MS2/Extra Alpha 3.0.3 20090627 (was 20090623)

This is a forum for discussing the development and testing of alpha MS2/Extra code. Documentation
(Runs on MS2 and Microsquirt)

Moderators: jsmcortina, muythaibxr

Re: MS2/Extra Alpha 3.0.3 20090627 (was 20090623)

Postby racingmini_mtl » Wed Jul 01, 2009 7:15 am

Thanks for the report. Nothing wrong with having a good running engine with EDIS or something else. And taking it step by step is a smart move when you can do it.

Jean
jbperf.com Main site . . . . . . . . . . . . . . . . . . . . . . jbperf.com Forum
Image
racingmini_mtl
Super MS/Extra'er
 
Posts: 8039
Joined: Sun May 02, 2004 6:51 am
Location: Quebec, Canada

Re: MS2/Extra Alpha 3.0.3 20090627 (was 20090623)

Postby warp21 » Thu Jul 02, 2009 5:03 am

So far all of the necessary board modifications are directed at completed boards and what needs to be removed from them to make this work. What if you're starting from scratch, could you completely omit the injector driver circuitry and tap into holes 2 and 4 of U4 for injector 1 and 2?
warp21
Helpful MS/Extra'er
 
Posts: 74
Joined: Tue Oct 30, 2007 2:56 am

Re: MS2/Extra Alpha 3.0.3 20090627 (was 20090623)

Postby Keithg » Thu Jul 02, 2009 5:50 am

warp21 wrote:So far all of the necessary board modifications are directed at completed boards and what needs to be removed from them to make this work. What if you're starting from scratch, could you completely omit the injector driver circuitry and tap into holes 2 and 4 of U4 for injector 1 and 2?


I would build the injector driver circuit so that you can use the stim to see if it is all working. If it does not bother you that the injection leds do not flash you can leave that out. You can wire some jumpers on the board to drive the outputs, or manually connect dome jumpers externally depending on how you have it wired. As for using the inputs to U4, I'd have to read the schematics, but I think that would be fine. Make sure you remove R14 and R17, though.

KeithG
Keithg
Super MS/Extra'er
 
Posts: 2413
Joined: Sun Mar 06, 2005 9:15 am
Location: Chicago, IL, USA

Re: MS2/Extra Alpha 3.0.3 20090627 (was 20090623)

Postby racingmini_mtl » Thu Jul 02, 2009 7:24 am

The instructions are not for already built boards. They are the same for new board builds except that you may omit all of the injector driver components. And you definitely can't use pins 2 and 4 of U4 because these correspond to the output of the NAND gate on the MS2 board. You need the input side which is why there is a need to perform the small mod on the MS2 board and why you need to connect the new injector drivers to the places mentioned on my site (and in the other posts here on the subject).

So for a new board build, you need the MS2 mod and you may leave out all the injector driver components but you need to connect the 4 injector drivers to the pads shown in the picture on my site. To test this, you'll need to connect the injector wires to some LEDs or you can use the ignition LEDs on a JimStim if you have one but you'll have to have some adapter to connect to the JimStim pin header. But you can also do what Keith suggest and build the injector driver circuit to test the whole thing with a stim in the normal way (or in semi-sequential mode). There won't be any interference between the 2 driver circuits but you'll have to connect the injectors only to the correct ones depending on the injection mode you configured (sequential or not).

Jean
jbperf.com Main site . . . . . . . . . . . . . . . . . . . . . . jbperf.com Forum
Image
racingmini_mtl
Super MS/Extra'er
 
Posts: 8039
Joined: Sun May 02, 2004 6:51 am
Location: Quebec, Canada

Re: MS2/Extra Alpha 3.0.3 20090627 (was 20090623)

