1991 Turbo 306 coupe MS3/PiMPxs tuning adventure

[a91what]

NORMAL
I will see if I can dig up my dyno sheet with that cam, peak torque is at 2500 (at least with my combo it was). This is what i was referring to in my previous post about a slight peak in the table.

 

[wiseguyk]

Ah ok I think I understand now. So it needs more fuel in the parts of the table where it makes peak power, that makes sense.

Awesome, I'd love to see that dyno sheet if you can find it to see what it looked like when you were running the E303.

Is it best for me to not adjust the cells that VE Analyze changed? And just make sure there aren't big jumps between changed cells and the unchanged cells next to them?

Sorry for all these questions about VE table values and smoothing, it's been really helpful though to ask these questions as I'm going through the VE table refinement phase of my tune.

 

[a91what]

 

[wiseguyk]

I was just going over the 2 datalogs I collected from today and have a couple questions. I took one shorter log on the way to work and then a bit longer one on my way home.

I ran the logs through the analyzer when I got home with CLT > 186* (after checking several recent logs and seeing it never drop < 185* once warmed up) and minimum RPM of 1200. This is what it recommended:

I"m curious about the low MAP area I outlined in red. I've noticed that VEA keeps wanting to add fuel in that area and I've been letting it, but when I check points like that in the logs that condition only seems to happen when downshifting or coasting while in gear.

Here's an example:

It seems erroneous to add fuel in those cells since in these scenarios I'm not on the throttle at all. Should I be excluding these low MAP cells from VE Analyze, or set them back to their original lower values?

One other question I have is that I've noticed some surging idle once fully warmed up with the fan on. Is it possible to add fan control on/off to the data log? It's tough to remember when the fan kicks on and off, though I guess if it can't be logged I could just try to watch the temp and press spacebar to mark the log when it happens...

 

[a91what]

1. set the minimum KPA to 30 thats the overrun portion of the fuel map
2. we need to identify why the idle is surging with the fan on, i will look in the datalog and see if something is amiss.
3. What output pin is the fan associated with?? in the MS3 manual there should be a table that shows the bit value of each output and what status it is tied to.

for example in MS2 extra Status 1 with a bit value of 64 is spark table switch
PORT AM with a value of 16 = ALED output active

 

[wiseguyk]

1. set the minimum KPA to 30 thats the overrun portion of the fuel map
2. we need to identify why the idle is surging with the fan on, i will look in the datalog and see if something is amiss.
3. What output pin is the fan associated with?? in the MS3 manual there should be a table that shows the bit value of each output and what status it is tied to.

for example in MS2 extra Status 1 with a bit value of 64 is spark table switch
PORT AM with a value of 16 = ALED output active

Thanks Steve, you're the man! The output I'm using for the fan is D15 (port PM5 if that helps) :

I went through all the outputs available in MLVHD and saw those "status" parameters and wondered what they're used for so it sounds like I just need to find out which status number and value MS3 uses for D15 - ON. Is that correct?

Thanks a lot for taking a look, I was going to ask you about over-run fuel cut in the last post since I thought it was related to the areas of the table hit when coasting or decelerating in gear. I believe we disabled it a while back and was curious if we would enable it again at some point.

I'll make the minimum MAP change to my VE Analyze settings. Do you think there are any other tweaks I should make now that my VE table is getting more dialed in?

 

[a91what]

http://www.msextra.com/doc/pdf/html...quirt3_TunerStudio_MS_Lite_Reference-1.5.html


Section. 4.6.1
Scroll down to portxx and start reading. Your output is watched by.
Portmj when you get bit 32 =D15 is active
Bits stack so if you see portmj=48 that means d15 and d16 are active.
It's all based on bitwise operation. This is also explained in the manual.

I hope this helps

 

[wiseguyk]

Thanks man, that's exactly what I needed! At every turn I continue to be impressed by all of the kick-ass and well designed features the Megasquirt platform has to offer. It really is amazing the amount of granularity and control it gives you!

 

[wiseguyk]

So my current approach has been to data log 2-4 drives each day, and then after getting home that night run the logs through VE Analyzer and apply those changes. Rinse and repeat the next day.

