1991 Turbo 306 coupe MS3/PiMPxs tuning adventure

[a91what]

When you are tuning this setting, there are two fields you must make note of.
1. TPSdot
2. AE (this is based as a simple PW [pulse width] adder)

Add both of these to your dashboard so you can see the tpsdot when modulating the throttle at different rates. I have not had time to look at the datalogs but I will say that the actual AFR number you reach when you need AE is not important.. give the car what feels best not the best numerical value that conventional wisdom would imply. To high of a value at low tpsdot rates (what kurtis is referring to) can cause stumbling issues and waste fuel.

 

[wiseguyk]

Man, it feels like forever since I've posted any updates!

On Saturday I finished up the Mark VIII fan mounting and wiring and installed the 180* thermostat. This made a HUGE improvement on temperature, going from an average of 215-220* in peak FL heat down to 190-200*. I"m using the fan control functionality of the MS3 to trigger the fan on at 198* and off at 190* and this has been working very well so far. I need to play with the fan control idle-up settings some more though because I do experience a pretty big RPM drop when the fan kicks on while idling. If there are any tips you guys have to help with this, I'm all ears

After the fan and thermostat were done, I drove the car a couple times and still noticed some hesitation in that 4k range. Since I'd swapped the MSD distributor and TFI out already but not the coil, I decided to try swapping the coil for the one Steve let me borrow just to see if there was any difference. The next drive when I punched it WOT from 2k rpm it just felt much stronger! There was no hesitation that I could feel, and the car just felt like it was pulling harder and smoother than before. I'm kicking myself for not swapping the coil already but the one I had was a brand new Screamin Demon coil so I just assumed it would be fine. Not so! lol

After that great news, I decided to drive the car back home! Woohooo!!!! She's been at the shop for going on 8 months at this point and can barely believe she's finally back in my garage!

Thankfully the rain held off during the day today so I drove the car to work and data logged 3 trips Since I believe the hesitation issue is finally vanquished, I loaded the previous tune before Steve pulled some timing and added fuel.

I loaded the multiple logs from today's drives into MLVHD tonight and did some VE Analyze action (Hard, lambda 3). I knew it was running kind of rich most of the time and it pulled some fuel out of the cells like I expected. I then went through and tried my best to smooth out the edges. Still not sure the best approach to doing this but I think it's ok...

I haven't made any VE changes in quite a while so I wanted to post the newest one from tonight in case one of you awesome fellas could spot check it for me. I plan to drive it again tomorrow and continue fine tuning. Please let me know if there are any things you think I should be looking into that I might be unaware of. Thanks!

 

[FastDriver]

Ha! It was ignition after all! Yesss!!! Anyways, glad/hope this has put that issue to bed. Again, consider downing the gap to ~.028".

 

[wiseguyk]

LOL yes, I hope it's been vanquished finally!

Thanks, next time I pull or swap the plugs I'll reduce the gap a little

 

[a91what]

Ok so under fan control you want to play with these two settings for now..
Idle up delay
idle up duty %
you want to set the delay long enough for the engine to get ready for the fan to kick on 100-200ms should be plenty long enough.
the idle up duty % will need to be increased so it counters the drop in rpm, try a setting of 10% first to get yourself started.

 

[wiseguyk]

Thanks @a91what for the fan control suggestions. I tweaked my settings for delay of 200ms and duty of 20% since at 100ms and 10% it didn't seem to try to catch the RPMs at all. This seemed to help keep the RPMs up a little higher than it dipped previously but I think it could still be improved. I have the idle up target increase set to 100rpm, do you think that needs tweaking as well?

Current fan control settings:

So this week was pretty good for driving the car to/from work and datalogging. I've attached the most recent logs zipped up and the current tune used for those logs.

I had more logs than these from earlier in the week and ran VE Analyze against them, loading multiple logs into MLVHD. Where I'm a little stuck/nervous though is applying the VE Analyze changes and smoothing the tables. I did this for the earlier logs and when I drove the car I was paranoid that it might have pulled more fuel than was safe since I did notice the AFR numbers leaner than before. So I reverted back to the previous tune and that's what the attached logs are from.

So with that said I'm hoping that I can get some advice on how to properly do the fine tuning phase of driving, datalogging, analyzing, applying changes, smoothing, repeat.

To maybe help people following along, here are some screenshots of the before/after VE Analyze for the logs and tune attached:

Before VE Analyze:

After VE Analyze:

As you can see, it's almost exclusively was pulling fuel out. What I'm curious about is if I should be careful about the changes it suggests under boost, and if I should be using a min-max MAP filter like I have set for the idle/RPM filter.

Additionally, is there a better way of smoothing out the VE table after accepting the VE Analyze changes? With the earlier logs I had been manually adjusting the neighboring cells to make smoother, but I also came across an excel spreadsheet someone made over in another forum that allows for automatic smoothing along with smoothing factor. I can attach it here or post the link if anyone is interested unless you think it's a bad idea to use a tool like that.

Anyway, thanks as always and I very much appreciate any help or advice you can give me on this part of my tuning journey now that the car is driveable again

 

[a91what]

The culprit!

Check your minimum clt temp....

Increase it to 190 and see what happens.

 

[wiseguyk]

Ah shiz, I didn't notice it was set to 160!

Here's the result with it set to 190*:

Oh and I should have mentioned that one of these logs were on a drive home after my Mark VIII fan decided to seize up so my temps got > 230*

Quickly comparing the 2 VE Analyze results I do see the fuel pulled is slightly less with the 190* filter set so great catch Steve!

