How to: Interface CA speedo to Vectrix Encoder

I've done some searching here and elsewhere and it seems that this is something no-one else has done yet - being able to read the speed of the Vectrix by interfacing with the motor encoder.

I think its a pretty essential modification, as it gives you your watthours per klm, a speedo that's less jerky than stock and higher resolution.

The motor encoder is a quadrature differential encoder, complete with differential A/B and differential hall outputs. The hall outputs fire 6 times per revolution, and output 3.6V. The A/B outputs are MUCH higher frequency. Pinouts were determined with a multimeter and oscilloscope and the pinout in the factory manual is completely wrong. Don't use it.

Knowing this, you can intercept the signal using a high impedance buffer (so you don't load down the signal - very important), and isolate it using an optocoupler (you must isolate it or you break isolation provided by the DC:DC converter).

I initially used the 5V rail (white wire, pin 13), ground (black wire, pin 10) and yellow hall (yellow wire, pin 3) as inputs, and ran these to an LM358 set up as a voltage follower, feeding an MCT9001 optocoupler (I used this part for two reasons - I had it in stock, and it has a good CTR - current transfer ratio). I then fed this to the speed wire of the Cycle Analyst, and used the opto to pull it to ground. This worked well till about 60Kph, when it dropped to zero - WHUT?!

Scoping the waveform showed a characteristic triangle wave of an RC (resistor capacitor) constant - Makes sense, since the speed pad is fed by a high value resistor, and has a capacitor to ground (to filter high frequency noise) - as the frequency went up, the waveform got smaller and smaller till it was no longer sensed as valid input by the CA's microprocessor.

Humph. The designer clearly never expected the frequency to get this high (300Hz) being pulled to ground. So - I ran a 2K resistor from the CA's 5V rail to the SP pad - and the waveform immediately improved - a nice square wave that the CA detected to the end of the speedo.

BUT - the additional load imposed by the 2K resistor meant at low speed (under 2Kph) the backlight was flashing dim/bright/dim as the hall sensor fired, and the voltage reg got very hot. This is a bad thing.

So - an auxiliary +5V supply was required - this consisted of a large series resistor (10K) and zener diode + cap, which now feeds the SP pad of the CA via a 1K resistor. Problem solved.

To calibrate the speedo, the values of 1600mm for the tyre circumference (measured with a dressmaker's tape around the tyre) poles of 28 (its a 12 pole motor, with a 4.6:1 reduction ratio - BUT the hall transitions happen 6 times per revolution NOT 12).

This was all tested with the bike on a jack with the rear wheel lifted, and the kickstand contacts shorted with a piece of wire. My CA is an older model V2.3, that had done a few years duty on ebikes already. Its fed via a relay that comes on with the bike ignition, and also when the charger is powered.