AbeFM

iTrader: (3)

Awesome, I can't edit my own posts.

Well, here's the link to the 2.1.0 stable release, I've yet to figure out if it's version N or version P or whatever, they are simply not labeled.

http://www.msextra.com/ms2extra/file...e_20090627.zip

My goal is to have the first post with a matching set of firmware and msq to run an average stock car. At this point, I still recommend doing it from scratch.


Here's Marc's 2.1.0 msq
Miata Turbo Forum - Home of the turbo Mazda Miata.


I'm attaching fuel and spark maps. They are probably not conservative enough. But it's a place to start, should get your idling. I'd recommend pulling a couple degrees of timing everywhere, maybe richening up top. You'll want a knocksense before you get too far into this. Just open the VE (or spark) table editor in TS or MT, and "import table", and use these files. I hope to post better ones later, but this is where I am at.

AbeFM

iTrader: (3)

Ok, people, what you may have all been waiting for.

I'm about to leave on vacation (I'll have some internet, no worries), so I finally made the switch to tuner studio.

Here's the deal... I took the "best" MSIIxe V2 software (the october 21st 2009 release of 2.1.1b), flashed it, loaded my own msq, brought it up to date, took some notes, and my car runs great. Quick test drive.

Issues YOU MUST BE AWARE OF:
This is for my car, with my wiring. It may not be yours. I will do my best to check this all later, but I haven't.

1) I have COP
2) I have 550 cc/min, low ohm injectors (this effects accel enrich, too)
3) I set it up to use barometeric correction, then set that correction to 0. I don't know if this worked. So turn it off.
4) I think you guys have VICS on a different pin. Check the write up in the beginning of this thread.
5) I get some knock over 16 psi. This might be a new brand of gas that I just put in.

Yeah, basically, there's lots to look at. I will try to look at it soon, but at least I'm dubbing this release of the firmware as "the official" for this board. Coming up with a solid MSQ to go with it will be a task for the next couple weeks.

Lastly, in case I have to email it, I rar'ed it. So, get winrar or whatever to 'unzip' it. The password is "abe".

Huh. I can't upload it. Well, fine. Email me if you need it. It's under a meg, and would be much smaller but there's a whole firmware release in it.

Zaphod

Send it to me, I upload it for you.

AbeFM

iTrader: (3)

Awesome, sent. I'll try to work out the rest of the wires/settings for your guys. Once I get all that good, I should be able to post an msq as a real test. If people like it, I'll make another zipped packet and that will be the official "base map".

In the mean time, use at your own risk. Hopefully by tonight I'll have an update posted.

AbeFM

iTrader: (3)

Ok, mailing you another. It's bigger, but it's everything in tuner studio and msq and firmware, etc. I think it's right. I'm not 100% on coils, or req fuel, but I think this should work pretty darned well.

Zaphod

O.K. here you go:

Abe's base map V 0.2

Marc D

iTrader: (7)

Abe,

Did you see that there was a new alpha code release yesterday? The new alpha code has a good amount of changes done to the closed loop idle code. I just flashed it and the idle seems to be working a bit better. It still has its quirks and i still gotta tune the PID, but its working better now.

thirdgen

iTrader: (26)

So Marc, what version is that?

muythaibxr

He's talking about 3.0.3n (http://www.msextra.com/viewtopic.php?f=91&t=34166). I changed the way rpmdot is calculated which makes the rpmdot calc more stable, and made it log to the datalog now, so you can actually figure out what's causing idle not to engage or causing it to engage when you don't want.

I also made it so that for the lockout settings, instead of waiting for your PID delay then checking the lockout settings (going into PID as soon as they're met after the delay), it checks them every time through that code during the PID delay, and if at any time the thresholds are not met, the PID delay timer gets reset.

These changes should make it easier to tune the PID disable RPMdot settings (more stable RPMdot helps there) and easier to tune the lockout settings (having rpmdot logged helps there).

I'm hesitant to call these bug-fixes really, as the lockout and PID disable settings could be made to work before, but these should make things easier.

Marc:

Did you make the overrun fuel cut, fuel, and timing settings changes I suggested? Last log I looked at, you were getting into fuel cut and idle speed control wasn't even engaging because of the oscillation fuel cut was causing in one log, and in the other, you were idling near 16:1 AFR, which will make it very difficult to catch sudden load increases, and will make the idle prone to oscillation even if PID isn't doing anything. I also suggested running idle advance so you could increase timing with increased load without having to mess with your main timing table.

Also, don't hesitate to call if you're fighting with it. I think we can get it working pretty well.

Check out this site for some more tips (this is for MS3, but the settings are the same):

2 Idle Control

Ken

AbeFM

iTrader: (3)

I would REALLY try to get a stable, universal, agreed upon set up working, first. I would be wary of anything called "alpha" unless "alpha" doesn't mean alpha, just like a release version of code should be as bug free as humanly possible with no new features.

I WILL NOT ARGUE SEMANTICS IN THIS THREAD.

Let's keep this thread about a BASELINE .msq file, a known starting position so that anyone can run it, and if it doesn't work, they know it's the hardware. I would be VERY happy to discuss the idle and new code more, but let's do it in the existing idle thread, or make a new one, or on the MS forums.

