MS2/extra beta 8 (20070818)

[TT350chevelle]

Well I had to return to MS I because I want to race this weekend and I just don't have enough time to debug my setup, MS II upgrade/testing will have to wait for me.

One odd thing I noticed is the injector PW (as reported by Megatune) was very close to being the same from MS I to MS II but the MS I is set to 4 squirts alt and MS II is set to 2 squirts alt.... that does not seem right.I also had to reduce my MS I VE map by a min of 35 points to get the engine to run and it's still pretty rich, not a big deal but expected my VE map to be close.

Hopefully I will beable to continue MS II testing at a later date to dig into some of the oddities I noticed and see if I can get it to work as well as MS I.

Brad J.

[muythaibxr]

This could easily have been an issue of the number of injectors you had configured or just about anything else that changes req_fuel.

I did the same comparison when trying to figure out some of the issues people were having, and the pulse-width given a certain VE is within .5-1 ms of each other from ms1-ms2.

Ken

[muythaibxr]

What I'd like people to do for this different numbers of squirts running richer/leaner problem is the following:

Go back to beta 6, and change the number of squirts, and see if the behavior is the same.

All my testing says that beta 6 and beta 7 are the same..

The only thing I can come up with is that when I made my opening-time changes, I might have messed up somewhere (or I might have fixed a problem that was there previously).

Ken

[Keithg]

Probably redundant information, but with Beta7, 2 squirts simultaneous and 1 squirt simultaneous yeilds the same AFR for me. All betas up to and including Beta 7 worked just fine in this respect. I have not loaded Beta8 as the only change did not affect me.

KeithG

On 8/28/07, muythaibxr wrote:

What I'd like people to do for this different numbers of squirts running richer/leaner problem is the following:

Go back to beta 6, and change the number of squirts, and see if the behavior is the same.

All my testing says that beta 6 and beta 7 are the same..

The only thing I can come up with is that when I made my opening-time changes, I might have messed up somewhere.

Ken

[muythaibxr]

It might just be that people did not retune their opening times after I made the changes to make that setting accurate to the microsecond then.

[tk_tech]

ok i think i missed some thing, could someone please explain retuning opening times. i have been using default times 1 ms, should i be changing with number of squirts, ie: shorting up time

sean

[muythaibxr]

If changing the number of squirts changes the AFR, then the 1ms opening time is not accurate, and you need to change it to make the AFR close to the same.

With a properly tuned opening time, you should be able to change between any number of squirts and maintain an AFR that's pretty close.

[muythaibxr]

And what is common problem: VE table calculations. It is too rough for nowadays resolution - without decimal fraction (?) Mainly it become as problem at lower side of VE map. Say at idle. In my case 54 or 55 VE change injection time enough for idle come unstable. It cange injection time on my VW engine from1.674ms to 1.717ms (0.043ms. my inj openning time is 1.1ms so it is 14% of injected fuel ). Yes I flatten map: got 0.015ms fluctuations and more stable idle. But this is only patching.

Gints.

The VE table calculations are to .1% units... If you're using multiply MAP, that can magnify pulse-width change caused by MAP fluctuations.

Ken

[muythaibxr]

Can't get to work rev limiter with spark cut. Fuel and spark works in strange manner - first hits upper and then run without spark or fuel until lower limit and so on.

Fuel cut seemed to work the way I'd expect, but spark cut is not working with distributor mode (seems to work fine in wheel mode). I'll have to look into why this weekend, but this is a good find!

Thanks.

Ken

[muythaibxr]

What I'd like people to do for this different numbers of squirts running richer/leaner problem is the following:

Go back to beta 6, and change the number of squirts, and see if the behavior is the same.

All my testing says that beta 6 and beta 7 are the same..

The only thing I can come up with is that when I made my opening-time changes, I might have messed up somewhere (or I might have fixed a problem that was there previously).

Ken

Can those who noticed differences when going between different numbers of squirts (ballistic, a couple others) try this:

change between various numbers of squirts, and instead of adjusting VE to fix any lean/rich problems on idle, only change the opening time.

