MS2/Extra 2.0 beta 16 (20071122)

Post by **muythaibxr** » Thu Nov 22, 2007 8:57 am

Time for the next beta. Fixes include:

1) More fixes for microsquirt
2) Added a prediction algorithm for ignition
3) Added another error condition that gets reported in bad configs
4) Fixes for PID idle, one of which should fix an occasional bug in PWM PID, the rest to fix problems with IAC PID.
5) Fixes for twin trigger
6) Some comments were updated to better reflect what the code is doing (not a functional change)

NOTE: In all of my testing of IAC, it is very slow, which may cause problems for it responding to sudden loads. I have a few ideas that may help this, and will tinker with those for the next beta.

As usual, please let us know of any probems.

The beta is here:

http://www.msextra.com/ms2extra/files/b ... 071122.zip

Ken

hassmaschine · Post by **hassmaschine** » Thu Nov 22, 2007 10:21 am

I'll test it out on my way to the grocery store. lol

thanks!

Keithg · Post by **Keithg** » Thu Nov 22, 2007 11:55 am

What problems with PID PWM were you chasing? Just curious. I have it working well and notice a few oddities, but it is imminently useful in Beta 14 and before.

What does the prediction algorithm do? Is there any way to make it sync faster during cranking. I have a hall which has clean output to 0 rpm, and it seems to take too many revolutions before it syncs. Even wasted spark seems to sync slower than MS1/E, or maybe I am imagining that.

KeithG

Post by **muythaibxr** » Thu Nov 22, 2007 12:39 pm

Keithg wrote:What problems with PID PWM were you chasing? Just curious. I have it working well and notice a few oddities, but it is imminently useful in Beta 14 and before.

I just noticed a few things that *could've* affected PWM idle when I was fixing IAC, so I fixed them while I was there. I don't know if anyone noticed any odd behavior, but it might've occasionally caused the valve to go full open or full closed at odd times.

What does the prediction algorithm do? Is there any way to make it sync faster during cranking. I have a hall which has clean output to 0 rpm, and it seems to take too many revolutions before it syncs. Even wasted spark seems to sync slower than MS1/E, or maybe I am imagining that.

KeithG

The prediction algorithm I believe is 1st derivative prediction that should work even with uneven wheels.

As I said before though in the 2.5 thread, faster sync/start for COP users is a 2.5 feature.

Ken

mops · Post by **mops** » Thu Nov 22, 2007 12:50 pm

Awesome timing.

tonight i'll attempt and upgrade to ms2/extra.

wish me luck

6040solder · Post by **6040solder** » Thu Nov 22, 2007 12:56 pm

Good luck :-)

hassmaschine · Post by **hassmaschine** » Thu Nov 22, 2007 1:13 pm

just got home - seems to work great. I did notice some wierd behavior with PWM idle before, where it would seemingly randomly open the valve too far, or close it too much. but that could have just been a settings thing..

FunnyBoY · Post by **FunnyBoY** » Thu Nov 22, 2007 2:35 pm

Really thanks, keep up the good work.

Post by **jsmcortina** » Thu Nov 22, 2007 3:31 pm

I added gauges to display EGT in Megatune if it is connected to either of the two spare ADC inputs with the AD595CQ chip. (like MS1/Extra)

I also put in a bugfix of the half rpm/spark issue that Gintsk reported.

James

Jon k · Post by **Jon k** » Thu Nov 22, 2007 10:00 pm

jsmcortina wrote:I added gauges to display EGT in Megatune if it is connected to either of the two spare ADC inputs with the AD595CQ chip. (like MS1/Extra)

I also put in a bugfix of the half rpm/spark issue that Gintsk reported.

James

That is cool news James

mops · Post by **mops** » Fri Nov 23, 2007 2:45 am

Howdy.

I succesfully managed to load the extra 2b16 code (upgrade from b&g 2.6xx code). and went for a short ride. I was very reluctant to do so beacuse i had a really good tune on the previous code.

observations:

checked timing, all ok.... all though with the revs going up it tends to retard timing.... not sure exacly how much but maybe 1 degree retarded at 3k rpm ? havent tried higher because i just got normal non-adjustable timing light and the only mark on the crank wheel is 0 degrees, so only way i can check timing is set it to fixed timing 0 degrees and rev it, it gets hot really quickly at 0 deg....

cant seem to be able to get injector opening time right. two squirts (sim) always seem to be richer. i tried 0.1ms increments from 0.7 to 1.2.... all though it might be just me. or maybe it has to do with vaporization of fuel....

had to richen up fuel req and adjust some parts of vetable. cruise areas seem to be lean, idle lean while wot areas rich...

as always i'm having trouble getting idle stable at the speed i want.... but that will eventually happen with enough mucking around...dont get me wrong, when it settles, it's rock stable, just never at the speed i want... either too high or too low...

tried eae and no surprise there... i got no clue how to tune it even after reading the guide.... i disabled t for now to get my fuel table right...

and quickly enabled just very crude tpsdot ae, so i can drive, and noticed following problem... all though my tpsdot treschold was 5 (then set to 10 and then set to 20), ae was cutting in and out all the time even at stable throttle... there was no mapdot based ae. i'm quite certain that ae was activeted even though it mapdot and tpsdot were below tresholds... see the log. notice tp and engine, engine is going into ae (17) and pw gets incerased even with no tpsdot.... any smart explanations for that ?

