MS2/Extra 2.0 bug list (OLD, for REFERENCE ONLY)

[muythaibxr]

Try setting it to PA1 but leave the wiring the way it is.

[UnaClocker]

PA1 is not an available option for a generic port.

[muythaibxr]

There is a port that PA0 is reversed with, I thought it was PA1, but it might be another. In any case, select the "other port" and PA0 should work.

In any case, we've not changed anything having to do with it.

Ken

[Keithg]

With the tach output, it is IAC0 and IAC1 which are reversed. I have it connected to IAC0 and need to select IAC1 in MT.

KeithG

[UnaClocker]

Well so far, I've had no luck getting PA0 to work (reliably) with any of the Beta's..

[jsmcortina]

This may be a code issue. PA0 is sparkD as well or possibly an input. So if the init code is in a tangle maybe it is being set as an input instead of output pin. Double check all menus for other uses of PA0.

James

[muythaibxr]

Ahh, I got mixed up.

Even though there aren't many extra ports, I still lose track!

Ken

[Keithg]

James,

I think I may have found a bug in the dwell after trigger code.

I know you put this in for me and I really appreciate it. There seems to be a problem with it, though. I cannot run 1ms. This is the setting I ran with MS1/E (If you look through the MS1/E code, you can see the bits I commented out that allow me to run this - Search on Saab). When I try this setting in MS2/E, it bucks violently at 3500+ rpm and WOT. It is fine at part throttle and idle, but at WOT it bucks it out of gear. I end up running 0.4. This seems to be a decent compromise, but it still bucks and hard at WOT but not as bad as 1.0 I thought it was too rich and blowing out the spark, but that does not seem to be it. In the log, you can see where my external boost controller is pulling boost because it senses knock, so it is actually leaner than it needs to be. The rpm trace shows clearly when it bucks. 0.4 is probably why mine idles harshly as well.

Can you put my MSQ in and run this on the JimStim and see if there is something affecting the 'time after' code? I am running the Beta from 7/3/07, though it has been present on the alphas from 6/10 as well.

KeithG

[msoultan]

Yeah, I'll look at the ini again just to be sure there's no problem there. I thought it might be a problem with the "canned" wizard, but I just took a closer look at the ini file, and it's got a max of "500" which can't be write seeing as how it's a U08.

Ken

Hey Ken,
I was looking at the megamanual documentation for the Acceleration Wizard and it says the following:

"If the car does not buck under acceleration, you are close to correct settings. If it bucks and stumbles, then it is going too lean and you need to richen that part of the table. Before tuning decel [or accel], make sure you have your VE table close to correct first! One way to do this is to get the VE table set up by setting the TPSdot threshold setting very high (250 kpa/sec or 250 %/s, something like that) so that TPS enrich/enlean never kicks in. Then, (in steady state) set up VE table."

So, it would be my guess that there is a x10 error in the code and/or something isn't allowing the code to go higher than 25. I just thought I'd throw that out there because it seems like 25 is the magic number, but it's just off by 10...

Thanks!
Mike

[jsmcortina]

Can you put my MSQ in and run this on the JimStim and see if there is something affecting the 'time after' code? I am running the Beta from 7/3/07, though it has been present on the alphas from 6/10 as well.

As far as I can see, it looks great.

I can't qualify the exact time as my scope is slightly broken and the timebase is out of calibration.

James

[Keithg]

James,

If you look at the MSQ, the rpm trace shows when it was misfiring. It is too violent to drive. It appears to happen at >3000 rpm and MAP > 160 (probably more like 180) If I sneak up on rpm by slowly opening throttle, it is less pronounced.

If I set dwell time to 1ms, the point that it bucks and misfires is lower in load, but at the same RPM, I believe. To get it mostly drivable, I need to set it to 0.4 which is actually less than suggested by Saab ( 0.5ms is considered minimum). If you run 1ms, you get a longe spark duration which helps lighting a wet mixture on the top end and helps idle misfire at the low end... I never experimented much with dwell time, but know that

As for the dwell time after spark, are you sure it is 1ms when it is 1ms? I have my MS1 back on the bench and I'll see what it is, actually. From the code mod that you suggested last year on MS1, it should be 1ms and I never had a problem to redline in the year I was running MS1/E with this mod (029 early on and HiRes MS1/E a bit later...)

(from thread http://www.msextra.com/viewtopic.php?p=114722#114722 )

Code: Select all

; don't use udvd32 - wasteful, only need 24/8bit divide 
clrh 
ldx #100T 
lda dwelltmpX 
div ;A rem H = (H:A) / X 
sta dwelltmpX 
; lda dwelltmpH 
; div 
; sta dwelltmpH 
; lda dwelltmpL 
; div 
; sta dwelltmpL 
; try to drive Saab DI ;kg 
clra ;kg 
sta dwelltmpH ;kg 
lda #10 ;kg 
sta dwelltmpL ;kg 

lda dwelltmpX 
beq dwelldiv_end ; too long, rail to max 
lda #255T 
sta dwelltmpH 
sta dwelltmpL 

dwelldiv_end: 

On 7/15/07, jsmcortina <jsm@jsm-net.demon.co.uk (jsm@jsm-net.demon.co.uk) > wrote:

Keithg wrote: Can you put my MSQ in and run this on the JimStim and see if there is something affecting the 'time after' code? I am running the Beta from 7/3/07, though it has been present on the alphas from 6/10 as well.
As far as I can see, it looks great.