You should eventually be able to find an opening time that makes it so that you don't need to retune VE.

I did fix what looked to me like a bug in the battery voltage correction that may be causing people to need to reset their opening times. I fixed this long before beta 6 though, so it might be worthwhile to try the other thing I quoted in this message.

Ken

[GintsK]

And what is common problem: VE table calculations. It is too rough for nowadays resolution - without decimal fraction (?) Mainly it become as problem at lower side of VE map. Say at idle. In my case 54 or 55 VE change injection time enough for idle come unstable. It cange injection time on my VW engine from1.674ms to 1.717ms (0.043ms. my inj openning time is 1.1ms so it is 14% of injected fuel ). Yes I flatten map: got 0.015ms fluctuations and more stable idle. But this is only patching.

Gints.

The VE table calculations are to .1% units... If you're using multiply MAP, that can magnify pulse-width change caused by MAP fluctuations.

Ken

Yes I use MAP multiplication. How can I migrate to no multiplication? Is current VE table is usable? What this feature realy do?

[muythaibxr]

Multiply MAP just multiplies the current MAP into the final fuel calc... which is good on boost, as you don't have to have your VE numbers as large, but is bad for resolution.

Turning it off will require you to retune pretty much the whole table, but will give you finer resolution, since MAP change no longer causes pulse-width changes on its own (without VE changing).

Ken

[GintsK]

Can't get to work rev limiter with spark cut. Fuel and spark works in strange manner - first hits upper and then run without spark or fuel until lower limit and so on.

Fuel cut seemed to work the way I'd expect, but spark cut is not working with distributor mode (seems to work fine in wheel mode). I'll have to look into why this weekend, but this is a good find!

Thanks.

Ken

Engine running with no firing from high limit rpm to low limit rpm I found on 60-2 trigger wheel equipped car. It is with Fuel&Spark cut limiter. Much smoother algorithm would be start some rare cutting at lower rpms and then maximize cuttings during rpms increase to upper limit.

Gints

[jsmcortina]

Can't get to work rev limiter with spark cut. Fuel and spark works in strange manner - first hits upper and then run without spark or fuel until lower limit and so on.

This is the way it is written in MS2.
We could get rid of that and use the single setpoint like in MS1 I suppose.

Much smoother algorithm would be start some rare cutting at lower rpms and then maximize cuttings during rpms increase to upper limit.
Gints

We had thought of that - an adaptive cut would be really nice, although a crude interpolated method could be developed I guess.

James

[GintsK]

Multiply MAP just multiplies the current MAP into the final fuel calc... which is good on boost, as you don't have to have your VE numbers as large, but is bad for resolution.

Turning it off will require you to retune pretty much the whole table, but will give you finer resolution, since MAP change no longer causes pulse-width changes on its own (without VE changing).

Ken

Is it same multiplication as MS1-E do?
PW = REQ_FUEL * VE * MAP * E + accel + Injector_open_time
What with boosted car? Can I switch multiplication off too?

[muythaibxr]

Yeah, same as ms1 there...

You can turn it off on a boosted car as long as you don't run out of VE ... meaning as long as the VE doesn't go over 255.

Ken

[rb26dett]

the rev cut should have hysteresis in it. i thought that at the lower rpm it retarded spark (if setup that way) then at the higher rpm it cut it, and didnt stop cutting it until the lower rpm again?

if thats the case, thats pretty much ideal.you can always set the two rpms close together such that the engine just hits a wall.

maybe it would be nicer if the hard hysteresis and the difference between soft cut and hard cut were different numbers seperately configurable?

just thinking out loud.

fred.

[muythaibxr]

Assuming I have a good way of testing, the next beta will include a fix for missing-tooth wheels with less than 12 teeth.

Ken

[muythaibxr]

OK, I've got most of the fix for the < 12 tooth wheels done. I'll be doing the rest of that fix tomorrow, and hopefully releasing a beta with that fix in.

Then I'll be working on fixing all remaining bugs during the evenings this week.

Ken

[UnaClocker]

How many teeth are on a Neon wheel? heh.. Probably more than 12, I imagine.