6040solder · Post by **6040solder** » Fri Nov 23, 2007 5:49 am

mops wrote:notice tp and engine, engine is going into ae (17) and pw gets incerased even with no tpsdot.... any smart explanations for that ?

Possibly because the variables aren't sampled at exactly the same time. Also, the code probably uses a value that's there and then by the time the calculation is done (and the pulsewidth changes) the initial value used to compute it isn't the same any more. That would be my guess. IE, you aren't seeing the tps changing, but it is changing and changing back before it is logged.

GintsK · Post by **GintsK** » Fri Nov 23, 2007 7:03 am

Hello!

Thanks for new version, but looks like dual wheel still do not work for me

More info here http://www.msextra.com/viewtopic.php?p=174081#174081

Gints

Post by **muythaibxr** » Fri Nov 23, 2007 7:26 am

mops:

Most of the retuning behavior you're seeing is expected, as is the slight retarding of spark as revs climb. It is expected that you have to retune due to the work we did in the 1.0.x codes to try to make the fuel calculations more accurate, and due to the .1% interpolation on the VE table.

You can correct the slight retarding of spark using the hardware latency feature.

For normal AE, I'm positive that it works, and just tested it on the bench with expected results. Most likely there you're having issues due to a noisy signal, or settings. ms2/extra samples the TPS much faster than the 2.6xx code, and samples the MAP sensor once for every ignition event, so also likely much faster than the 2.6xx code. This means you may see noise when you didn't before.

There may be a problem with decel though, I didn't test that, so I'll be looking at that as I get time this weekend.

Ken

mops · Post by **mops** » Fri Nov 23, 2007 11:40 am

Hi Guys.

Thanks for prompt reply.

Yes, I'm aware i'd need some degree of ve retune to make it work, thats no problem....

I'm told that ae might be related to everything being more sensitive, so i'll just upp the tresholds, incerase tps lag and keep trying...

other thing i noticed... nonlinear baro correction.....
normally (everywhere else in ms), 100% = no correction... 90%= -10% correction and 110%=+10%... that is not the case in that baro correction screen... 100%=+100% correction... i was driving like that for a while and i checked gamma erenich and it was pegged at 195%..... never mind....

keep up the good work ;)

Asteris · Post by **Asteris** » Fri Nov 23, 2007 3:02 pm

Any news on the always powered on IAC (which heats my actuator) ?

mattd860 · Post by **mattd860** » Sat Nov 24, 2007 5:57 am

Now that it's very cold in the morning (20°F this morning) I can really see a problem with the closed loop idle (pid pwm). So far every morning this week when I started the car with a very cold engine, idle shoots to 1700-1900rpms and ignores my idle table which has a max rpm of 1100. It really bothers me because it's not good for an engine that is 20° to immediately idle close to 2000rpms. If I had to live with this I’m sure my poor engine wouldn't last too long.

Throughout the summer and fall, PID settings of 8,3,0 were perfect and maintained a smooth idle. However, this week I raised them to 11,6,0 just to see if that was the problem but this had no effect. I also raised the pid lockout rpm threshold to 100rpms (from 50rpms) but again, this had no effect. Lastly, I put the car in 1st gear, held the brake firmly, and slowly let out the clutch which forced the rpms down to 900 for about 10 seconds to force the PID loop but again, no effect - RPMS when back up to 1700.

Once the car warms up to 130°+, everything is fine again except now that the PID values are higher, the idle valve oscillates. I'm running Beta 15 but didn't read anything that fixes this in the Beta 16 release notes.

Anybody else having this problem too? Ken, any suggestions?

6040solder · Post by **6040solder** » Sat Nov 24, 2007 6:24 am

Nothing to do with these settings is it? :

http://www.msextra.com/ms2extra/General_Pix/PWM_IAC.GIF
http://www.msextra.com/ms2extra/General ... C_Duty.GIF

http://www.msextra.com/ms2extra/MS2-Ext ... l.html#wue

Just a thought.

hassmaschine · Post by **hassmaschine** » Sat Nov 24, 2007 11:51 am

6040solder wrote:Nothing to do with these settings is it? :

http://www.msextra.com/ms2extra/General_Pix/PWM_IAC.GIF
http://www.msextra.com/ms2extra/General ... C_Duty.GIF

http://www.msextra.com/ms2extra/MS2-Ext ... l.html#wue

Just a thought.

no, those settings are only for IAC warmup mode, not PWM closed loop

I had the opposite problem from matt - when I start up and it's freezing, the PWM valve refuses to open far enough and my engine wants to die without throttle input. I haven't done a cold start like this with the new beta yet - but sometimes it seemed to ignore what I had in my target table, and idled at 1400-1500rpm (once the valve finally opened) rather than the 1200rpm I had as the target.

I thought maybe the valve was frozen, but I've had this car for years and never had a problem with it when I had the stock ECU starting on much colder days.

6040solder · Post by **6040solder** » Sat Nov 24, 2007 1:14 pm

ok, so when closed loop is selected here, the pid loop is expected to keep the idle constant as per the usual settings?

http://www.msextra.com/ms2extra/General_Pix/MS1_IAC.GIF

I thought that there would be a base to work from even in closed loop that was temperature dependent. Some engines dont like to idle slowly when cold, and some people would want their engines to idle a bit faster than normal when cold anyway to speed the warmup process.

Are you saying that that is not achievable at the same time as closed loop at the moment?

On the 20 valve 4age, it from factory has closed loop pwm idle and also a warm up pwm ramp over time. The rb20de skyline has the same setup also (2 wire instead of 3 though).