Postby md95 » Sat Jul 04, 2009 11:42 am

I've got a really strange issue i'm trying to figure out...I mostly notice this on the highway, but at a certain steady TP position, under slight accel, my AFR's go crazy back and forth, mostly real rich. I datalogged it, and didn't see any accel or decel events going on, but you can see the AFR and EGO going nuts, and also MapDOT. If someone can take a look and maybe see something i'm not? It's very strange. I space bar marked it with like 5 or 6 hits right together so you can find it. Time is at roughly 8645 seconds. I really can't see what's causing this. Never had this issue with non-sequential code btw.
-Matt-
1998 Mitsubishi Eclipse RS 420A Turbo- MS2 Extra- 2.1.0 Release
1992 Plymouth Laser RS 4g63 AWD Turbo MS2 Extra- 3.0.3s
md95
Master MS/Extra'er
 
Posts: 721
Joined: Tue Mar 15, 2005 6:44 pm
Location: Grand Rapids, MI

Re: MS2/Extra Alpha 3.0.3 20090627 (was 20090623)

Postby Keithg » Mon Jul 06, 2009 8:04 am

Uploaded 20090627 this morning. Seems fine. I halved the crank PWs and it started right up.

One thing I noticed but do not yet have a log to illustrate. If the car is partially warmed up(edit: or completely warm), it can have an erroneous EGO reaction as it comes off ASE. EGO will lean way out then correct. I have also noted that it will grab some odd value in AFR which will cause an EGO amount (97% or whatever) and it will retain this for a few seconds after it catches.

I think this is a remnant of when Ken removed the EGO lockout from the code. I do not want the EGO lockout back in the code as it was. I have my LC-1 set to report 14.7:1 until it is warmed up. So, it would be nice if EGO would be off during cranking and whatever AFR is reported that GEGO would do nothing, but that when AFRs start deviating that it would react and do something. The old way of waiting 30 seconds is not really needed, IMO. Also, this allows me to key on to warm up the LC-1 and see what is reported during startup.

KeithG
Last edited by Keithg on Wed Jul 08, 2009 11:10 am, edited 1 time in total.
Keithg
Super MS/Extra'er
 
Posts: 2413
Joined: Sun Mar 06, 2005 9:15 am
Location: Chicago, IL, USA

Re: MS2/Extra Alpha 3.0.3 20090627 (was 20090623)

Postby Keithg » Wed Jul 08, 2009 11:09 am

I have a log of the odd startup with GEGO and AFR.

It primes, I crank and it starts. AFR is 14.7 because the LC-1 reports 14.7 until it warms up. But, right after startup, GEGO goes immediately to 98% and holds there for a while. It is probably because the value of AFR as it transitions from cranking is 14.6 which would generate about 2% correction. It should hold at 100% until the AFR deviates from 14.7, right? At about 48919s (in the first log) is where the LC-1 finally warms up. And then the MS begins to regulate. The second log is worse as it goes to 94%. Like I said, I do not want it to wait an undefined period of time which I cannot change, but it would be helpful if I could set it to wait 1 second instead of 30 before it tries to adjust GEGO. It appears that the EGO correction routine is grabbing the AFR value right after the crank bit is cleared and the MS is reading it at 14.5 instead of 14.7 and it does not seem to react when it comes back to 14.7. I will change my LC-1 ground to see if I can make it so that its output does not deviate during cranking, but that may be a tall order. We are talking about 0.1V deviation on what the MS is reading of the output of the LC-1...

KeithG
Keithg
Super MS/Extra'er
 
Posts: 2413
Joined: Sun Mar 06, 2005 9:15 am
Location: Chicago, IL, USA

Re: MS2/Extra Alpha 3.0.3 20090627 (was 20090623)

Postby md95 » Thu Jul 09, 2009 8:37 am

Anyone have a look at my datalog yet? It seems that PW1-4 start to deviate from one another more than when the AFR is flat/smooth. I dont know if its a code issue or not, but I dont see what could be causing a crazy AFR back and forth and EGO going nuts. It only happens under slight accel, at highway speed, just like to accelerate slightly with a slow throttle advance, but holding the throttle down as it accelerates(basically giving it slightly more throttle then holding it there to accelerate). It goes uto normal if I just back off the throttle slightly. I can't figure it out!
-Matt-
1998 Mitsubishi Eclipse RS 420A Turbo- MS2 Extra- 2.1.0 Release
1992 Plymouth Laser RS 4g63 AWD Turbo MS2 Extra- 3.0.3s
md95
Master MS/Extra'er
 
Posts: 721
Joined: Tue Mar 15, 2005 6:44 pm
Location: Grand Rapids, MI

Re: MS2/Extra Alpha 3.0.3 20090627 (was 20090623)

Postby PaulS » Thu Jul 09, 2009 9:24 am

I've not studied your log, but I had a similar problem and it was due to AE.

