TD_D mod journey. From stock to the 'bastardmobile'

Forums New posts
Td_d

Td_d

Commander In Chief
Ok - I managed to flash, and do a quick 15-20 minute drive. Woha... from the graph I'm busy crunching now, it's thrown it completely in the opposite direction now - it's pulling fuel hard at the bottom end - top end is within 5%-7% negative trim now. It's basically shifted the whole graph down by about 25%-30%. Slope still highly positive...

Here's the primary log: http://dl.dropbox.com/u/3099895/TD_D_Log_4_Feb_11_newMAF scale.zip
 
HolyCrapItsFast

HolyCrapItsFast

Drinks beer!
Okay let me have a look
 
HolyCrapItsFast

HolyCrapItsFast

Drinks beer!
My slow ass computer is still compiling the data.

Did you remember copy and paste the maf table from RomRaider into the results tab of the Maf Tool?

This could skew the results quite a bit.
 
Td_d

Td_d

Commander In Chief
HolyCrapItsFast said:
My slow ass computer is still compiling the data.

Did you remember copy and paste the maf table from RomRaider into the results tab of the Maf Tool?

This could skew the results quite a bit.
Click to expand...

Nope :oops:, my bad!

I'll do so with the logs off the new run tomorrow morning. Thanks again for all the help!

Rerunning the data now, and yes indeed it did skew it markedly - other than a serious neg trim at the very low end (which is why I gather you want to change the latecies) the top end whilst erratic has flattened out quite significantly (but now all in neg trim). Your logic makes perfect sense - the 10% latencies and MAF scaling changes should bump the low end up and lift the whole range closer to fluctuating around the X axis, which is what we want.
 
Last edited:
Td_d

Td_d

Commander In Chief
Just flashed the revision (won't get the chance to drive now, it's midnight - I think I might get some dirty looks from my wife!). What was particularly gratifying to notice was that AF stored 2, 3 and 4 (which is the most important in transitioning to open loop I gather) were within 1%. Stored learning A? A whopping 25%. We're definately making headway here!
 
HolyCrapItsFast

HolyCrapItsFast

Drinks beer!
Okay. Let me have a look. We may just start scaling the maf at this point.

Oh I wanted to ask you if you have a map of when you had the stock intake on. I would like to compare it (with a 1.30 multiplier) to the current maf scaling. I was comparing it to one I downloaded from the web and I was seeing some funny things that would explain what we are seeing but I am not 100% sure that the one I downloaded is stock maf scaling.
 
HolyCrapItsFast

HolyCrapItsFast

Drinks beer!
IGOTASTi

IGOTASTi

System Operator
Staff member
There goes Holy's brain!!! LOL.
 
HolyCrapItsFast

HolyCrapItsFast

Drinks beer!
This one is a challenge I will tell you that much. It is going slowly but I believe we are making good progress. It is looking like the maf scaling is compensating for latency and add on top of that the change in MAF housing and injector scaling issues and you can see how difficult this can be at times. Sometimes is hard to figure out where to start first when one thing has drastic effects on another but I got this figured out. I just wish you could see inside my mind so you can get a better picture of it all.

I'm committed to this.
 
IGOTASTi

IGOTASTi

System Operator
Staff member
I wouldn't except anything less from you Holy! You da man.
 
HolyCrapItsFast

HolyCrapItsFast

Drinks beer!
Okay here ya go...

I kind of started over from the beginning with the MAF. I reverted to stock but applied the 1.30 multiplier to compensate for the MAF housing. I also put the latency back up to something a little more normal.

The reason I did this is becaase I felt the original MAF scaling was trying to compensate for maladjusted latency so the top end was made leaner than the bottom. This became obvious to me when I compared the stock curve with the current curve. Even if I'm not right there is a clear difference between the stock scale and the current, and the errors coincide with the differences in curve

I'm hoping this will flatten the slope and allow us to continue on to MAF calibration. I think if this doesn't do much to smooth out the slope we are just going to move on to MAF calibration because this simply could turn out to be the nature of your very custom setup.

http://www.drift-ready.com/Files/TD_D/2-05-2011.zip
 
Td_d

Td_d

Commander In Chief
HolyCrapItsFast said:
This one is a challenge I will tell you that much. It is going slowly but I believe we are making good progress. It is looking like the maf scaling is compensating for latency and add on top of that the change in MAF housing and injector scaling issues and you can see how difficult this can be at times. Sometimes is hard to figure out where to start first when one thing has drastic effects on another but I got this figured out. I just wish you could see inside my mind so you can get a better picture of it all.

