3SGTE Gen3 MR2 Weird Ignition map

[JesterTiger]

Got a customer with a G4X that was recently tuned by a garage.
been hunting a funky idle issue. 
have gone through all the basics, vac leak, boost hose leak, IACV issue. 
even replaced the IACV with a known working one. (off my own Gen3 turbo)
but then went into looking at the tune done. 
I'm a bit curious about the Ignition map. i use a MS3X on my own Gen3. (i know Megasquirt.. but )
the Ignition map on that starts at 0 KPa and goes up from there.  simple enough stuff. 
the one that's on the G4X tune starts at -100 KPa and goes up from there. the last two rows are the only ones in boost ie above 100 KPa.  and the variations of ignition timing at the lower levels, its like no wonder its hunting. the engine is constantly changing the timing at idle or into negative vac. 
Hoping someone on here can provide a reason why its been tuned that way and if it could very well be the root cause of the weird idle because of this map. 

steven-mr2-turbo-gen3.pclx

[Adamw]

Your ignition table is set up with MGP on its load axis (Manifold Gauge Pressure), so -100kpa is total vacuum, 0Kpa is atmoshere, any positive value is boost.  

If your megasquirt table only started at zero kpa then that would likely be MAP rather than MGP.  For ignition MGP Vs MAP on the load axis wont make a lot of difference to most users that dont drive over large altitude changes.  MAP would be the more correct load reference for ignition as that is what affects ultimate cylinder pressure (knock).  Some users (mostly americans) prefer to use MGP just so -ve numbers is vacuum and +ve is boost.

MGP is good to use on the fuel table load axis as you get a more correct VE look up with altitude or baro changes.   

So the ignition table looks ok with a flat 15deg around the idle area.  Be aware however when idle conditions are met the ignition timing will be controlled by the idle ignition function anyway - ignition table will not be in use.  This is set up a bit slow in you map but I dont think it would cause hunting.  Most other settings dont look too bad.  

It would be best to do a PC Log showing the issue, then we may be able to offer some direction.  

 

 

[JesterTiger]

Thanks Adamw. 
so that's the difference between MAP and MGP. gotcha. need to do more reading up on that then. 
i have attached a log of Stevens tune. just need to see what others think. Steven in convinced there is an issue with his hunting and idle. if you watch the log you can see its not stable, but how stable its supposed to be im not sure. not on that Link anyway. just off to sort a G4+ on a supra thats lost its crank/rpm count. (wiring loom took a hit in the RH wheel arch and chomped through a load of the wiring. will prob run new crank and cam sensor shielded wiring to the ecu.)

PC Datalog - 2022-09-17 12;57;55 pm-steven.llgx

[Adamw]

The idle looks pretty smooth to me for most of that log.  I dont see any hunting so I assume the problem just wasnt happening when this log was taken?

I do have a couple of tune suggestions though that will improve idle control. 

The first one is the idle base position table is not even close.  In the pic below where my yellow cursor is you can see RPM is sitting right on target, with the idle position at 31.9%.  Yet there is a value of 47% in the base position table.  So you can see just to the right of the cursor when the throttle is blipped, the idle position jumps to 47% (the commanded base position), this makes the RPM settle above target on the way back down and it takes a solid 30 seconds to slowly come back down to target.   So two changes needed here: 1. base position should be 32% for this temp, likely the whole table is wrong, not just this one cell.  2. You need a lot more integral so that the closed loop works faster to bring the RPM to target as quick as possible.    I dont have much experience with the MR2, but your integral gain is currently set to 0.1, I suspect you will be able to increase that to at least 0.4, possibly a lot more.  

 

Secondly, the idle ign is not doing much, the proportional is too low.  It wont be fast enough to catch sudden load changes or stall events like this.  Your proportional gain is currently 0.1, 1.0 is more typical.   I would also increase idle ignition max clamp to 30deg as that will be much closer to MBT than the current 20.  

The other thing that looks odd in the log is after the throttle blips the lambda pegs at max lean for quite some time before it recovers.  After the second blip it takes ~15seconds for the lambda to come back down to target.  Yet the RPM, MAP & injector PW is stable the whole time (ie air flow and fuel fuel should be stable), so something is fishy there.  It could be a wideband issue, fuel pressure issue, air leak?   Since it still appears to be idling ok in that area I suspect maybe the wideband is most likely since it would unlikely keep running if the lambda really was 1.25

[JesterTiger]

Cheers Adam, will have a look at those the next time the MR2 is in. Need to make another post as i have an interesting 2JZGTE one for you to have a look at.

[koracing]

You should log the car from cold start to fully warm and see what idle duty cycle is actually required to hit idle target as it warms up.  I like to set the base position just a few percent (3-5%) over the normal required amount as it seems to make a slightly smoother return to idle in most cases without dipping below target.   This makes it function similarly to the dashpot offset, but I think it's a little smoother operating this way.  It also will nominally make it idle just a touch above target while driving until you get below the speed lockout which I also prefer personally.   

Here's the settings in the G4X I use on most relatively stock MR2 3SGTEs.  Some settings will obviously vary for the individual motor, but should be a decent starting point: 

[JesterTiger]

On 10/12/2022 at 5:26 PM, koracing said:

You should log the car from cold start to fully warm and see what idle duty cycle is actually required to hit idle target as it warms up.  I like to set the base position just a few percent (3-5%) over the normal required amount as it seems to make a slightly smoother return to idle in most cases without dipping below target.   This makes it function similarly to the dashpot offset, but I think it's a little smoother operating this way.  It also will nominally make it idle just a touch above target while driving until you get below the speed lockout which I also prefer personally.   

Here's the settings in the G4X I use on most relatively stock MR2 3SGTEs.  Some settings will obviously vary for the individual motor, but should be a decent starting point: 

I'll look into that. 
Currently dealing with a few things with it. 
Found the AEM Wideband sensor was placed badly. who ever fitted it didnt read the instructions that came with it. Its got a Berk Downpipe. so they cut the body of the original Narrow band sensor off, welded the Boss for the wideband onto the 2 bolt plate. then fitted it into the stop top position. the wideband isn't a long body sensor.  that with the plate, the gasket, the stand off that the stock one had, meant the poor sensor wasn't anywhere near being in the flow. although that prob saved it from being overheated being a few inch's from the turbo. my own one is at the bottom of the downpipe and ever that is really a bit too close. 
in process of welding in a new bung at the lower end of the downpipe at the correct angle. 
it was causing lean conditions. watched it on the laptop after AdamW mentioned about things going lean and not recovering quickly. also the owner doesn't have a AEM Gauge to show what's going on. just a simple Autogauge Mixture gauge, which does nothing but peg lean with a warning light.  i turned it off cause its useless. 
however everything im doing with the MR2 Turbo right now might be a mute point. its in for a Pre MOT inspection and i have found so much rot on the rear arches that a slight dodgy idle is the least of his issues. 
(the idle btw happened after it was tuned by the rolling road. im just trying to sort things after the fact. and dealing with the fallout.)

[koracing]

Lots of tuners are good at dialing in the rolling road/dyno tuning but not so good at making a good stable and repeatable idle.  End of the downpipe is very commonly used in my installation and doesn't seem to cause pre-mature sensor failure.  I also often add a relay to power the wideband, however, controlled by an ecu output so you can wait for engine to start before the heater turns on to limit thermal shock.