[FW MOD] ms2extra 3.2.1 beta gslender v2

[gslender]

where do i find the option for rpm smoothing?

Under idle enhancements.... Note that it only smooths the rpm input used by pid, which in turn smooths pid and in turn the idle.

G

[muythaibxr]

As far as this smoothing goes... if it doesn't hurt response and does a good job filtering compared to the lag factors, then we might make it smooth RPM (and MAP, and TPS, and...) for all operations and get rid of the lag factors (as they won't be necessary anymore).

Ken

[Zaphod]

As far as this smoothing goes... if it doesn't hurt response and does a good job filtering compared to the lag factors, then we might make it smooth RPM (and MAP, and TPS, and...) for all operations and get rid of the lag factors (as they won't be necessary anymore).

Ken

I can't believe the dynamics that the MS2 (and off course 3) development has got over the last months. Great work - all of you!
It's nice to see how far we have got with this cooperative work of you in the last time. Big kudos!

[gslender]

Oh, I forgot to explain that this release adds something I've called Adaptive idle advance - it is essentially rpm error to target and allows advance to be adjusted right from cold through warmup and onwards, as it will track to target idle rpm.

If idle goes -100 rpm below target it can add 10 deg advance, and if idle goes 100 rpm above target, you can soften with -8 deg advance. When coupled with a little idleup rpm and duty adder, you get solid catch of the AC load.

Side benefit is super stable idle... I'm getting idles where rpm varies only +/- 25 rpm from target, with a good solid % of the time within 10 rpm from target. When coupled with pid rpm smoothing, I now find the iac valve sits with virtually no movement at all... Which means pid isn't jumping around trying to keep the valve and idle on target.

It is really good and I'd encourage you to try Adaptive and lower idle advance clt entry to start from cold.



G

[93white3400z]

Some of these changes and the feedback people give almost makes me want to downgrade to ms2 just to try them Hope some (all?) of them will make their way to ms3

[muythaibxr]

Many of the changes actually came from Grant (with my permission) backporting MS3 code, so those are already in the 1.1 alphas. Really the only one not there is his adaptive idle advance (which I requested he change from the original method) and I guess the RPM smoothing for PID. I will be reimplementing both for ms3 unless they are easily portable from his ms2 implementations as well as backporting more of the ms3 fixes to ms2.

We are also working on boost control improvements, MAP sampling and smoothing changes, a new, more responsive and easier to tune "accel-pump ae," alternator control, and a number of other enhancements.

Ken

[juansh2385]

Many of the changes actually came from Grant (with my permission) backporting MS3 code, so those are already in the 1.1 alphas. Really the only one not there is his adaptive idle advance (which I requested he change from the original method) and I guess the RPM smoothing for PID. I will be reimplementing both for ms3 unless they are easily portable from his ms2 implementations as well as backporting more of the ms3 fixes to ms2.

We are also working on boost control improvements, MAP sampling and smoothing changes, a new, more responsive and easier to tune "accel-pump ae," alternator control, and a number of other enhancements.

Ken

Alternator control?

I hope I am understanding this correctly if I am this mean that the voltage regulator that i had to buy to use only ms in my eclipse 420a won't be necessary any more?

[muythaibxr]

If you're using MS3... alternator control won't be backported most likely. The rest of the features have a good chance of being backported though. I'm mainly going to start controlling it so that when it needs to be in higher output mode, I am controlling it and can feed that forward to idle control to give better response to electrical loads

Ken

[Greg G]

As far as this smoothing goes... if it doesn't hurt response and does a good job filtering compared to the lag factors, then we might make it smooth RPM (and MAP, and TPS, and...) for all operations and get rid of the lag factors (as they won't be necessary anymore).

Ken

Already started preparing for the removal of lag factors. Here's a comparison of RPM lag 40 vs RPM lag 100. RPM smoothing factor 10 for both. The RPM shows a lot of small jaggies, and the spark naturally follows. But these are not felt...if anything, the engine feels smoother. The RPM seems to be held to target more strongly. Note the super tight RPM scale



This is steady state idle when hot.

[Greg G]

Some of these changes and the feedback people give almost makes me want to downgrade to ms2 just to try them Hope some (all?) of them will make their way to ms3

The 'new stuff' in the gslender modified firmware is largely backported stuff from MS3. The devs have been incredibly helpful in this process, especially Ken, who has steered us in the right direction several times. So props to the dev team for helping keep MS2/Extra relevant. Having said that, the future does lie with MS3...soon enough there will be stuff that can't be done with the MS2 hardware

[muythaibxr]

There already are several things:

SDcard, VVT, using several of the features that require lots of I/O at once, V8 sequential and COP, rotary support for 3 and 4 rotors, much less jitter on the outputs (since we have the XGATE controlling the time-sensitive outputs) and the list goes on and on.

We'll be able to do advanced features like angle clock as well which should free up a lot of time that is now being used for angle<->time calcs to do other things.

Ken

[gslender]

Some of these changes and the feedback people give almost makes me want to downgrade to ms2 just to try them Hope some (all?) of them will make their way to ms3

The 'new stuff' in the gslender modified firmware is largely backported stuff from MS3. The devs have been incredibly helpful in this process, especially Ken, who has steered us in the right direction several times. So props to the dev team for helping keep MS2/Extra relevant. Having said that, the future does lie with MS3...soon enough there will be stuff that can't be done with the MS2 hardware

Spot on Greg - as you know, I struggled for ages trying to get a new memory page sorted in MS2, but couldn't get it to work (and I almost gave up). Luckily Ken stepped up and volunteered his time to help me figure out the bugs (and this would have been at the cost of MS3 development work).

