MS2/extra beta 9 (20070905)

Post by **muythaibxr** » Wed Sep 05, 2007 8:44 pm

Time for the next beta. Changes include:

1) Fixes to allow wheels with < 12 teeth to get proper dwell. Dwell may not be as stable on these wheels, but should no longer be lost completely in jump-back situations
2) Fix to 1ditable that should allow it to handle negative numbers. This should fix the reported problem with table lookup doing odd things when negative with closed loop fast idle. (I have not had time to test this fix. Whoever reported this should test it, and I'll test it on the bench when I have time again).
3) Fix so on-off fast idle hysteresis works.
4) Fix so that JS11 should work as an input for launch control (untested, but configured as per hcs12 docs, testers should let me know if this works).

The beta can be found here:

http://www.msextra.com/ms2extra/files/b ... 070905.zip

Ken

GintsK · Post by **GintsK** » Thu Sep 06, 2007 2:27 pm

I've tested launch on JS11 today. Now it triggers. But in spark cut mode works only retard. Not spark cut. It is on time based dizzy setup. D14 is spark output.

Should trigger return must work?

Post by **muythaibxr** » Thu Sep 06, 2007 2:39 pm

Yeah, I knew spark cut revlimiting didn't work in dizzy mode, I think there's a bug in the bug list for it already.

I already know what the problem is and how to fix it too, and will try to have it fixed in the next beta.

Ken

easyBob2 · Post by **easyBob2** » Thu Sep 06, 2007 6:54 pm

The <12 tooth (w/missing tooth) still has a problem...It's not as bad, but it's still there.

I guess I should ask, should I start looking for a different toothed wheel or can this dwell issue be fixed in the code? I'll just start looking for more wheel options, just in case.

Oh, and the PWM idle code still only has the 2 low options; us mazda/ford folks need a higher Hz.

-Chris

Post by **muythaibxr** » Thu Sep 06, 2007 8:01 pm

I have a mazda, it doesn't have any problem with the lower hz option.

That said, I'll add it when I get to it, I never said it was there in the first place.

As far as the dwell problem, I tested from crank to 10000 rpms, and didn't see any missing, so your problem might be caused by something else.

At exactly what RPM, advance, and conditions (quick accel? slight accel? steady state?) are you seeing the problem?

As far as I can tell on the scope, the dwell problem *is* fixed, but I'll check under your exact conditions to be sure.

Please post your latest msq as well.

Ken

Post by **muythaibxr** » Mon Sep 10, 2007 7:18 am

muythaibxr wrote: At exactly what RPM, advance, and conditions (quick accel? slight accel? steady state?) are you seeing the problem?

As far as I can tell on the scope, the dwell problem *is* fixed, but I'll check under your exact conditions to be sure.

Please post your latest msq as well.

So, did you want me to try to look at the problem you're having? If so, could you post the info I asked for? I'd like to make sure that all remaining problems are fixed, and do a "release candidate" for everyone to test.

Ken

easyBob2 · Post by **easyBob2** » Mon Sep 10, 2007 9:01 pm

muythaibxr wrote:
So, did you want me to try to look at the problem you're having? If so, could you post the info I asked for? I'd like to make sure that all remaining problems are fixed, and do a "release candidate" for everyone to test.

Ken

Yeah, sorry. It happens at the same RPM ~2000 under all conditions.

This should be my most current MSQ file.

Thanks,
-Chris

Post by **muythaibxr** » Tue Sep 11, 2007 6:45 am

Alright, I'll take a peek later.

Just FYI though, I did test at that rpm. What advance was it at when it happened, or did you happen to notice?

Ken

teg · Post by **teg** » Tue Sep 11, 2007 7:45 am

muythaibxr wrote: 2) Fix to 1ditable that should allow it to handle negative numbers. This should fix the reported problem with table lookup doing odd things when negative with closed loop fast idle. (I have not had time to test this fix. Whoever reported this should test it, and I'll test it on the bench when I have time again).

Tested - works fine!

Post by **muythaibxr** » Tue Sep 11, 2007 8:08 am

Thanks for trying that out.

Ken

easyBob2 · Post by **easyBob2** » Tue Sep 11, 2007 9:24 am

muythaibxr wrote:Alright, I'll take a peek later.

Just FYI though, I did test at that rpm. What advance was it at when it happened, or did you happen to notice?

Ken

I'll get a datalog today to get some solid numbers on the advance, but as a guess, and because it happens at any load, I'd say that it can happen at any (or most) advances in the spark map.