The truth is, if you get your car set up and working 100% with V2.x of the code, and the PWM Warmup idle works, it will then be easy to tell what things are better or worse running the V3.x code.

But again, before people try to do some untested, unproven thing, then come back and say the board is too complicated, megasquirt will never work, and all the other things I see from people running alpha code with and without my board.... Let's make a stable baseline to start from.

So please, untill you're 100% sure everything works how you want, let's all run the v0.2 package Zaphod just posted and once you are sure VICS is switching how you want, the timing is how you want, etc... Then we'll try the new ones.

Ken, since you're here - any suggestions on MAP sampling window settings? I'll say I found the 2.1.1 oct-29 version to be smoother than the 2.1.0 release version. Could be in my head, but I'd like to think all that hard work of yours made a difference. :-)

Also, I'm still running into issues with accel enrich - things get better and better as I add more fuel, until I hit the cut off for the highest numbers allowed. Related - with different injectors, I assume you have to retune this? Am I right? I was thinking since it's in "ms" and not in percent that it would - all my tuning has been with double sized injectors and I'll up the numbers before releasing an MSQ if that's the case.

Let me know how it's going. :-)

muythaibxr

Alpha code is alpha code... Marc D is using it so I responded to his post. There are bugs there in that code... we let a new developer (Jean) do sequential for that code, but over the course of his rearranging to get it to fit, he introduced a few bugs. That said there are people using it just fine, and I am unaware of any bugs that should affect miata users (after my last bugfix release).

Use what works really. My suggestion is to pick the start timing that you think should yield the lowest MAP, then adjust the window to "catch" it.

The way those setting works is that the timing setting sets the time where sampling starts. The MAP is sampled every .128 ms until the end of the window. The lowest value found during the window is passed to the mainloop. So it's timed at the exact same time in the cycle every time, and just in case of acceleration or deceleration, it picks the lowest value in that window to try to get the most consistent value possible.

Just about anything you choose here should give a more stable MAP signal though since no matter what you set it's timing very close to the exact same time in the cycle (2x per rev on a 4 cyl).

That feature was actually fairly easy as far as these things go. You should be able to tell if it's in your head if you compare a log on the old code to one on the new code though.

Yeah, with ms2 you do. We've made those % of reqfuel in ms3 though, so you shouldn't have to mess with them once they're set.

IF you're having trouble with response, you can also increase the accel shot time. Worst case if you're not already running 4 squirts, you can increase the number of squirts per cycle to 4. If your injectors aren't huge and your opening time is set correctly, this can make tuning much easier and response much better.

I personally on the 20v 4age use EAE to get slow to medium speed accel really smooth (feels as good as my friend's factory-tuned 20v 4age), then use a little regular AE to speed up EAE response to quick blips.

Ken

AbeFM

iTrader: (3)

Ok, so it looks like I left the settings for the coils shy - i.e. it'll rev to 2k then stumble to death. Solution: Set the dwell higher. I've been using 5 ms, you might need more, I can look this up.

Also, change the polarity of the warmup LED outputs, so the fans are ON when hot and OFF when cold. :-)

So far, it's running ok. Will have a tune soon. then I'll post a msq-V0.3.
-Abe.

AbeFM

iTrader: (3)

I was figuring everyone would be all over this. :-) I guess it's because its out a couple months AFTER the board instead of before, eh? :-) Either that or this "winter" thing...


Quick informal vote: Who here hates their innovate wide band? It seems everyone I know if having problems with theirs. I certainly had problems with mine for a while.

Marc D

iTrader: (7)

Holy crap, 5ms? Are you using stock coils?

I have COPs, but im using 2.6 ms for dwell.

AbeFM

iTrader: (3)

Yeah - stock coils need a lot of dwell. Someone else had this problem before, and when my friend called me I flashed on it - too little dwell = not revving over 2,500 :-)

redrider706

iTrader: (4)

Oops, just saw this. Yes, I got the pins. Just tested the whole setup using the JimStim including the OEM connector configuration for a 91 and it all seems to work OK.

I will make a post later on some changes for the next edition of the board based on my experience on mod-ding it for a 91.

Hoping to put it in the car in a week or two.

The rar file with the base map, etc.. seems corrupted.
Winrar says the archive is password encrypted.

Can you export you AFR map and post it?

-Raj

AbeFM

iTrader: (3)

The password should be included earlier in the thread - I know it works. The password is "abe"

Cross

iTrader: (4)

So abe I have been playing around today with this version posted here around last years time frame...

But I have on of your units and this seems to be the best info thread... I was hoping Maybe I could get it moving again for some help.

I am working with a 2000 and the more I ready through this the more I notice talk of VICS which I obviously do not have.

I am having an issue even after flashing over to your setup here from what I was using and changing the dwell as well as the ve, afr and spark tables to my na ones.
My idle is 3-4k jumps or sits depends on what I am playing with and is not going down.

I am including my latest msq from right now, hopefully it helps.

But basically I would like to find out whats going wrong so I can get over this hurdle. I was about to give up after a scare tonight with the car heating up... but I really want to get back to a MS setup.

Here is my MSQ:
http://www.fileden.com/files/2009/6/...9_23.35.57.msq

Any help with board settings and/or what they should be for a 2000 in terms of the jumpers on your PNP unit and anything else you can help me with would be great!

Thank you.