Sometime during the development of 2.1.0 code, the AE settings were multiplied by ten, presumably to increase resolution. This has carried through into 3.0.3, but it was not in 3.0.2 that I started with.

In my case TPSDot was set at 20 and was initiating AE at very light throttle. Any noise on the TPS signal and it would not even idle.

Check TPSDot is at 200 or so.
PaulS
Helpful MS/Extra'er
 
Posts: 81
Joined: Mon Dec 31, 2007 10:57 am
Location: Northants UK

Re: MS2/Extra Alpha 3.0.3 20090627 (was 20090623)

Postby md95 » Fri Jul 10, 2009 5:27 pm

PaulS wrote:I've not studied your log, but I had a similar problem and it was due to AE.

Sometime during the development of 2.1.0 code, the AE settings were multiplied by ten, presumably to increase resolution. This has carried through into 3.0.3, but it was not in 3.0.2 that I started with.

In my case TPSDot was set at 20 and was initiating AE at very light throttle. Any noise on the TPS signal and it would not even idle.

Check TPSDot is at 200 or so.



That's what I thought originally before I took the log, but it shows no TPS accel. Right now TPSdot threshold is at 75.0, any higher than that then I dont get any AE with light throttle stabs, and the engine stumbles very slightly. It really does seem like AE is triggering, but the log doesn't show that, at least what I see anyway.

Why was TPS made more sensitive anyway..I never had any issues over the years past with any accel issues or tuning. It worked so good I tuned it once, then forgot about it.

The strange thing was I never had this issue until I used the sequential injection code. I didn't have this issue using the latest beta code with the more sensitive TPS settings. Jean, is there a way to maybe double check the code/math for AE, just for peace of mind? It just seems that it's even MORE sensitive than it should be. I have to use 57% decel fuel to not have any rich studdering too..I've never had to use any decel!
-Matt-
1998 Mitsubishi Eclipse RS 420A Turbo- MS2 Extra- 2.1.0 Release
1992 Plymouth Laser RS 4g63 AWD Turbo MS2 Extra- 3.0.3s
md95
Master MS/Extra'er
 
Posts: 721
Joined: Tue Mar 15, 2005 6:44 pm
Location: Grand Rapids, MI

Re: MS2/Extra Alpha 3.0.3 20090627 (was 20090623)

Postby racingmini_mtl » Fri Jul 10, 2009 9:08 pm

I'll have a look but that's a completely separate part of the code than what I have touched with the sequential code. Also, you seem to be the only one suffering from it so I would think it's a different issue. I'm not saying it's not a code issue but I doubt very much it's an AE code issue.

Can you post your msq so I can see if I can reproduce it on the bench.

Jean
jbperf.com Main site . . . . . . . . . . . . . . . . . . . . . . jbperf.com Forum
Image
racingmini_mtl
Super MS/Extra'er
 
Posts: 8039
Joined: Sun May 02, 2004 6:51 am
Location: Quebec, Canada

Re: MS2/Extra Alpha 3.0.3 20090627 (was 20090623)

Postby md95 » Sat Jul 11, 2009 5:24 am

racingmini_mtl wrote:I'll have a look but that's a completely separate part of the code than what I have touched with the sequential code. Also, you seem to be the only one suffering from it so I would think it's a different issue. I'm not saying it's not a code issue but I doubt very much it's an AE code issue.

Can you post your msq so I can see if I can reproduce it on the bench.

Jean


Msq attached. I just can't seem to figure it out..It just doesn't make sense that the issue is only present with sequential..weather it's a code issue or not, I have no idea.. I just find it odd that I have to cut back all my ms/% values severely, over what they were running latest beta code and non sequential fuel. That and I have to run so much decel fuel cut..Since running sequential, i've spent many hours with AE trying to get it perfect. It's close, but still with the issues. Other than that, it runs beautifully!

