wires enter at the bottom . Am I right in thinking that there are magnets mounted on the bottom bracket axle ? Do all pedal sensors work on the hall effect principle , or are there other types ?

 

[shemozzle999]

Normally the control pin is connected to the open collector of an internal switching transistor that is turn on via the hall sensor in the presence of a magnetic field and effectively shorting this pin through the emitter of the transistor to ground or 0v.

 

[neptune]

@Shemozzle 999 .So when a magnetic field is present , there will be zero volts between the control pin [signal wire ] and the negative supply lead . Am i right in thinking that in the absence of that magnetic field that the voltage will then be 5 volts ?

 

[Deleted member 4366]

@d8veh . From the information in your last post you obviously have a lot of knowledge about these things . I worked on washing machines for many years and so I know a bit about fault finding and multimeters , but I am new to electric bikes . So , faced with the above problem you have three wires . The problem is that you have to work out which wire does what ,unless you have a circuit diagram , which is unlikely . So , would it work if you did the following . Unplug the three wires . Test the three wires coming from the controller with a voltmeter [and the bike switched on ] and find which 2 wires have a voltage of 5 volts between them . The remaining wire is then the signal wire . Find a way to disconnect the signal wire only , leaving the supply wires connected to the sensor . Connect an analogue volt meter between the negative supply wire and the signal output of the sensor. When the pedals are turned , analogue meter will swing between zero and 5 volts .Is this a valid test or am I talking rubbish ?
On my 2003 model Powabyke , the pedal sensor is hidden inside the bottom bracket , and the wires enter at the bottom . Am I right in thinking that there are magnets mounted on the bottom bracket axle ? Do all pedal sensors work on the hall effect principle , or are there other types ?

Spot on! except it's normally a bit easier than that. In most cases the wires to the pedal sensor are coloured: Red is +5v; black is ground; the third wire (green, yellow or white mostly) is the signal wire. You can use any 5v supply: a phone charger, a USB port; 3 AA cells.
I don't know about your particular pedal sensor on the Powerbike. There are different ones. The early 8Fun kits (Sunlova) were optical. Some use torque. You'd have to look inside to see what you have.

 

[shemozzle999]

It will not be exactly zero as measured at the control pin even though the transistor is turned on (effectively a short circuit) there is a slight resistance across the junction of the collector and emitter so a small voltage is developed across it. When there is no magnetic field the transistor is turned off (effectively open circuit) and yes you will see whatever voltage is coming from the controller on the control pin.
As d8veh said the other two pins are simply the supply for the whole hall sensor device.
Put simply it is just a magnetic operated switch.

The same principle applies to hall effect controlled motors, they have 3 separate devices connected to the controller and they detect the position of the motor and fire the main supply wires to the drive coils in the motor in the correct sequence to produce rotation. I think if you reverse the sequence of the hall control wires and the motor wires the motor will run in the reverse direction ie: backwards.

 

[karl101]

I suspect this may be unhelpful, but on my Wisper 905SEL the sensor has a built in red LED that illuminates as the magnet goes past. I spotted it the other day.

Karl

 

[z0mb13e]

I suspect this may be unhelpful, but on my Wisper 905SEL the sensor has a built in red LED that illuminates as the magnet goes past. I spotted it the other day.

Karl

Thats interesting. I don't think there is one on my sensor though - would be handy for diagnostic purposes.

Edit: It also looks like the mounting is different.

 

[z0mb13e]

It will not be exactly zero as measured at the control pin even though the transistor is turned on (effectively a short circuit) there is a slight resistance across the junction of the collector and emitter so a small voltage is developed across it. When there is no magnetic field the transistor is turned off (effectively open circuit) and yes you will see whatever voltage is coming from the controller on the control pin.
As d8veh said the other two pins are simply the supply for the whole hall sensor device.
Put simply it is just a magnetic operated switch.

The same principle applies to hall effect controlled motors, they have 3 separate devices connected to the controller and they detect the position of the motor and fire the main supply wires to the drive coils in the motor in the correct sequence to produce rotation. I think if you reverse the sequence of the hall control wires and the motor wires the motor will run in the reverse direction ie: backwards.

I might be over analyzing this, but still want to know the type of hall effect transistor/sensor (NPN/PNP, voltage range and switching current). This is relevant to setting up the sensor on a bench rather than on the bike. Then again, if I know the type fitted and it turns out to be not doing anything, I can replace it with a 99p part rather than a £15 part. I suppose I could try testing it with a multimeter or dragging the scope down to the garage...