I can't qualify the exact time as my scope is slightly broken and the timebase is out of calibration.

James

[jsmcortina]

As for the dwell time after spark, are you sure it is 1ms when it is 1ms?

Like I said, my scope is broken. But when I wrote the code it was.

I can't immediately see a problem on the bench, so it is over to you to replicate it, I'm afraid.

This isn't real dwell time - seeing as it is after the spark.

James

[muythaibxr]

So, it would be my guess that there is a x10 error in the code and/or something isn't allowing the code to go higher than 25. I just thought I'd throw that out there because it seems like 25 is the magic number, but it's just off by 10...

Thanks!
Mike

There is no error that I can tell. 255 kPa/sec is the maximum for MAPdot threshold, and 25.5 is the max for TPSdot. I will most likely be changing the variables used so that larger numbers can be entered, at least for TPSdot.

Ken

[Keithg]

James,

I'll do my best. If I set MS2/E it as single spark output and distributor input, it will show all sparks falling on the single output. I can verify this for time base and compare tom MS1/E on the stim.

I cannot enter 0.1ms in MT (testing to see if the bucking goes away here to do a simple test w/o scope to see if there is a x10 issue with the time base.)

Can I edit the ini and allow 0.1ms? Hopefully, I'll have time tonight to do this on the stim.

KeithG

As for the dwell time after spark, are you sure it is 1ms when it is 1ms?

Like I said, my scope is broken. But when I wrote the code it was.

I can't immediately see a problem on the bench, so it is over to you to replicate it, I'm afraid.

This isn't real dwell time - seeing as it is after the spark.

James

[muythaibxr]

I should be able to verify this on my scope tomorrow.

[Keithg]

thx, Ken!

I am at a loss as to what it really is. It does not feel like a rev limiter as it is so violent and you can actually go through the rpm range at low throttle with no to few apparent issues. It does not feel to be fuel related as I am now much leaner than before and it knocks and my controller pulls boost. When I was rich, I thought I was blowing out the spark, but I do not know how violent that would be or if it would feel more like a rev limiter or not. The character does change when I put in different time numbers in the bin. At 0.4 it is mostly drivable, though I can still hit it. At 1.0ms in the bin, it seems to move down to a lower RPM and MAP value which would be consistent with running out of time to spark before the CDI can charge for the next spark. One thing I could try would be to try wasted spark and see if that has any effect. It is definitely repeatable on the car, so we should be able to see something on the stim. The prob is that I do not have a jimstim and have to configure it a bit differently to see any sort of output... I'll put it on my scope, hopefully tonight.

KeithG

[Keithg]

The kids are in bed and the scope is warmed up. Flat lines means it goes to 0V on the output.

MS2/E 20070701
Configured as 'basic trigger' with time after set to 1.0 ms it measures 0.7ms (on my scope) and is very steady. It flat lines at 8500 rpm. 2ms measures 1.4ms and flat lines at 4300 rpm. 2.2ms yields the same trigger time I get with MS1/E, and it goes flat at 3500 rpm. I cannot run more than the one output as I do not have a JimStim.

MS2/E 20070706b and 20070708
Woah! time after spark does not appear to work in these versions. The waveform is inverted and it appears to be normal dwell. The time spent at 0V varies throughout the RPM range. I reverted to 20070701 and all appears well again, so neither of these versions appear to work with 'time after' dwell.

MS1/E HR09C
Configured as 4 spark outputs with stim for wheel. The code 'should' give 1ms, but my scope measures it as 1.6ms. When I reduce the spark outputs to 1, I get the same thing... Ditto when I set it as MSnS mode with spark A only. At 13krpm, the time it is pulled down to 0V lengthens by 0.1ms to 1.7ms.

I believe the problems I am having with 20070701 is due to the fact that I cannot run enough duration and get it to go through 3500 rpm. With short duration (0.2ms) I get weak spark which causes misfire under load and when I get enough duration to have decent idle, I get an undesired rev limiter. The other 2 versions I have not loaded in the car, but from the spark trigger waveform, it does not appear that they will appropriately drive the CDI unit.

KeithG

[rb26dett]

The code 'should' give 1ms, but my scope measures it as 1.6ms

what battery voltage do you have? that could be explained by dwell voltage correction thats present in the new code. i saw someone else post about this, and that was the response.

fred.

[Keithg]

Really, all I was trying to do was to do an even compare. Even if my time base on my scope is off, I wanted to duplicate what the MS1/E was putting out as I never had a spark problem with that.

The other thing is that this 'dwell mode' is supposed to be a time based pulse with no relation to battery voltage.

KeithG

On 7/17/07, rb26dett <fred.cooke@gmail.com (fred.cooke@gmail.com)> wrote:

Quote: The code 'should' give 1ms, but my scope measures it as 1.6ms

what battery voltage do you have? that could be explained by dwell voltage correction thats present in the new code. i saw someone else post about this, and that was the response.

fred.



ms2,v3,cop,innovate,mazda fe3/fe-dohc 2l 4cyl with stock 10:1 pistons,4 stock coils,4 stock ignitors,rx7 550cc injectors maxed@6600rpm&17psi,custom everything,holset he351cw turbo,44mm ext gate,nis gtr bovs,nis gtr intercooler,70mm lexus throttle,chinese fpr,10may v2 ms2e alpha code

[rb26dett]

ok, didnt know that. possibly a bug then? batt correction included when it shouldnt be?

i'll stay out

fred.