I was also thinking it would be nice to have separate threshold settings for accel and decel, instead of only one. I may also try using map based AE and see what happens.
-Matt-
1998 Mitsubishi Eclipse RS 420A Turbo- MS2 Extra- 2.1.0 Release
1992 Plymouth Laser RS 4g63 AWD Turbo MS2 Extra- 3.0.3s
md95
Master MS/Extra'er
 
Posts: 721
Joined: Tue Mar 15, 2005 6:44 pm
Location: Grand Rapids, MI

Re: MS2/Extra Alpha 3.0.3 20090627 (was 20090623)

Postby Keithg » Sat Jul 11, 2009 6:04 am

Matt,

FWIW, I was never able to get regular AE to work well on MS2 and went to EAE. I have not tried AE in a long time, though and it may be much better in 2.1.

KeithG
Keithg
Super MS/Extra'er
 
Posts: 2413
Joined: Sun Mar 06, 2005 9:15 am
Location: Chicago, IL, USA

Re: MS2/Extra Alpha 3.0.3 20090627 (was 20090623)

Postby md95 » Sat Jul 11, 2009 6:49 am

Keithg wrote:Matt,

FWIW, I was never able to get regular AE to work well on MS2 and went to EAE. I have not tried AE in a long time, though and it may be much better in 2.1.

KeithG


Really? I've *never* had any issues with TPS based AE in any MS installs..I've done quite a few now. It's always worked perfectly, and maybe spent 15 minutes at most to get it dialed in, then thats it. Of course, i've only done 4 cylinder cars, nothing odd or complicated. Maybe it's because sequential injection is that much more efficient at fueling that I had to cut way back on AE settings..I'm baffled at the moment.
-Matt-
1998 Mitsubishi Eclipse RS 420A Turbo- MS2 Extra- 2.1.0 Release
1992 Plymouth Laser RS 4g63 AWD Turbo MS2 Extra- 3.0.3s
md95
Master MS/Extra'er
 
Posts: 721
Joined: Tue Mar 15, 2005 6:44 pm
Location: Grand Rapids, MI

Re: MS2/Extra Alpha 3.0.3 20090627 (was 20090623)

Postby UnaClocker » Sun Jul 12, 2009 4:12 pm

racingmini_mtl wrote:But as I said, the 2.2/2.5 wheel has a real issue which needs to be corrected because it does affect both ignition and injection timing.

Any progress on this? I know Ken & James are hard at work on the MS3 code, but seems like this should just be a quick fix, just a logic error or something. It's definitely needed because there are many cars running this trigger mode now, and I imagine it's partly broken in the 2.1.0 release code as well..
Brian
'84 Dodge Rampage
User avatar
UnaClocker
Super MS/Extra'er
 
Posts: 1933
Joined: Fri May 07, 2004 12:59 pm
Location: Tacoma, WA

Re: MS2/Extra Alpha 3.0.3 20090627 (was 20090623)

Postby racingmini_mtl » Sun Jul 12, 2009 5:20 pm

Thanks for reminding me and I'll look into that ASAP. Sorry for the delay.

Jean
jbperf.com Main site . . . . . . . . . . . . . . . . . . . . . . jbperf.com Forum
Image
racingmini_mtl
Super MS/Extra'er
 
Posts: 8039
Joined: Sun May 02, 2004 6:51 am
Location: Quebec, Canada

Re: MS2/Extra Alpha 3.0.3 20090627 (was 20090623)

Postby racingmini_mtl » Sun Jul 12, 2009 8:00 pm

Ok. I have checked this out and the issue I was seeing was a mismatch between the JimStim signal and the wheel definition in the code. Now, I assume that the wheel pattern in the MS2/Extra code is the correct one but from James' comment in the JimStim wheel pattern (Unaclocker's wheel), Brian, you should be the one telling me.

