Megasquirt Support Forum (MSEXTRA)

Folks,

Man it's annoying when attachments disappear here. Anyhow, I replaced it, see the first post.

No, but I don't run MLV on the laptop. I run both TS and MLV on my desktop and have never experienced the problem on that machine. The laptop I think uses vacuum tubes for storage. :lol: Kidding aside, it's an SSD but definitely not one of the faster ones. Crucial M4 128GB with SATA interface. Machi...

Not quite sure what you meant by "clearing the folder" but that file is present, dated 8/3 (last time I used TS), not marked read-only, and my user, "system", and "administrators" have full control on it. TS version is 3.0.60.61. It's weird... I just tried turning off t...

Not sure if it's related but I'm seeing the same on my Windows laptop. I'm 95% sure I'm running the latest beta TS but will check and post back.

Not to be rude, but why do you need a sensor that "looks like a normal vr/hall sensor"?

It sounds like what you need is called a "photointerrupter": http://heliosoph.mit-links.info/make-yo ... y-encoder/

Well... um... about that... This is a German (Audi) engine with a rather weird (by the standards of the rest of the planet) coolant-temperature-control setup. The thermostat is located in the return line from the radiator! The effect being that the temp is not controlled as tightly as on other engin...

Ah, I'd actually forgotten I owed you guys a better log.

Here is a longer file and the .msq that it was created with: https://www.thelawsplace.com/2019-01-29 ... nd.msq.zip

Elaw: 1) I think your AE settings need work 2) Testing in some steady-state situations would help. 3) Your PID settings, mainly your P term seems to be pretty far off as there isn't that much response from it. I'd want to see the VE, AE, and EGO settings get closer before we make any more changes t...

What did the other tables look like? if you look at the gain change with error its pretty much the same across the table (2:1 from top to bottom), so that could almost be one parameter. so it could be boiled down to one curve vs rpm, and a gain rate, or "gain break" as we used to call it ...

That is interestingly similar to the scheme I was using for boost control a while back... a scheme which worked very well. There was a small table for each of P, I, and D, where you could set the coefficients based on absolute value of error and RPM. It also had a bias table which I don't think woul...

Ken sent me new firmware a few days ago and asked that I take a datalog with the P, I, and D parameters. So here it is! Along with my current tune. EGO control itself seems to be working fine with the new firmware - I can't discern any change from the previous, but I haven't driven the car much on e...

ego resetting here at what appears to be a steady state WOT pull If you're not using the new test firmware, the EGO correction will reset every time accel enrichment or decel enleanment is triggered. That's what appears to be happening in that log, and to a lesser degree in the other log you posted...

For the situation you're describing, adjusting the lag factor probably will not help. The lag factor affects the delay between the sensor's actual output and what the Megasquirt "sees" (and also smooths the measured value, which is the real goal). The AFR you see in your log is that smooth...

It's not the code that's creating the delay, it's the combination of the engine and the sensor. If you have a 500ms delay with the engine and sensor, even if the code samples/corrects every 1ms, it'll be applying correction now for the AFR half a second ago. And with regular closed-loop correction, ...

Elaw, ego doesn’t stay active, or shouldn’t, it just shouldn’t go to 100% when the even occurs. It should just pause changing correction, then resume. If it is actually changing value during the event, I have some more coding to do. Yeah it looks like I misspoke (miswrote?) above. I examined my log...

So... I received some test code from Ken, and installed it just now and thought I'd report back. The good news: it definitely works as advertised! EGO correction remains functional during AE events. The bad news (for me): EGO correction remains functional during AE events. :lol: I think having EGO c...

This is probably a little over the top, but I always thought it would be cool if you could run a wideband and a narrowband sensor at the same time (connected to different inputs). When the mixture is at stoich (as reported by one sensor), the ECU could cross-check for agreement between the sensors. ...

There's one way I can envision it working better/faster, or let's just say "differently"! :) One thing that could be perceived as a fault in a PID controller is the amount of correction depends on the error. So as the error approaches zero, the rate at which the controller corrects decreas...

The one time it gets a little tricky is at WOT when you may want the correction to go to 100%. I handled that by having the code save the current EGO correction value when WOT was detected, then setting the correction to 100%. When the WOT condition ended, it would restore the previous correction va...

It wouldn't be as much of a problem if it didn't reset to 100 every time AE was triggered or any of a number of other conditions was met. When I was running my self-modified firmware, instead of resetting to 100 at those times I had the correction "freeze" where it was, then it would just ...