Ken's also been really good working with me in ironing out AC idleup bugs - that now will not only benefit my work with MS2, but will also now get into MS3 and generally help everyone.

Without doubt, all the best work is in MS3 and I'd encoruage everyone to look into that platform... from a developer perspective, living in MS2 ain't easy and I can't see it lasting forever with improvements going fwd. Sure, it does what it does well, but trying to get some of the things done is really hard work and I can totally understand and support why Ken and co are focused on MS3... it is just that much better in every regard.

G

[93white3400z]

Yeah good job gslender and everyone else that worked on this.

I have a question for you that have an ms2 and are developing for it. I'm wondering what is the motivation to develop on ms2 instead of upgrading and develop on ms3 (and use already existing feature w/o having to backport). Is it a $$ issue or there is something else I'm missing ? Please don't take this the wrong way, I can understand if it's a $$ money, but I'm just wondering if I'm missing something

[Greg G]

Well I just feel the diypnp does everything I need, especially now with the improved idle. I would just love a couple more inputs- then I could request a separate feedforward idle up for headlights and power steering. But as it is, I am pretty happy with it.

I just got into MS a few months ago, so I'm still learning. Am I looking at the ms3? Yes, maybe in the future.

[gslender]

Yeah good job gslender and everyone else that worked on this.

I have a question for you that have an ms2 and are developing for it. I'm wondering what is the motivation to develop on ms2 instead of upgrading and develop on ms3 (and use already existing feature w/o having to backport). Is it a $$ issue or there is something else I'm missing ? Please don't take this the wrong way, I can understand if it's a $$ money, but I'm just wondering if I'm missing something

A few things come to mind... 1) there seems to be more experimental tuning required on the ms3 in terms of pnp to the car I own (MX5), 2) the ability to develop for the ms3 has only recently happend (the source was 1st available Oct 26th), and 3) the total cost for MS3 vs MS2 was something that I initially weren't comfortable paying extra for when I didn't really see how all those extra bits would be needed (not a great big deal).

In saying all that, now that I know what is required to code in ms2 vs ms3, I'd gladly upgrade tomorrow - with the only constraint being that the developers aren't releasing the code for every release like they do for ms2... which whilst I understand why, I think is a shame as that for me was one of the big attractions to megasquirt in the first place.

I personally believe more can be gained by being more open and allowing and encouraging contributions than by keeping it locked up and hidden. There is less minds looking at the code and implementations, less chance of bugs being found, and less thinking occuring as not as much is challenged. So whilst having it open might allow competititors to steal, the real competitive advantage is having a better product and offering - and if you look at some of greatest open software projects out there (mySQL, Snort, Java, GCC, Linux, Firefox, OpenOffice, Apache, etc ) they kick butt not because the best and brightest people do the work in secret, but because they work mostly in an open way, strongly fostering the community around them to make the products the best in the world!!

If there is anyway Ken, James and Jean can get the software more open and still keep the license control they need, then I think MSExtra has the potential to rock the world in the ECU space... as the features and capabilities will surpass anything closed... and this has been proven time over time.

I'm not knocking the guys, I'm just stating my opinion and mine is fairly immature in the ECU space... I've only started looking at this world for the last 6 months, but have a fairly long history developing software in the industry and worked on both closed and open sides.

G

[Zaphod]

Hi I can't use my old MSQ with the new firmware - as the bins in the new functions are all filled with unsensible values like -2.234325243545465466 and when I try to change the value I get an error for the next bin and the error pops up endless when closing. I suppose I will have to make a new MSQ from scratch...?

[Black99rt]

Yeah good job gslender and everyone else that worked on this.

I have a question for you that have an ms2 and are developing for it. I'm wondering what is the motivation to develop on ms2 instead of upgrading and develop on ms3 (and use already existing feature w/o having to backport). Is it a $$ issue or there is something else I'm missing ? Please don't take this the wrong way, I can understand if it's a $$ money, but I'm just wondering if I'm missing something

I'm not devleoping for ms2, but I'll answer anyway. In addition to the money I already have wrapped up in the ms v3.0 and ms2 processor card, as well as wiring, wideband sensors, mistakes, re-wireing, and of course time- I'm not looking forward to shelling out hundreds more for ms3. The entire project would quickly approach the cost of an off the shelf standalone when looked at "big picture" wise. That being said, It is likely within the next two years I will upgrade. No nice dashboard options has me hesistant to spend that money though. I'm hoping to see a mod where the ms3 can make data availible on the CAN line instead of needed to be asked for it.

[jsmcortina]

Hi I can't use my old MSQ with the new firmware - as the bins in the new functions are all filled with unsensible values like -2.234325243545465466 and when I try to change the value I get an error for the next bin and the error pops up endless when closing. I suppose I will have to make a new MSQ from scratch...?

Sounds like you are doing this offline? That would be the problem most likely.
Do this online (i.e. MS2 connected and powered up) and it comes with the default settings pre-loaded with sensible values.
Then load your old MSQ on top and the existing (old) settings will get loaded in leaving the new ones as defaults.

James

[Zaphod]

Ah, thanks. Car is in "Winter-sleep" mode now, so yes I did this offline. Gonna try in the Car.

[Greg G]

Upgraded to the latest version 2 release- take note that you may need to check your output window. Idleupcnt was replaced by tpsADC as the trigger value, must be some memory juggling. Idle is perfect, but I still have some niggling issues on throttle lift situations, especially with AC. Would appreciate more data from others