Today when I analyzed a few logs from yesterday and this morning, I noticed a pretty big adjustment VEA suggested which caused the value between two cells vertically to be pretty wide, like 20+ point difference:

@a91what Steve, what do you make of this and is it something to worry about smoothing or tweaking?

It's interesting to me because it lowered VE in the cells below, and raised it in the cells above. It raised the blue cells by ~10 and lowered the red cells by ~2.

I'm going to smooth the values above them it that weren't touched by VEA since now they're jagged but curious if you think it's OK to have a jump of over 20 points or if there's something else I should do. Thanks man!


Edit:
A visual showing my concern

 

[a91what]

looks like transient fuel to me.... can you find this spot in the datalog so we can look at it?

 

[wiseguyk]

looks like transient fuel to me.... can you find this spot in the datalog so we can look at it?

I tried to find some way to do a combination search on the log so I could search for RPM and MAP together but can't seem to find a way to do that and find those spots easily Is that a feature that I'm just missing?

I've attached the log files and tune file the VE Analyze output was generated from in case you get a chance to take a look. I'm not running the updated version of the tune since I'd like to get this anomaly figured out first. Thanks!

 

[wiseguyk]

@a91what Steve, I've just been continuing to datalog each of my drives using the same rev 11.1.5 tune and plan on loading them all and then running VEA to give it a nice large data set to use.

Is there a way to search for multiple parameters in the log(s)? Such that I could put in an RPM and MAP threshold to find these possible "transient fuel" areas that you mentioned? Or if there's another way to find these spots other than manually scrolling through to find them that would be excellent. Thanks man!

 

[a91what]

Not that I can think of off hand, now there is the chance that fuel is truly needed there but it is slim... I'm thinking is more ae related. I will have to scroll and look for this spot.
Now I like to use the histogram and scatterplots as well, the histogram would be very useful here. Once you have it built filter the transient's and see what happens.

 

[wiseguyk]

Thanks dude! So I set up the histogram and scatter plot to have RPM vs MAP and set up the ranges to match my VE table values. This helped in a HUGE way to see which areas most of the drive time is spent

I realized I could click and drag to quickly scan through and this was the first spot I found which seems to verify your thought about it being AE related:

Is there a way to filter out AE events from VE Analyze or do I just have to manually reset those cells after it recommends a new table?

Here are shots of the histogram before and after enabling the transient filter:

With the transient filter enabled, I see the hit counts definitely drop in the suspect VE cells. What would you then do with this info?

 

[a91what]

I can't give good advice from that log, it looks like you blipped the throttle to downshift. Not sure, my gut says it is simply a transient and should be ignored.

 

[wiseguyk]

@a91what Here are my latest logs from my most recent (11.1.6) tune. I haven't tweaked the fuel or timing in the low MAP ( < 40) range between 1600-2400 where I mentioned it's been bucking.

If you get a chance to take a look and suggest any VE/Spark changes in those areas along with anything else we might want to tweak at this point in the tune that would be AWESOME Thanks so much man!

 

[a91what]

Try this.

 

[wiseguyk]

Try this.

Thanks man, I see you bumped the timing some like we chatted about earlier

Could you please take a look at my VE table too?

I've been unsure of how to adjust that same range where the bucking occurs and was able to find in the logs when it was happening that the AFR was swinging lean. I've been excluding that range from VE analyze but it does seem to need more fuel in there (comparing actual vs target AFR) I'm just not sure how much and curious to see how you would adjust and smooth it...

 

[a91what]

What i did was adjust the fuel rows, this will make each step cleaner. I also turned on overrun fuel cut, this will save fuel. keep this in mind as you tune.

 

[wiseguyk]

When I compare to the previous tune, I see almost across the whole table fuel was pulled out. Since 11.1.7 increased timing I'd expect we'd pull some fuel out of those areas of the table but the diff shows it's pulled out of pretty much the whole table.

Is that a result of purely adjusting the fuel rows like you said, or did you use VE Analyze or anything else for these changes?

I don't want to just apply your changes without understanding them or I'll never learn!