[daemon]

Originally Posted by WolfPlayer

I know that Daemon on this board ended up giving the higher deltas more fuel too.

FWIW, I'm not using that mapping anymore. It appeared to improve some things, but I was never 100% happy with it. Then I ran out of time to play (especially with lots of base map changes). I'm eagerly anticipating your new values.

[WolfPlayer]

Is it me or has this board picked up some speed lately? Nice

[ride5000]

Originally Posted by RainMaker

The problem also is that as far as I can tell there is the single dimension of TP delta... but I think there is a temperature based dimension out there too, since my car definitely has tip-in issues for the first 3-5 minutes of driving that are absent after that.

there is definitely a 3-d tps/ect map on my ecu image. i had to use xmlwrite to pick it up off my base a4tc101l rom (my03 usdm wrx).

incidentally, it is NOT zeroed out, as are some of the other tps-based maps (ie tps/rpm). i have not altered it yet.

ken

[RainMaker]

Originally Posted by Rainmaker

What you say appears to be totally true.

Please consider my Tip-In values "Under Construction"

Ive added another 20 to the lower 5 values and then blended that in.

It hasnt made a huge difference. I'll add some more, but with these JECS, its not making the difference it used to when you added a mere 10%.

On another note, I have had a couple cases where on tip-in it went lean, and logging didnt pick up any tip-in enrichment at all. Out of the blue, some moments dont actually trigger tip-in, which makes the car actually stumble for a change (according to both my feel and logging). Very odd.

chris

[RainMaker]

Originally Posted by WolfPlayer

(not that the car felt any different - but I felt better because the logs looked better - roflmao).

Boy is that the truth. Sometimes I wonder if the LC-1/XD-1 guage is a curse. The car probably runs quite well, but I cant stand the guage swinging around and I probably "feel" what it is saying even when its not noticible.

The curse of the over-informed.

[ride5000]

Originally Posted by ride5000

there is definitely a 3-d tps/ect map on my ecu image. i had to use xmlwrite to pick it up off my base a4tc101l rom (my03 usdm wrx).

incidentally, it is NOT zeroed out, as are some of the other tps-based maps (ie tps/rpm). i have not altered it yet.

ken

for example, my untouched coolant temp comp map:

deg C	% cor.
-40	3.21875
-30	2.96875
-20	2.71875
-10	2.5
0	2.25
10	2
20	1.78125
30	1.59375
40	1.40625
50	1.25
60	1.15625
70	1.0625
80	1
90	1
100	1
110	1

this is called "throttle delta fuel comp. (coolant temp)"

[RainMaker]

My current values (that still arent enough in the lower 5 cells):

141-156-169-256-333-417-489-560-623-703-778-859-939-1008...

[RainMaker]

I just went another 15%. Hopefully we'll have something a little richer this time. I'd be ok if I went too rich, because I'd know I could come back a few percent.

Chris

[LOLSTi]

This thread rules, I haven't even started playing with my ST, but the discourse is so damn informative .

[RainMaker]

So tuning a blow-through is a little strange, in terms of Tip-in. Im getting situations where I perform the same operation twice and once the car registers tip-in and fuels accordingly, another time it somehow doesnt register the tip-in and bucks lean.

Once I get the A:F ironed out, which I think will require some different column numbers (the low end is where all my issues are... .12-1.0 TP Delta), I'll try and get further into logging why certain situations log differently.

Also, tip-out seems to be effected by the tip-in numbers. My fuel is going rich on tip-out now that I have raised tip-in a bit.

You can raise the lower end of tip-in, but eventually the idle and the the RPMs settling into idle begins being effected. Since the idle is a function of our DBW, the ECU seems to register a little tip-in when the throttle controls the idle :-/
Again, I need to rename some of the smaller columns to increase the resolution down at the low end.

[WolfPlayer]

Well, I see a lot of weird stuff too. Sometimes I see one thing. Other times I see another. Furthermore, I haven't quite figured out how exactly to tune tip in. I log delta TPS, TPS, STFT, AFR, and vehicle speed at 10/s. Problem is, the delta TPS just makes no freaking sense to me. I usually always insert a new column and create my own delta TPS column. This doesn't even come close to being the ECU delta TPS (even though it is more accurate for delta TPS). I can open the throttle plates 10% and delta TPS will read like 0.85. WTF? I've been delaying talking about this (even though I have been aware of it forever). I guess I was hoping to figure it out. However, I can't. Any help?

