Grinfineon Controllers

 

1 Flash: Brake Cutoff Engaged

2 Flashes: Regenerative Braking has been Engaged

3 Flashes: Battery voltage is below the low voltage cutoff

4 Flashes: Controller shut off because motor had stalled

5 Flashes: Throttle voltage fault, or already high when controller powered on

 

If there is an actual controller fault, then the LED will stay on for a long time and only briefly turn off.

 

1 Flash: Motor Short Circuit Protection has Engaged

2 Flashes: 15V Regulator Circuit is Damaged

3 Flashes: Throttle Signal Fault (>4.2V)

5 Flashes: Microprocessor is damaged

 

Notice that there is no flash sequence for an invalid hall sequence (as happens when one signal is down) or when the controller loses lock in sensorless mode because the electrical RPM was too high. So in practice, the LED codes aren't all that useful in some key fault situations.

 

Notice too that there is a bit of a 'staircase' pattern between the input throttle voltage and the controller output. It turns out the throttle signal is being discretized into about 50 steps from 0 to max power. In most situations with a throttle control this is not very noticeable, but with higher power setups that use the CA's throttle limiting features, it can result in a 'rough' feeling as the controller is repeatedly jumping between two different output states rather than settling at the ideal value in between. The CA limiting still works, but not as nicely as we would like, and we are working to resolve this issue.

Since Feb 2010, all of our Infineon controllers have been supplied with regenerative braking already enabled by default. Simply plug in an ebrake lever to the 4-pin connector on the controller, and the controller will brake regeneratively when the lever is depressed and charge the battery in the process. The amount of regen torque is roughly constant, which means that the amount of regen current flowing back into the battery is a function of the speed at which you are travelling. If you are moving faster, the braking torque is the same, but you have more watts going into the battery because the power (torque*rpm) is greater. Similarly, with a higher voltage battery, the same power into the hub (Volts * Amps) equates to less current. The graph below shows the amount of regen current measured against wheel RPM with the 2806 winding hub.

While the regenerative braking can bring the bike to a stop reasonably fast, it has nowhere near the stopping power of fully applied mechanical brakes, which should always be available for emergency stops. To disable regen and have the ebrake lever act merely as a cutoff, then you need to open the controller and desolder the jumper that connects pad P22 to Ground.

With our controllers shipped before Feb 2010, regen was not enabled by default. To enable regen, it was necessary to open up the controller and then short the pad labelled BK to Ground with a jumper. On controllers shipped before Feb 2010, there were two levels of braking force: maximum regen was achieved by squeezing the ebrakes while engaging the throttle, while a lesser regen current was possible by using the ebrakes only and no throttle. The graph below shows our measurements of the regenerative braking current as a function of the wheel speed, both with and without the throttle applied.

In our current controller offering, both the 25A and 40A controller models will work with 60 and 72V batteries as-is.

 

The Infineon controllers sold prior to July 2011 have a somewhat limited design for their 12V linear regulator, which makes it difficult to have the same controller work with both 36V packs and 72V batteries. If you want to run with more than a 48V battery pack, then it was essential to replace the 2 watt power resistor R1 on the PCB with a higher series resistance. For 60V operation required 400 ohms, while 72V operation needed about 600 ohms. There is approximately 50mA of current through this resistor, so we recommend putting two 2W resistors in series to handle the heat.

 

One other issue is that the controllers were designed to have a fixed 57V regen cutoff voltage. That meant that regen would not work with a higher voltage battery pack unless you tricked the controller into thinking everything is running at a lower voltage. This was achieved by changing the small 0805 surface mount resistor R12, from 2.4kOhm down to about 1.6kOhm. This mod would increase the regen cutoff up to 84V, and it will also raise the low voltage cutoff from 27V to 40V.

We've made the pinouts of the hall and power leads consistent between any of the motors purchased from us and these controllers, be it eZee, Nine Continent, or Crystalyte. They will just plug together and work, even though the actual colour sequence may not match up. If you have a 3rd party motor, then it will almost certainly have a different connector and pinout standard, and it is up to you to figure out and hook up the correct wiring between the hall signals and the phases. There are in general 36 possible ways to connect a sensored controller to a hub motor (6 pairings for the hall signal leads and 6 pairings for the power leads). Of those, only 3 will work correctly, the rest will either shake the motor or cause it to rotate slowly with a heavy growl.

If you are attaching to a 3rd party motor, you'll find that there are 3 colours for the phase leads and the same 3 colours for the hall signals. Suppose the motor uses orange, purple, and white wires, while our controllers have the more common colours of yellow, green, and blue. As a first guess, you could try:

 

Yellow <-> Purple

Green <-> Orange

Blue <-> White

Do this exact mapping both for the power leads AND the phase leads. If the motor spins correctly, then great. If not, you then swap any pair in the mapping. So, for instance:

 

Yellow <-> Purple

Green <-> White

Blue <-> Orange

 

Where the orange and white were swapped, and then it should spin fine. You could also have swapped the purple and orange, or purple and white, instead. That means that if your motor also uses the yellow, green, and blue colour coding, you should as a first attempt try just matching the colours. If that doesn't work, then swap any arbitrary pair of them, just be sure to swap the same pair both with the hall signals and the power leads. If it happens that your motor uses a different set of colours for the phase leads and the hall signal leads, then you'll have to resort to a lot more trial and error.

 

As it is now, the infineon controllers have a direct colour matching with the Nine Continent and BMC hubs. However, with both the Crystalyte and eZee motors, a swap is required. With our connector standards it is the blue and the green that are swapped.

 

Note that the 3 working sequences are all just cyclic permutations of each other. If you take a working sequence, and then rotate all of the phase and hall signals over one, then it will still work:

 

If this works	So would this	and this
Yellow <-> Yellow	Yellow <-> Blue	Yellow <-> Green
Green <-> Green	Green <-> Yellow	Green <-> Blue
Blue <-> Blue	Blue <-> Green	Blue <-> Yellow