 

[shemozzle999]

I might be over analyzing this, but still want to know the type of hall effect transistor/sensor (NPN/PNP, voltage range and switching current). This is relevant to setting up the sensor on a bench rather than on the bike. Then again, if I know the type fitted and it turns out to be not doing anything, I can replace it with a 99p part rather than a £15 part. I suppose I could try testing it with a multimeter or dragging the scope down to the garage...

The type of hall switch sensors are totally dependant on the circuitry they have to drive in the controller.
From my experience the standard controllers I have played with operate as I described i.e.: they sink the current to ground.

You can get ones that operate in reverse way i.e.: they source the current, where the internal transistor is connected to the positive supply and is grounded by the controller - but I have not come across one.

Typically the ss41 devices found in motors have an operating voltage from 3.8v to 24v and are rated at 20mA with a maximum of 50mA current carrying capacity.

If your Wisper has a standard controller I would expect it would operate that way

Another problem is that they are made to switch at different levels field strengths (60 gauss, 140 gauss etc), so unless you can identify the device you could be dabbling in the dark.


Link to ss40 data sheet:

http://docs-europe.electrocomponents.com/webdocs/002e/0900766b8002e123.pdf

 

[Deleted member 4366]

I might be over analyzing this,

You are over-analysing it. Just put a meter on it and rotate the magnets. you'll soon know if it works. If you're lucky, you can get a probe on the signal wire while it's connected, so you might not need an external power supply. easier to do with two people: One to hold the probes and the other to turn the crank.

I connected one of mine to a pic chip and had it working an LED. It just worked like an on/off switch: The first magnet switched it on and the next switched it off, or there were two halls in the sensor and it switched on then off as a single magnet passed. i can't remember which.

You've just reminded me. I need to get back on that project to make a programmable pedal sensor power control.

 

[shemozzle999]

That is another variant you can get d8veh. It is probably a latching version which holds its last state until until acted upon again.

 

[z0mb13e]

The type of hall switch sensors are totally dependant on the circuitry they have to drive in the controller.
From my experience the standard controllers I have played with operate as I described i.e.: they sink the current to ground.

That would be an NPN type. Certainly more common when looking at the available transistors.

You can get ones that operate in reverse way i.e.: they source the current, where the internal transistor is connected to the positive supply and is grounded by the controller - but I have not come across one.

And that would be PNP. Again they certainly seem to be more rare.

Typically the ss41 devices found in motors have an operating voltage from 3.8v to 24v and are rated at 20mA with a maximum of 50mA current carrying capacity.

I would expect, from a manufacturing perspective that all the hall sensors would be the same (though I can think of many reasons why they wouldn't be).

If your Wisper has a standard controller I would expect it would operate that way

Another problem is that they are made to switch at different levels field strengths (60 gauss, 140 gauss etc), so unless you can identify the device you could be dabbling in the dark.


Link to ss40 data sheet:

http://docs-europe.electrocomponents.com/webdocs/002e/0900766b8002e123.pdf

Surely the field strength can be varied by moving the magnet further away from or nearer to the transistor?

Thanks for the link to the datasheet. I will work on the assumption that that applies.

 

[z0mb13e]

You are over-analysing it. Just put a meter on it and rotate the magnets. you'll soon know if it works. If you're lucky, you can get a probe on the signal wire while it's connected, so you might not need an external power supply. easier to do with two people: One to hold the probes and the other to turn the crank.

I connected one of mine to a pic chip and had it working an LED. It just worked like an on/off switch: The first magnet switched it on and the next switched it off, or there were two halls in the sensor and it switched on then off as a single magnet passed. i can't remember which.

You've just reminded me. I need to get back on that project to make a programmable pedal sensor power control.

Yes, that sounds like the integrated or encapsulated type of hall sensor rather than just a bare transistor. Farnell do something like this but it will set you back £30...

And I know that I do over analyse things, but then I like to know the details before I hook stuff up and accidentally destroy it!

 

[flecc]

Surely the field strength can be varied by moving the magnet further away from or nearer to the transistor?

Greater distance would lead to less precision in the switching point though, it could become very vague.

 

[z0mb13e]

Greater distance would lead to less precision in the switching point though, it could become very vague.

Precision isn't a huge issue for testing purposes.

If testing the sensor on a bench then all I would need would be enough to see some deflection on a scope. Might be an issue testing with a multimeter though.

 

[shemozzle999]

New Honeywell sensor product list/specs link:

Honeywell Sensing and Control Product Search

Just type in markings shown on the device prefixed with "ss" in the search box i.e. 41f = ss41f