After fracking (BSG reference) with this Tip-in for several base maps, I am actually back to my 'rich then lean' scenario. And NO, this is not a function of the STFTs (Air/Fuel learning immediate). If I reduce Tip-in then it's pure lean. In the current form I get slightly rich then lean (and sometimes all lean with no richness) <-- And this is with a tip-in map with substantially more tip-in. The good thing is that this map feels better than the leaner tip-in map.

If anyone knows how this table works in the context of determining a delta TPS and/or using the ST delta TPS value, please post.

t

[RainMaker]

One thing that I noticed is that with the recent adjustments to Tip-in, it seems to work about half the time.

If I stare at the logs long enough, its appears that the half of the time that tip-in works is when the STFT is in "my favor"... and when it isnt, the car goes lean.

Perhaps thats a sign I havent "gotten there" yet, and perhaps its a sign that I'm tuning tip-in too close since PCM reset... Im not sure. But it sure makes things difficult.

I understand that tip-in can be a bit of a "impulse" to the STFT feedback system, so I guess I'll go and continue adjusting tip-in until the "impulse" doesn't upset things as much.

Still, what Wolf said above is very true. I cannot for the life of me figure out why certain motions make a TPS Delta one time but not the next.

Chris

[ride5000]

is this perhaps a function of the DBW system?

[WolfPlayer]

An update on tuning tip-in.

This table has 18 cells that start at 0 and go up to 31.x. This table was originally confusing to me because no matter what I did, I never would see anything above like 6% TPS delta in the logs ... definitely no double digit numbers. Most of the time, no matter what you are doing, the car is between 0 and 1% TPS Delta (even in some cases where you crack the plates open by like 40%). After several base map flashes and some realtime tuning/tracing, it is sorta clear. Basically, a rescaling is in order. So much time is spent between 0 and 1% (the first 2 cells) that there is just no resolution (RainMaker already talked about this). I've been slowly focusing my reflashes on just the first 3 cells of the table and have had good results with doing that. It's a very slow process because there is no resolution. However, after enough flashes you will realize it is getting better and better (in the logs only, the car feels good as long as it isn't an initial lean spike).

The past few reflashes were done by logging A/F learning immediate, delta TPS, Wideband AFR, and vehicle speed. I log lots of throttle transitions. Then I sit down in Excel and begin highlighting each transition in yellow. I then put a description next to each section (i.e. lean, rich then lean, rich, etc). What I found is that I could correlate the reported delta TPS with one of my descriptions most of the time. Sometimes things just didn't make sense (as reported by RainMaker and me previously). However, I could make some correlations and tune the first 2 values better.

Overall, the table definitely needs to be rescaled. That is my next plan of attack.

t

[sti752]

I have spent some time tuning tip in lately as well. I was using the stock tip in for the JECs (with a good bit of enrichment added in the bottom three cells) and I would get the rich, lean, rich situation upon tip in. I tried datalogging like Wolf said but that just ended up frustrating me as I could not focus on one specific cell very easily. My next method was to have a friend ride with me on the highway and watch the tip in table using live tracing as I tried to move the throttle to hit certain cells. We found that while there is no direct correlation in the numbers in the tip in table to change in throttle position you can actually make it in to the high teens of tip in. I would constantly hit the gas and watch the XD-16 as my friend watched the tip in table using live tracing. We focused on the lower cells first and moved on to the higher cells if we were satisfied with what we had seen. As we went through the table we entered new values into a second ST window and saved this base map. This method took about 30 mins but I found it more productive as I was able to focus on certain areas until I could be confident in what was going on in that area of the table and what changes need to be made. Like riftswrx said my values have ended up being a good bit higher in the lower end of the table and a good bit lower in the upper end of the table. Just thought I'd let you guys know my method of tuning this table as I found it much more productive than datalogging and trying to make adjustments based on those.