Manual Cam Checklist Configuration

Syvecs Staff - Cleaner
Posts: 356
Joined: Fri May 23, 2008 9:23 am
Location: Out there... somewhere

Manual Cam Checklist Configuration

Post by pat »

I have on a number of occasions been asked to explain the operation of the Manual Cam Checklist, so rather than keep repeating it, below is a description of how it works:

For crank patterns which have a unique feature and can go into 360 sync without further information from the Phase Sensor (eg 36-1, Subaru G11 etc), the Manual Cam Checklist allows phase detection with fairly arbitrary cam sensor patterns. In order to do so, it is necessary to find something "unique" about the cam signal, something that can distinguish one revolution from another.

First thing to do is to find all the teeth on the cam and prevent the engine from stalling if incorrect detection happens whilst still gathering information. Enable the Manual Cam Checklist feature, but under Starting set the minimum engine speed for 720 sync to a high value, say 7000 RPM. This will stop the ECU going into 720 until we are ready for that.

Next thing is to make sure that the Phase Sensor signal is clean and reliable. For VR sensors choose the fastest edge (normally falling) and trigger thresholds that match the signal (ie notional 0V for low trigger threshold with falling edge, high trigger threshold above any noise present). For Hall Effect sensors, normally the falling edge tends to be the fastest edge, but it may not be ideal if the signal inverts as it sometimes does with some sensors.

Now with the engine running in 360 sync, bring up a gauge for camRaw, as a text display updating at 1 or 2 Hz. It helps to have some patience and know how many teeth to expect; the values will cycle fairly randomly but there will be "clusters" eg 19.5, 19.75, 19.25, 19.25 etc then it may quickly show another value, say 555.50 and then move on. Try to write down the "centre" of each cluster, corresponding to each tooth. Don't worry about the fact that the numbers may fall in anywhere between 0 and 720, the ECU just had to "guess" at a phase when it fired the engine, but because it is in 360 sync it doesn't matter if it guessed the wrong phase. For now all that is needed are the tooth angle numbers.

With the list of angles, now subtract 360 from any numbers that are bigger than that, to "normalise" them. We'll look at 2 examples here - a) a pattern with 3 evenly spaced teeth and b) a 4-1 pattern :

For a), let's say there were readings at 180, 420 and 660 degrees. That translates into one tooth at 180 degrees in one revolution and two teeth at 60 and 300 degrees in the second revolution. Every time a tooth happens, a counter is increased. At each Check Angle configured in the Manual Cam Checklist, this number is read and reset. So with this example, if there was a Check Angle at 90 degrees, then in the second revolution there would be one tooth, but in the first revolution there would be none. If another Check Angle is added at 210 degrees then in the first revolution one tooth would have been seen but in the second revolution no new teeth have happened since the last Check Angle (90 degrees) so it would count zero. Finally, if another Check Angle is added at 330 degrees then in the first revolution no new teeth would have been seen so the count would be zero, but in the second revolution a new tooth would have been seen making the count 1. This information is used in the Check Counts. So for revolution 1 we would have counts of 0, 1 and 0, whereas for revolution 2 we would have counts of 1, 0 and 1. Of course, we COULD have just one Check Angle at 330 degrees - in the first revolution there would only be 1 tooth and in the second there would be 2, so why bother with the extra work of adding more Check Angles ? Safety. If we only check once per revolution then anything that happens that is wrong will only be noticed at that point, it isn't being checked before. Three Check Points are fairly sensible here, but it would be possible to put more in, ahead of each tooth, to make sure that no phantom teeth have been seen.

For b) let's say there are teeth at 120, 300 and 480 degrees. That translates into two teeth at 120 and 300 degrees in one rotation and one tooth at 120 degrees in the next rotation. We could define Check Angles at 150 and 330 degrees. In the first and second rotation, at 150 degrees a tooth should have been seen, but only in the first rotation would the tooth at 300 degrees have been seen by the 330 degree Check Angle. Here the Check Counts would be 1 and 1 in the first rotation and 1 and 0 in the second rotation.

Once the cam angles have been measured, and the number of Checklist entries configured along with the Check Angles and Check Counts, the minimum RPM for 720 sync can be dropped and an attempt made to go into 720 sync. Remember that the phase was guessed so it may be wrong! If upon entering 720 sync the engine stalls then this is exactly what has happened (assuming coil on plug) and the Check Counts need to be swapped between the first and second rotation to get the correct phasing.

What if the cam is also a VVT cam ? In that case it will be necessary to take into account what the tooth may do - if the cam is already fully retarded then the check angles recorded above may well be fine (so long as they are not too long after the tooth), but if the cam could retard then this must be factored into the Check Angles. Taking the 3 tooth example above, if that cam could retard 50 degrees then those angles would become 230, 110 and 350 degrees. Clearly Check Angles of 90, 210 and 330 won't work because the cam retard will cause teeth to move from before to after those angles. The check angles could however be retarded 30 degrees to 120, 240 and 360 degrees. 360 degrees is of course the same as 0 degrees, so that goes to the start of the list. Note that this will fall in the next rotation. So the Check Point at 0 degrees in rotation 1 would show 1 tooth (because it was originally at 330 but it has been delayed into the next rotation), whereas the Check Point at 0 degrees in the second rotation will show 0.

Special care must be taken when a Hall Effect sensor is used for Phase detection and its signal is prone to inversion. The choice of the correct edge to use may be limited such that the edge moves from the trailing edge of a tooth to the leading edge, rather than the leading edge of one tooth to the trailing edge of the previous tooth. Under these circumstances it may be a necessary evil to use the slower rising edge than the fast falling edge. If there is another phase-capable cam signal available that could work as falling edge it may be better to move the Phase Sensor selection across to that sensor.

Hopefully this sheds some light on how the Manual Cam Checklist feature allows arbitrary cam patterns to allow phase detection. NB: There is a special case for cams that have LOTS of teeth. Here, a count of 7 in any Check Count means 7 *OR MORE*. This may be useful to limit the number of check angles where there is a clear distinction between phases (eg just 1 tooth in one rotation, but 10 in the second).