-Chris

EDIT: Added datalog. Tried to make the missing visable by going through the rpm slowly...You can kind of see it that way (little dips in rpm).

Post by **muythaibxr** » Tue Sep 11, 2007 10:12 am

OK, if it's happening at any advance, that's a bit unusual. I'll take a look, hopefully tonight.

The fix I did basically ensures that you'll get dwell even if the main scheduled dwell misses from a jump-back. It also looked like it was working on the scope before I posted, so I'm thinking there's something pretty specific to your configuration causing the problem.

Ken

Post by **muythaibxr** » Tue Sep 11, 2007 6:58 pm

I'll get a datalog today to get some solid numbers on the advance, but as a guess, and because it happens at any load, I'd say that it can happen at any (or most) advances in the spark map.

I found and fixed the problem. It wasn't happening with my test msq, but was happening with yours, so I guess the conditions had to be "just right" to make it happen.

I did notice that dwell goes long on spark A at the same spot you were missing, so I need to find a fix for that, but it shouldn't miss anymore.

I'll do a release as soon as I've had a chance to fix a few other bugs.

Ken

easyBob2 · Post by **easyBob2** » Thu Sep 13, 2007 9:38 am

muythaibxr wrote:
I'll get a datalog today to get some solid numbers on the advance, but as a guess, and because it happens at any load, I'd say that it can happen at any (or most) advances in the spark map.
I found and fixed the problem. It wasn't happening with my test msq, but was happening with yours, so I guess the conditions had to be "just right" to make it happen.

I did notice that dwell goes long on spark A at the same spot you were missing, so I need to find a fix for that, but it shouldn't miss anymore.

I'll do a release as soon as I've had a chance to fix a few other bugs.

Ken

Thank you! Thank you! Thank you! Thank you! Thank you! Thank you! x50!

-Chris

Keithg · Post by **Keithg** » Thu Sep 13, 2007 11:51 am

Ken,

Is there any thought to getting a COP install to run as W/S during cranking then switch over to COP as it drops out of crank? Also, is there any possibility of adding a 'ramp closed' instead of step closed in closed loop idle? If you lightly roll on the throttle (especially with A/C) it is easy to stall when the AIC snaps closed.

KeithG

Post by **muythaibxr** » Thu Sep 13, 2007 11:59 am

Keithg wrote:Ken,

Is there any thought to getting a COP install to run as W/S during cranking then switch over to COP as it drops out of crank? Also, is there any possibility of adding a 'ramp closed' instead of step closed in closed loop idle? If you lightly roll on the throttle (especially with A/C) it is easy to stall when the AIC snaps closed.

KeithG

Both of these are on the plate for 2.1... there is a workaround for the idle issue that you can use for 2.0 (set the rpm adder a bit larger).

Ken

rb26dett · Post by **rb26dett** » Thu Sep 13, 2007 12:29 pm

muythaibxr wrote:Both of these are on the plate for 2.1

sweet, the long time cranking thing could get annoying in the long term i guess. you are making me homesick with all this great code ;-)

Ballistic · Post by **Ballistic** » Fri Sep 14, 2007 7:51 am

muythaibxr wrote: I found and fixed the problem. It wasn't happening with my test msq, but was happening with yours, so I guess the conditions had to be "just right" to make it happen.

Any chance this is the same problem causing the oscillation I experienced?

Post by **muythaibxr** » Fri Sep 14, 2007 8:04 am

I don't think so... the fix for dizzy with your settings worked perfectly as far as I could tell. The breakage on the 6-1 had to do with the fact that he was running wasted spark and dwell for one of the teeth was jumping over the missing tooth back to where dwell was starting for another coil... we can't start 2 dwells from the same tooth, and that's what caused his problem. The fix for his problem was roughly the same as the fix for yours though. I just extended the rtc dwell section to cover more coils, and made the function that dwells the coils able to dwell more than one at a time.

I think the problem you're having with oscillation might be more related to fuel...

Did you ever get a chance to try adjusting the opening time and battery factor?

Ken

Ballistic · Post by **Ballistic** » Mon Sep 17, 2007 4:30 am

muythaibxr wrote:Did you ever get a chance to try adjusting the opening time and battery factor?

I'm sorry Ken, but I have not. I am totally swamped at the moment. It will be another month before I get some relief.

Megasquirt Support Forum (MSEXTRA)

MS2/extra beta 9 (20070905)

MS2/extra beta 9 (20070905)

Re: MS2/extra beta 9 (20070905)