With the same pattern the code expects, the timing is perfect with either ignition or injection. The interesting thing is that the injection timing was fine when using start-of-pulse timing with the mismatch but not when using end-of-pulse or mid-pulse. And since ignition uses something similar to end-of-pulse (or rather the other way around), that was the reason ignition was also not behaving correctly.

Since the mismatch was not that big and resulted in problems similar to what was seen by ashford, it makes me think that his problem was either caused by noise or that there is actually a discrepancy between the actual pattern and what the code expects. The best thing would be to get a composite log to see what the signal received looks like.

Jean
jbperf.com Main site . . . . . . . . . . . . . . . . . . . . . . jbperf.com Forum
Image
racingmini_mtl
Super MS/Extra'er
 
Posts: 8039
Joined: Sun May 02, 2004 6:51 am
Location: Quebec, Canada

Re: MS2/Extra Alpha 3.0.3 20090627 (was 20090623)

Postby ilotalo » Sat Oct 10, 2009 5:09 am

md95 wrote:I've got a really strange issue i'm trying to figure out...I mostly notice this on the highway, but at a certain steady TP position, under slight accel, my AFR's go crazy back and forth, mostly real rich. I datalogged it, and didn't see any accel or decel events going on, but you can see the AFR and EGO going nuts, and also MapDOT. If someone can take a look and maybe see something i'm not? It's very strange. I space bar marked it with like 5 or 6 hits right together so you can find it. Time is at roughly 8645 seconds. I really can't see what's causing this. Never had this issue with non-sequential code btw.



Hi!

Did you got any results for this problem? I got the same issue. In slow acceleration afr start jumping. I was thinking is it squirt with all injectors at same time. That makes sens to me. So bug might be in sequental code.

janne
ilotalo
MS/Extra Newbie
 
Posts: 2
Joined: Mon Jun 15, 2009 11:26 am

Re: MS2/Extra Alpha 3.0.3 20090627 (was 20090623)

Postby md95 » Sun Oct 11, 2009 8:34 am

ilotalo wrote:Hi!

Did you got any results for this problem? I got the same issue. In slow acceleration afr start jumping. I was thinking is it squirt with all injectors at same time. That makes sens to me. So bug might be in sequental code.

janne


PM sent..
I'm still having the issue as well, and i've been doing lots of testing and driving in the meantime. I will say that I have not yet tried any updated releases of alpha code to see if this issue has "resolved itself".. I have noticed that by changing the injection timing point(end of pulse timing), that I can change the pressure point to where this happens. I've found that with negative timing values, the issue happens in vacuum( say 15-100 kpa). If I set it for positive values, it happens slightly above 100 kpa. Unless you have a wideband gauge display or are watching the AFR gauge on the laptop while driving, you may not even notice this happen. When I had the timing point set at a certain negative value, I was able to get it to happen at around 25-30 kpa light load, and you could really feel it happen in the engine(lulls, studders in the engine). When it set so it happens in higher load points, its much less noticaeble. All I know is that it's not a tunning issue. I've offered up as much info as I can, and am just living with it for now. A bit odd that another 4g63 user sees the same issue..
-Matt-
1998 Mitsubishi Eclipse RS 420A Turbo- MS2 Extra- 2.1.0 Release
1992 Plymouth Laser RS 4g63 AWD Turbo MS2 Extra- 3.0.3s
md95
Master MS/Extra'er
 
Posts: 721
Joined: Tue Mar 15, 2005 6:44 pm
Location: Grand Rapids, MI

Re: MS2/Extra Alpha 3.0.3 20090627 (was 20090623)

Postby UnaClocker » Mon Oct 12, 2009 10:55 pm

You really should try the new code (and this thread should get unstickied...). There's some new map sampling routines in the latest alpha that does wonders for smoothing out the map. Should help your problem out.. :)
Brian
'84 Dodge Rampage
User avatar
UnaClocker
Super MS/Extra'er
 
Posts: 1933
Joined: Fri May 07, 2004 12:59 pm
Location: Tacoma, WA

PreviousNext

Return to MS2/Extra Development

Who is online

Users browsing this forum: TurboBob and 1 guest