Worn starter brush prevents starting

This week we had a 1996 Ford Falcon that would battle to start. A garage had changed spark plugs and leads and coil and PC valve while trying to solve the issues. It would take a while to start, if at all, and then run rough for a few seconds then run well after that. We removed the spark plugs and they all looked clean enough, then checked for spark while cranking. There was no spark on any of the leads while cranking but when the key was let go it gave a few sparks as the engine slowed down and stopped. We checked the ignition switch wiring and there was no fault there. [Sometimes the ignition switch can lose contact with main ignition while cranking but regain it when the key is back in the normal position].

The next step was to scope the CAM and crank sensor patterns as that is what the ECU needs to calculate its spark timing. We had found that we could get the car running by ‘flicking’ the ignition switch a few times, the good spark from the ‘off’ time managed to get the car running after a few tries. This is a scope pattern of the CAM [yellow] and crank [red] sensor patterns. In the first part of the pattern the engine is running - then turned off, then the second half is cranking with no-start.


Look at the hash in the crank sensor signal when cranking and then the clean pattern at the end when the ignition switch is let go, and the starter is no longer engaged.


Now look at the battery positive [red] and engine earth[yellow] being scoped while cranking - lots of ‘hash’!

falcon crank.png

The hash on the crank sensor is directly related to the starter motor - time to pull the starter off and open it up. Look how badly worn the one brush was while the others were the correct length.


A new starter motor was fitted and we scoped the CAM and crank sensor signals again. The first half of the pattern was with the engine running, then shut down halfway through the pattern, and then cranking and starting - it cranked and started good every time after that. The bad brush contact was creating an emp, or many of them as the brush made and lost contact, that was interfering with the CAM and crank sensor signals while cranking so the ECU couldn’t work out the right time to fire the spark. A well programmed computer would be able to work that out and give a fault code - not this one.