DIYEFI.org Forum

Users of MS2Extra 3.X lower than beta 3.2.1 using sequential mode are advised to proceed with caution. This code contains severe bugs that can and has damaged engines. 3.2.X contains a rewrite of the affected parts, however may not be extensively tested yet. You may not be affected depending upon your exact configuration. The issue arrises with specific combinations of input patterns and output timing. Suggested fixes by the developers at the time of buying larger injectors are not considered appropriate by myself and others. If you're using this code be very careful if/when you update your injection timing parameters.

More information is available in the MS2Extra development forum:

http://msextra.com/forums/viewforum.php?f=91

Specifically the following threads:

• 3.1.1 had some good stuff in it, but I can't find it? Swalled by post volume? Censored? No idea.
• 3.1.2 beta release thread http://msextra.com/forums/viewtopic.php?f=91&t=41714
• 3.1.3 beta release thread http://msextra.com/forums/viewtopic.php?f=91&t=42417
• 3.1.4 beta release thread http://msextra.com/forums/viewtopic.php?f=91&t=42672
• 3.2.0 release thread http://msextra.com/forums/viewtopic.php?f=91&t=42893
• 3.2.1 beta release thread http://msextra.com/forums/viewtopic.php?f=91&t=43262
• 3.2.1 release feeler thread http://msextra.com/forums/viewtopic.php?f=91&t=43315

Some great quotes:

From: http://www.msextra.com/forums/viewtopic ... 20#p300220

So bigger injectors and correct timing are what you need to be looking for. With those, there won't be any issue.

FreeEMS handles this just fine:

So sometimes it can't schedule the pulse because the injector is still on (uses the same timer) so it starts at the next tooth which is just slightly later.

Avoid 3.1.4 too:

I just flashed 3.1.2 and had a run down the road. 3.1.2 runs the same as 3.1.1 for me. i.e. 99.9% ok with rare lean spikes/misfire. 3.1.4 runs rough all the time.

Eating their words, note, they DID the rewrite in 3.2.X:

When I said it's not a bug, I should have been clearer. It is not the ideal situation but it is due to limitations in the implementation. It could be corrected but that would involve completely rewriting the sequential injection code. I am not ready to do that but if someone else is up for it then be my guest.

OK, enough time spent protecting MS users from their firmware providers.

Fred.

Just to explain what is likely happening there, and I'm not 100% sure as I've not looked at the code, if the injection timing swings around enough it could split a pulse width across two cylinder intake cycles and make one overly rich (combined with a normally timed pulse) and the other lean (only a fraction of the pulse provided). There are a thousand other bad bad things that could be happening there, and this is purely speculation, however, it's probably a good guess.

My understanding is you need to limit the rate of change (per cycle) of the injector timming.

Depending on your inj timing map, large changes in MAP and RPM can cause your injector pulses to overlap. For very large duty cycles, you may need to limit the rate of change to a few degrees per cycle.

Con

That's true too, but these people didn't have wild injection timing maps, their timing was reasonable, and the scheduling code went "Oh, we can't do it right now with our current algorithm, we'll just leave it till later, it'll probably be fine!", but no, no it wasn't fine, it killed engine(s).

Re timing, read this:

viewtopic.php?f=54&t=1114

Thanks