Looking at the results of the analyzing, what are your thoughts about allowing it to make changes under boost and the proper smoothing technique to use after accepting the VE Analyze changes? Thanks dude!

 

[a91what]

the VE analyzer is correct.

Its actually not going to pull quite enough but it is averaging in multiple pulls. My math says it is accurate, this is normal after moving the IAT sensor.

to blend the edges just pull a line at a time from what has been changed and interpolate, most of the map that is unchanged by the analyzer the car will never be in, so dont worry to much about them just put in numbers that make sense based on what the values are in the used portion of the table.

looks like the idle is giving you issues once the fan turns on huh... dont increase the RPM at all leave that at 0
increase the amount of time for it to get ready to 300ms and up the adder to 30 and see what happens... you want just enough for it to catch and not idle up to far.

 

[wiseguyk]

Its actually not going to pull quite enough but it is averaging in multiple pulls. My math says it is accurate, this is normal after moving the IAT sensor.

Ok great, thanks for double checking it for me. I'm planning on continuing to data log whenever I drive this week so I'll run those logs through VEA and see if it pulls out a little more.

to blend the edges just pull a line at a time from what has been changed and interpolate, most of the map that is unchanged by the analyzer the car will never be in, so dont worry to much about them just put in numbers that make sense based on what the values are in the used portion of the table.

When you say to pull a line do you mean like this and then interpolate vertical?

Just want to confirm that's the right technique. Then I'll take a stab at it and post my results.

looks like the idle is giving you issues once the fan turns on huh... dont increase the RPM at all leave that at 0
increase the amount of time for it to get ready to 300ms and up the adder to 30 and see what happens... you want just enough for it to catch and not idle up to far.

Yessir, it's still dipping quite a bit when the fan turns on. It is better than it was after first installing the fan though. I was previously just using a programmable output to trigger the fan but then realized I could be using the actual "fan control" functionality with the additional settings. Duh! lol Anyway, it used to almost die when it would kick on so I know we're on the right track I've made those changes and will see if it helps.

 

[a91what]

Ok great, thanks for double checking it for me. I'm planning on continuing to data log whenever I drive this week so I'll run those logs through VEA and see if it pulls out a little more.

When you say to pull a line do you mean like this and then interpolate vertical?

Just want to confirm that's the right technique. Then I'll take a stab at it and post my results.

Yessir, it's still dipping quite a bit when the fan turns on. It is better than it was after first installing the fan though. I was previously just using a programmable output to trigger the fan but then realized I could be using the actual "fan control" functionality with the additional settings. Duh! lol Anyway, it used to almost die when it would kick on so I know we're on the right track I've made those changes and will see if it helps.

Yes you can just pull verticle and horizontal lines from the affected blocks and interpolate them, always make sure the numbers make some kind of sense based on the value in the adjacent cells

 

[stanglx2002]

Easiest thing to remember is the higher the rpm at that same Map reading the larger the VE number (ie: the more air the engine intakes). This is the same as Map increases as well. This should help you smooth out your VE table.

Example at your 3500-4500 rpm at 140 map are lower than the neighboring cells. These need to be smoothed.

 

[wiseguyk]

Yes you can just pull verticle and horizontal lines from the affected blocks and interpolate them, always make sure the numbers make some kind of sense based on the value in the adjacent cells

Awesome thanks! Here's my attempt at smoothing using the horizontal and vertical interpolation:

I also tried using that excel spreadsheet I mentioned as well and did a version smoothing out 1100 RPM onward. I'm curious what your thoughts are on it since it certainly seems to do a good job at smoothing everything, like fixing spots where there's a single larger value between two smaller ones.

Am I worrying too much about table smoothness? lol

 

[wiseguyk]

Example at your 3500-4500 rpm at 140 map are lower than the neighboring cells. These need to be smoothed.

Ah yes, and I missed that in my manual table smoothing...

I would imagine after I do a bunch more data logging and repeat the VE Analyze and smoothing steps some more, eventually I'll hit a point where VE Analyze won't try to change much since my values will be adjusted enough that I'm close to target?

Obviously VEA will only try to adjust for areas I actually hit in the table, but the "normal" areas should be pretty well tuned after doing this a bunch?

 

[a91what]

In short yes.

now i will say this about that.... you will not always see higher ve numbers in the higher rpm ranges... I see this most often in cammed engines, they will need less fuel [ slightly nothing crazy ] after they cross peak torque. in your case [turbo] this should not happen. but you will see a slight peak in the cruise range somewhere around 2500 with that cam where fuel will be lower on each side of the table in the transitions.

that excel table was far to aggressive with the smoothing = do not suggest

 

[wiseguyk]

that excel table was far to aggressive with the smoothing = do not suggest

Ok good to know, thanks Steve! I shall smooth by hand then

 

[wiseguyk]

I did some more work, addressing areas on the right side of the table where there would be a few values out of the normal slope:

@a91what and @stanglx2002 What do you guys think?

 

[a91what]

Fix the strangeness at 1100rpm column

 

[wiseguyk]

Fix the strangeness at 1100rpm column

Ok, how about this? Definitely feel like I'm getting the hang of it. I think... lol

 

[wiseguyk]

I did 4 drives today running my smoothed VE table shown in my last post and wanted to get @a91what your opinion on whether or not you think the VE Analyzer changes it recommended today need smoothing or are good as is.

In particular, there are a few spots where the values have been adjusted such that there will be a larger value in between two smaller values or vice versa. Is it OK to have this in certain places or should the values always be increasing when moving up or to the left in the table?

Here's an example:

Those cells are larger than their surrounding values so looking for guidance on if that's OK or if I need to smooth it out at all. Thanks!