This is how a car should idle with the A/C on.
 

I'm still dialing in AFR and enrichments.

But god damn if I've got A/C idle-up working better in two days than the duration of ownership of my other ecus.

Very impressive, nice job.

Going from the Adaptronic to the ms2 I too am AMAZED at how much better the ms handles ac engagement.

How much better than the adaptronic or how much better than MS1?

I still struggle with A/C at times. Nice job.

I confess I have been unwittingly spreading lies about allowable PWM duty limits of the stock idle valve. For whatever reason, my hydra used 15% and 85% as the limits. So I reused that in the adaptronic. Travis told me I was nuts and that it didn't do much below 65%. I had clearly remembered testing the max and min values and watching the RPM change as I went down as far as 15%. It could just be that it wasn't true duty cycle on the Hydra...

So Brainy came over and schooled me on the minimum duty cycle test for the MS3 using Warmup Only PWM idle. Once we did that, shit got reeeeeeeeal.

still...

There are some nice idle features that the MS3X has like an independent VE table for just around idle. It also links the functionality of the condenser fan and A/C so that 1) they both come on at the same time and 2) you can set the amount of duty cycle and delay before they engage as well as 3) set limits on their operation like TPS and mph restrictions.

I have barely scratched the surface of the feature set but the car runs 90% as good as anything I've tuned before.

Here's what I've done and % completion so far:

Ported over my maps from previous engine management systems
Tuned idle VE and A/C (85% need to fix my VSS and that lean spike when A/C kicks on)
Tuned cranking, after start, and warmup (80% -- it starts in 30F weather ok)
Accel Enrichments (60% -- I literally just threw some big numbers in and it is quite smooth)
Autotuned the fuel map (80% -- street tune for cruise mostly but it's a good system!)
Fans work properly
Open loop VVT (mimicks my ideal map but only uses full advance and full retard)
Running full sequential fuel and full sequential coil on plug
VTCS works for cold start (<60C or <1800rpm)
NB Alternator controlled by the JasonC SBB circuit hold steady at 14.3V.


Stuff that still needs work:
VVT obviously. The MS3 roadmap suggests this is in the works for the next release (1.0 RC4 just came out--VVT is planned for 1.1. I just have to wait.)
Launch control needs setting up properly
Tweaking of AFR targets
Enhanced Acceleration Enrichments need to be set up and tuned
Boost control and subsequently Closed Loop BC
Water injection (why not? the MS supports it--I will have no-water tableswitching)
Knock Sensing. This will be a while but is expected for MS3 firmware 2.0 which will supposedly have an angle clock that allows windowing.


I think that's about it.

Plans have changed here a little, we're combining what we'd planned for 1.1 with 2.0...

So the next major release will be 2.0. First thing I'm going to work on there is making comms more robust, then angle clock, then phasing in scheduling of everything using angle clock, then I'll probably start on VVT.

It means that it will be a while before I work on VVT, but I will do it, and when I'm done it will support up to 4 cams at once.

Glad to hear things are going well for you though. I'm assuming that you're using AC idleup with closed loop idle speed control?

Ken

Ken, you just ruined Christmas.

yes, closed loop idle control. once I got the mixture and timing dialed in a bit better, it smoothed out the idle speed at higher load. It now doesn't drop to 800 and shudder.

Haha sorry about that. I decided that I'd rather just support the VVT feature fully the first time so I don't have to go back and rewrite parts of it after doing the angle clock.

Realistically it means your knock feature will get in a lot sooner, VVT might take a little longer than it was going to, but overall a lot of features will get done sooner than they might have, and the MS as a whole should work better since the angle clock gets rid of a lot of the overhead we currently have.

Ken

That's cool. I'm looking forward to all of the above. You guys have a rough timeline yet?

It's hard to say at this point. I don't have any idea how long the comms work will take. My estimate is that it'll take 2-4 weeks, then angle clock will likely take 2 months start to finish. VVT after that is likely 2 weeks or so.

Before I start on all that James and I need to at least finish documenting all of 1.0's features as well.

Ken

Isn't a big chunk of the VVT work getting the loop to work nicely and stabley?
Then making it capable of 4 cams is just replicating it?

Part of the work is also measuring the current cam position. Doing that with our current code requires an expensive bit of math in an interrupt handler to convert time to angle, and ends up being different for every engine. If I do that for 4 cams, that is 4x the math in interrupt code.

With the angle clock, I'll already have an up-to-date angle when the interrupt for the cam comes in, so all I'll have to do is latch the angle so the PID loop can use it. Doing that 4 times isn't really that much more expensive than doing it one time since it's done in the main loop, and not done on every mainloop iteration.

Getting it to work in a stable manner is something I already have a good grasp of as well as several users. It's just a PID loop.

Ken