TD_D mod journey. From stock to the 'bastardmobile'

[Td_d]

Well, have to say I love the opensource community... just discovered that there are extensive experimental (basically not-official Romraider disassembled) logger definition files with a shit load more data for the '08 sti, including such gems as EGTs! Will try them this afternoon, but this could be a goldmine of logging data.

 

[Td_d]

Here's the link to the experimental RomRaider logger files: http://www.romraider.com/forum/viewtopic.php?f=34&t=6681

And here's what one of the members managed to disassemble: http://www.romraider.com/forum/viewtopic.php?f=34&t=6058&start=15

We're unfortunately SOL with EGT's - the STI does not have a EGT sensor installed, damn (only the WRX's). What caught my attention was the cruise/non-cruise switch in the above link...

On a happier note - my injector scale and MAF hack have worked like a charm! Yeeeeehhhhaaa... had to clean up the scale a little (did 5 WOT runs over lunch, a little hair-raising in traffic) - but no longer leaning out at the top - in fact 4.69 V was at -0.8% - basically spot on! The load axis hack has also made the car feel a little smoother - you can feel it's a little less urgent (obviously, since the effective timing is lower) - but no knock at the top.

 

[HolyCrapItsFast]

Well - at least I know my reasoning is sound! The table is indeed limited to a maximum of 4 (on the 08 STI's) due to the formula, which means that I will indeed have to

a) Reduce the injector flow down by X%
b) Reduce the entire MAF scale by same percentage
c) Reduce the load scaling on all load axis (fueling, timing, etc.) by the same percentage.

In other words, artificially reducing the load the engine calculates to stay under 4. I can already confirm that this works - reduced the MAF scale by 15% and the injector flow by 15%, and load is not exceeding 4.06 now.

Yeesh, I wish I'd know this upfront...

Check out this thread on openecu:

http://www.romraider.com/forum/viewtopic.php?f=15&t=7254&p=68059#p68059

Fucking ATR messes with me every time. I had a really hard time wrapping my head around this because in ATR when you reduce your injector scale you reduce the amount of fuel. In RomRaider it is opposite.

So you reduce the injector scale to increase fuel by 15% and the reduce MAF scale to remove 15% fuel to even it back out and then you rescale your load axis?

 

[Td_d]

^ - yup, exactly. I built a spreadsheet (looking at this thread, I'm amazed how much I've learnt in the last 4 months!) that basically took the current tables, calculates a ratio between the old flow scale and the new one (which was 15% less), scales the MAF scale proportionately. I took it a step further for load - since I had a known set of timing and AVCS that was good from the first tune, which was at a injector flow of 714, I reduced the load axis in a ratio relative to the new flow (which is 652, btw) - so effectively my load cells came down by 9%.

I've got to tweak the timing advance down a little at some areas to deal with some minor knock (-1.4 at 3000rpm, low loads and 4200 at high load), but other than that it seems to be working very well.

 

[HolyCrapItsFast]

Very interesting indeed! This is great to know!

 

[Td_d]

Luckily, it seems that the closed loop tuning we did upfront, and the load compensations, have held up relatively well (using learning view, CL is pretty good, and no stumbles).

 

[Td_d]

Another interesting little morsel - the latest RomRaider beta (0.5.3.RC8) has a built in road dyno (much the same as virtual dyno - based off Airboy's log interpolation spreadsheet) that now also includes the 08 Sti's - so you can tune and check for power gains using the same logs. Awesome!

 

[Td_d]

I can barely believe it, but basically that's the MAF scaling done. Open scale is now all under -1.5% AFR differentials. A small stumble has crept in again at 2900 RPMs, so I'll have to do another load compensation run, but that's pretty straighforward (and I was expecting it with the injector / MAF hack). Only knock I have is at around the stumble (-1.4 at very low loads - 0.9 to 1.1) due to leaning out, but new load comps will sort that out. All other knock dispatched with some tweaks to ignition advance at the high load columns.

So, new load comps, and then I think a couple of good road dyno pulls - I'm keen on seeing what the power and torque numbers are.

 

[Td_d]

I still have one nagging question. Given that I have a meth system that triggers off IDC and boost, would changing the flow scaling have changed when it triggers? Its still spraying perfectly at high loads / boost (I can see by the display) but it feels like the meth injection is starting later. Maybe I'm being paranoid - my gut tells me that despite the changed scaling, all I've done is fool the ECU into reading lower loads (and adjusted fueling and timing to match) - but the real load and fueling have not altered.

Too much thinking? :tard:

 

[HolyCrapItsFast]

In the end of it all you are still flowing the same fuel as you were before so the resulting IPW is the same. IDC will remain the same.

