MicroSquirt.com read only forum (Jan.15,2018)

mops wrote:Thanks Al, good info there ;)

moreover, i'm noticing following things... with just standard (mapdot, tpsdot) AE, when I releasse the throttle then ego immidiately reads rich for up to even a second (10.something afr's)

I've spent about a week trying to tune this exact problem... at first I figured that setting the very low VE values to small numbers might fix the problem (ie MAP below 20). But this didn't seem to fix it...

It seems as though there's a different fueling requirement for when you're accelerating through the VE bins than when you're decellerating through them.

Is this where something like "Overrun fuel cut" would help out???

Justin

Pardon the delay,

Here are three little chunks pasted into one log.

TPSacc represents the added fuel pulse in mSec. after subtracting 100.
(if less than 100 is %decel correction)

The AE is pulse (rpm) dependent.

Instead of going into much discription, I will just post the log.
Ask if there is a particular area or detail you want to know about.

[I just modified exhaust plumbing. Replaced 1.125 DIA restriction stock muffler with a 2.25 DIA homebrew. Now overall fuel needs to be tweeked a bit. Idle area seems most effected.
]

ok, ok, ok... hang on there tiger....

you are saying that tpsacc is actually in miliseconds of pw added ? this is great news !

you you are saying that (tpsacc-100)/10 = added pulsewidth in miliseconds? (negative value = decel)
that can be incorporated as calculated field into logviewer and be displayed directly.. actually just done it... directly to logviewer config file:

APPEND_FIELD_AE\ PW=([TPSacc]-100)/10

I'll try tunbing with that today....

[edit]
the datalog is cool. i can see that you are pretty much maintaining afr's within min/max range without 'clipping' my afr's are all over the show while freereeving the engine.... oh well... i'll give it a shot tonight...

{were did the datalog I posted go? are .xls files removed?}

you are saying that tpsacc is actually in miliseconds of pw added ? this is great news !

This is true for MY datalog and if memory is right is default behavoir on MSII. MSI_extra is another subject. I saw all sorts of odd numbers from different versions of code...

that can be incorporated as calculated field into logviewer and be displayed directly.. actually just done it... directly to logviewer config file:

For the attached screen shot I used this function in Excel (for PW-AE collumn):

=IF(O3>100,Q3-(O3-100)/10, Q3*100/O3)

where O3 is the TPSacc collumn and Q3 is the injected PW collumn... I realise those details are useless without the .xls file. But like you discovered, better yet is to do it directly in the megasquirt logging section.

So in the log image, the yellow PW line is the actual injected fuel pulse were as the blue PW-AE line is the "base" fuel pulse from the VE table, req_fuel and static corrections.

I marked a section I found interesting with the red label. I liked how it showed the lag time between a fuel event and the EGO sensor. (The delay here is about 0.3 seconds.)
It also seems that in this spot I might be able to save a bit of fuel by reducing AE specificaly in the area of:
MAP ~24 kPa
RPM ~2400
TPSdot ~5
(I can do that with my code
The challenge as you probably have found, is getting good responce from a gentle throttle movement off of idle (like starting from a stop light) were quite a bit of AE fuel is needed relitave to TPSdot but not giving too much AE while making the same gas pedal movement during cruse.

[grrrrr...
The Attachment/Image must be less than 1200 pixels wide and 900 pixels high
]

A few more notes on the log file for the curious...

... AfrTarget ... CyclsTX FuelFlow

AfrTarget is handy to log so you can see how close the tune is to the AFRTarget table setting. I really should divide by 100 so the scale is normal.

CyclsTX is how many engine cycles happen between log entries.
It is usually 1 or 2 but in extreem driving (were you need the details the most drops to 5-7. I use it for code diagnostics. When I see big glitches in the log, this delay might be part of the problem.

Also an experimental FuelFlow... still needs work on the scale factor and a "FuelUsed" would be nice to help with this calibration.
When i have the time and a helper I hope this will alow me to hold a cruse speed on a flat road and adjust timing and AFR for the best ecconomy. Lowest FuelFlow for a given power/road speed not neccisarily a given MAP.

AE is a pain. I've been playing with this forever. But progress is progress. I don't know what it is with these 4 cylinders.....

A.J.

bluetrepidation wrote:AE is a pain. I've been playing with this forever. But progress is progress. I don't know what it is with these 4 cylinders.....

A.J.

try it on a V8 ;) ....

just a short update.

found whole heap pf rust on my e30, so fixing that atm.

i think fuel flow is not neccesery factor.

at given speed you have given gear and stable rpm. so they only thing you can tune for is the shortest possible PW and that will give you best economy, doesnt matter what map. I found my car need an amount of fuel to maintain cruise whether it is at 40kpa or 60kpa... if i try to lean it much beyond 16.2 afr then engine becomes really weak.


re accelenrich.... i got the tps one tuned two nights ago. I basicallyed used live log function od logviewer. on the highway, 45 kpa cruise at low map... - give it a stab... watch the log... noticed it going lean... incerased a bit... repeat few times and to got to some pretty good values. it is interesting that my slider sits now at 38%, and at lower than 35% and higher than 41% i get bad behaviour, like missed (lean or rich)... it appears to be a very delicate setting. so it's all tuned but the extreme throttle movements, but i'll get that sorted.

what else... re off-idle response.. the way i deal with that is that i richen up the bins abouve idle bins and that seems to help alot.... i know it's not a perfect solution, but hey... it works ;)

So mainly you guys are using the stock settings but using the slider to scale the AE pw response. Now for the areas where you still get AFR clipping X Tau comes into play.

Am I correct on this?

A.J.

Jedrik wrote: TPSacc represents the added fuel pulse in mSec. after subtracting 100.
(if less than 100 is %decel correction)

]

Maybe its different for MS-I v3 board running 29q2? My TPSacc is very strange - when AE kicks in (engine code = 17), I get a negative number.

And I sometimes get AE for no reason at all - TPS didn't change, neither did RPM/CLT/MAP. Not sure what is going on. Anyone experienced these?