Monday 10 June 2013

Almost there

Before progressing any further with the injection side of things I wanted to make sure that the ignition control was working as expected throughout the range. That meant actually going out and riding the thing with the laptop plugged into the ECU datalogging while riding.

The first time I took the bike out I noticed straight away that it didn't seem to be revving as high or as freely as before. As I had no dash connected, I wasn't able to tell what the issue was until I pulled the datalog. The bike seemed to refuse to rev above 12,500rpm. After riding and logging some more times I was getting a clearer idea of what was happening. The power appeared to flatten out more as the revs grew higher and eventually there was just not enough power to pull anymore after 12,500rpm. At first I thought it was spark timing retard as the revs grew (hardware latency) but a quick test with the timing light debunked that theory. The small amount of retard I was seeing as the revs grew could easily be put down to lag in the timing light itself.

I left the problem simmer with me for about a week and while doing some research on dwell duration on the CBR coils and typical spark durations, I figured I had the root of the problem. The MS documentation suggested a typical max dwell duration of 3ms for most coils, so unable to measure the dwell on my coils I went with that number. Additionally, as I had not investigated the spark duration, I left it at its default value of 0.2ms. I discovered that the actual max dwell on the CBR coils was 6ms instead of 3ms. And I also found that typical "normal" spark durations run between 1ms and 2ms. MS works by trying to fit the max dwell duration int a revolution and if the period of a revolution reduces to below the sum of the max spark and dwell durations, it will reduce each duration proportionally to fit in the period. I plotted what was happening with the original numbers and the spark duration was quickly being squeezed down to numbers which would barely allow a spark to occur.

So I reset the dwell and spark durations and went for another ride. The difference was apparent straight away. The engine pulled as hard as it should to higher revs. Although I did stumble into another issue then. The RPM signal dropped at about 14,800rpm which cut spark. The signal was gone for 4 seconds before it reappeared and ignition kicked back in rather suddenly. This is likely to be just a noise filter setting in the microsquirts software but I have not had a chance to play with it since.

Datalog showing 4s period of RPM dropout and also a momentary signal dropout that happened just before the main one. That small downward spike just felt like a misfire at the time


I have not installed the CBR600RR COP just yet as I need to make sure ignition is working properly before ditching the HT leads that the timing light works off of. 

2 comments:

Unknown said...

Awesome work my friend. Looking forward to this project as i'll be converting my MC22 this winter as well. The guys on the MS2 extra forum were not very confident about spark timing accuracy at the RPM's this engine runs at tho...

Thanks for sharing

motthomas said...

Hi! Since this post, I've got the engine running without a hitch on ignition control. While, I am not able to accurately measure the spark timing at the top end of the rev range, the engine runs smooth throughout the range with no unexpected power loss or knocking so I can only assume that the ECU is doing its job perfectly!