I'm committed to this.
Click to expand...

And much appreciated. This certainly proves the merit of undertaking the tuning as a parallel process to the mechanical modifications - otherwise, it become a multidimensional, complex problem. I look forward to hearing your thoughts on this - as I'm learning about how the various variables interact, I'm beginning to realise the difficulty of troubleshooting this when one set of variables may (subtly) affect another set...
 
Td_d

Td_d

Commander In Chief
HolyCrapItsFast said:
Okay here ya go...

I kind of started over from the beginning with the MAF. I reverted to stock but applied the 1.30 multiplier to compensate for the MAF housing. I also put the latency back up to something a little more normal.

The reason I did this is becaase I felt the original MAF scaling was trying to compensate for maladjusted latency so the top end was made leaner than the bottom. This became obvious to me when I compared the stock curve with the current curve. Even if I'm not right there is a clear difference between the stock scale and the current, and the errors coincide with the differences in curve

I'm hoping this will flatten the slope and allow us to continue on to MAF calibration. I think if this doesn't do much to smooth out the slope we are just going to move on to MAF calibration because this simply could turn out to be the nature of your very custom setup.

http://www.drift-ready.com/Files/TD_D/2-05-2011.zip
Click to expand...

Thanks George - I'm glad you managed to compare with a more stock MAF scale - I think what you're suspecting is exactly the problem - rather than dialing in the latencies properly first, I think the tuner moved too quickly to try and compensate for the errors by tuning the MAF scale on the opposing edges - which means as we normalise the latencies, we get for lack a better word, a see-saw effect. I'll flash and get some logs in - keeping my fingers crossed!
 
HolyCrapItsFast

HolyCrapItsFast

Drinks beer!
Td_d said:
Thanks George - I'm glad you managed to compare with a more stock MAF scale - I think what you're suspecting is exactly the problem - rather than dialing in the latencies properly first, I think the tuner moved too quickly to try and compensate for the errors by tuning the MAF scale on the opposing edges - which means as we normalise the latencies, we get for lack a better word, a see-saw effect. I'll flash and get some logs in - keeping my fingers crossed!
Click to expand...

Me TOOO!
 
Td_d

Td_d

Commander In Chief
Well - believe it or not, I just took a drive (I'm just too damn curious - and it helps that my wife's passed out ;-) ) - I think you nailed it. From the numbers I'm crunching now, the graph is now looking more like a sine wave around an axis below zero. There's still a big dip very early (which biases the slope upwards - so there might still be scope for adjusting latencies which I gather affects the bottom end much more than the top) - but the curve is much flatter now - see for yourself:

http://dl.dropbox.com/u/3099895/TD_D_5_Feb_logB.rar

Yup - just finished crunching the numbers (man, it takes forever!) - here's the graph - and frankly, the positive trim up at the top end is already well above 60g/s, which is basically going into open loop - so I think this did the trick. Untitled-1.jpg
 
Last edited:
Td_d

Td_d

Commander In Chief
Hi again - struggled to get enough driving time today - here are two logs - about 10-12 minutes worth driving, hopefully it will show the trend.

A question of interest - with the car in s# and a new map, idle is very poor, yet in S it's good (in fact a little high) - with all the maps to date. What's the mechanics behind that? Any idea?

http://dl.dropbox.com/u/3099895/TD_D_6_Feb_logB.zip
 
You must log in or register to reply here.

Similar threads

D
EJ257 - 2008 STI Hatchback, engine replacement?
Replies
0
Views
456
deraeler
D
A
WOT stumble/lean
Replies
0
Views
680
Abaddon04
A
J
2014 wrx sti; 90k miles; columbus, oh $19.5k
Replies
0
Views
785
jhestyr
J
Alin
State Of The Forum In 2024?
Replies
1
Views
316
IGOTASTi
IGOTASTi
cyclefiend1123
Horn Stuck On?
Replies
0
Views
353
cyclefiend1123
cyclefiend1123
Top