ms2 extra ignition

[EssEss]

I know they don't use output compare for their ignition (well, I was warned of this) .. and so it was ripe for inaccuracies .. this is why I wanted to check it first. I've been told many times by other folks that ignition control is solid - I just wanted to find out for myself. I didn't expect much, nor am I really impressed, nor do I call it solid. This is why I preach using output compares/match's in hardware for doing something like this. You obviously do embedded dev to understand this so well (as stated above). This was steady state testing and the spec couldn't even be met - jitter exceeds +/-.1* not even when an input capture is present.

If this was shortcuted, maybe I should spend more time actually digging into the code looking for other red flags. My doubts are starting to grow again, maybe I just expect too much. I have my doubts of even mentioning this to their forum .. it took a few weeks to get a simple response about their code plans for ms3 (this was after direct email attempts to the developers for weeks before that) - so they're already not receptive to me or know I comment here ocasionally.

[johntramp]

You should take it to their forums even if you don't expect them to be receptive of it. At least then other users are aware of the situation.

[cpuTorture]

IMO You'll just be bashing against a wall. Face it MS was a hack, is a hack and always will be a hack. Have a look at the lovely code layout the perfect naming of variables and constants for settings. The only thing intuitive about it is a dog ate it at some stage. MS is great for what it is. We are just expecting too much.. I just don't think your going to get .1 degree accuracy at 10krpm.

looks like a spider on crack weaving a web :-

Code: Select all

    if((flash4.EgoOption == 2) || (flash4.EgoOption == 4))  {
        if (flash4.ego2port < 0x7) {
            if ( ((adctest & 0x80) & (flash4.ego2port == 7))
                    || ((adctest & 0x40) & (flash4.ego2port == 6)) ){
                conf_err = 2;
            } else {
                if (flash4.ego2port == 6) {
                    adctest |= 0x40;
                } else if (flash4.ego2port == 7) {
                    adctest |= 0x80;
                }

Regards
Jay

[EssEss]

Most people allow themselves to live with a certain amount of bugs and/or relaxations in specs for the things that they buy for a lower cost. I live to a bit of a higher standard which is why I'm so critical of specs and correct operation. If a manufacturer says he can do it, then he better damn well do it. msqrt costs have gone up but their standards haven't followed.

you're preaching to the choir - I've seen beter code - but watch out who you say that to, they don't like their baby being called ugly .. even if it really is ugly. Call me a masochist - but I do enjoy reading bad code, if you can read it, you can read anything. sometimes its the 'hard path' of wierd constructs and poor formatting that really open your eyes and force you to concentrate. The 'good looking stuff' trips me up a lot, you get complacent during reviews and gloss over the red flags easier.

maybe as step 2, I'll look over the code and offer a fix (if possible) before I post on their forum. my time is thin, so I don't think this will happen.

[Fred]

Jay/cpu, I edited your post to use code tags to wrap the code. I hope you don't mind. You might wanna use them in future, they help a lot :-)

EssEss, magic numbers, mostly, are evil, that "setting.blah = 3.14159" shlt is unacceptable, really.

Fred.

[Fred]

Bump for a ROFL worthy thread! Especially cpuTorture's comments two posts above! I've been giggling about that comment for a week now! :-)

[tpsretard]

i am about to loose cylinder #4 on my fresh STI engine before of "features*cough, cough* i mean bugs" in the megasquirt.

been saying for ever there was a problem, but only now is it been looked at
i agree we just expect to much.

looks like a spider on crack weaving a web

Jay you are a fucking star!!!!

[Fred]

This thread was recently linked on msextra.com forum and some moron was ranting on about it and ignoring all of the, for a change, accurate things that James, Ken, and Jean were saying. Firstly, I don't know if he's even a registered user here. Secondly, I hope he's not. Thirdly, he's smoking some great drugs, if anyone has his IP (to ban) or dealer's number (to buy), I'm available at the usual email address!

http://msextra.com/forums/viewtopic.php?f=131&t=39578

What EssEss found above is benign and not dangerous, unlike the sequential fueling bug that killed a few engines :-(

Everything they say about single/dual core and IC/OC/multiplex/dynamic timing, etc is all right. Reality is that when you free rev a normal engine with a moderate number of teeth you don't have a large error anyway. And the error you do have under acceleration (high load) is LESS advance than requested, so it too is benign.

Fred.

[EssEss]

I'd love to know who that troll 'Benkku' is .. I'd swear he knows what he's talking about.

[Fred]

You really think so? I think he believed what he said! :-/