 

[Td_d]

Yeah, I thought as much. She's running absolutely beastly - it's winter here now, so the weather has cooled down nicely, so no heat soak and the response is brutal (well, when I say it's cold I don't so much mean NY or midwest cold, chuckle - we suffer with 4deg celsius mornings! Africa has it's advantages ).

I think to acknowledge all the help I've received on this amazing forum, besides doing a road dyno, I need to make a short video, with sound for y'all

So what the hell am I going to do with my time now... Oh, I know - put in new injectors! Seriously - is it fairly straightforward to install new injectors - if it is, I think I'll do it myself, rather than all the usual time wasting at the shop...

 

[HolyCrapItsFast]

Yes it is rather straight forward and with the smart fuckers we have here we will sure be able to help out if you run into a snag.

I have the utmost confidence in you.

Heck you should document it and make a how-to!

 

[IGOTASTi]

Yes it is rather straight forward and with the smart fuckers we have here we will sure be able to help out if you run into a snag.

I have the utmost confidence in you.

Heck you should document it and make a how-to!

Yes Holy we do have some very smart people in this forum. I'm so glad IGOTASTi is doing something positive for the community. It's really nice to see people helping each other. :thumbsup:

 

[Td_d]

I guess if I go ahead, I'll have to start bugging some of the mechanical gurus on here like Reks - as opposed to harrassing George all the time

I've already got quotes from Jordan for the 1200 and 1400 five-o's, actually reasonably priced.

I definitely want to do this sooner, rather than later - firstly to get IDCs down - I'm at the edge already, but secondly - for powa! All in all, I think I'm relatively conservatively boosted for a Dom3XTR (1.55bar - 22.5 PSI peak boost held), and given that my block not only has forged pistons, rods etc. but also Darton sleeves I can certainly go up to 26 psi to get the full potential out of the turbo.

Quite keen to see the road dyno results, butt dyno tells me I'm probably at the 400whp mark.

 

[Td_d]

Ok, managed to do a quick run on the highway, not the best stretch, slight uphill so I think it's probably mildly understated (never realised how difficult it is to find a good stretch of road...).

Using virtual dyno - uncorrected, so what the car is actually putting down at an altitude of 5600 feet - - 366WHP, 345lbft. And the curve is now looking so much better - rising smoothly, and holding power all the way past 7000 with ease (I think it's slightly overfueled at the top, there's a small dip). I'd like to do some more runs before I post a graph, but I think this is near the mark.

If I have to do an SAE correction on it (so that it's comparable with my very first dyno), in other words compensate for altitude, but also drop for the fact that it's cold - peak of 423WHP and 397lbft. Yeeha!

I'm just recalling Jared asking me on IW what my target was - this was way back... 420 to 430WHP. This car was dyno'd stock at 227whp (and that's corrected!). So basically a 200whp increase, nearly 100% increase from stock.

I guess that explains my smile everytime I drive it :tard:

 

[Td_d]

Eh, alright! I'm not trusting this one too much in terms of curve, but I think in terms of the peak, it's probably thereabouts. Just need to find a better spot to pull, and see if I can also do an insane 4th gear on the road, which is always fun...

Not liking that dip at the top - I think it's a little too rich:

 

[Td_d]

Oh - and this is without the wideband being logged (SD card, didn't have the time to hook up the netbook). I guess I really should do that to get a good idea of what's going on at the top.

 

[Td_d]

Managed to do a couple more runs - power is roughly the same - but looking at the wideband AFR curve, I need to do some cleaning up (also edging up to 11.8 AFR right at the top, I'm targeting 11.47 - it was however previously tuned for 11.8).

 

[Td_d]

I'll have to drop the injector scale and related load / fueling axes by a further 5%, dammit. I'm still hitting calculated loads of close to 4.4 with the current scale - need to get it below 4, so injector flow needs to come down to 615. I think that's why it's leaning out at the top - interpolating the data relative to the fueling targets, the dip is actually rich at 4200-4300 in the graph above - what looks like a lean dip at 4800 is actually tracking target...

 

[Td_d]

Hmm... I think I solved the mystery with the leaning out towards the top. I found an error in the code in the OlMAF2.2xls spreadsheet - I think I better inform peeps on the forums. I could not understand why the top end was persistently 4% to 5% off (which is really not a good thing when the MAF sensor is reading 4.82v !). I started looking at the data manually, and lo and behold, the code that categorises the data as per the row columns stops short (4.65V in my case). So what was happening was that all the data above that was being erroneously bunched into that one cell - and the values that I put in up higher - basically raised by what seemed a reasonable % - were actually to low. So, effectively I had an overshoot on AFR, and then an undershoot.

Busy fixing it now, we'll see tomorrow when I do some more WOT runs.