TD_D mod journey. From stock to the 'bastardmobile'

[Td_d]

So we're basically going for the same kit, chuckle... You can check out the tune with pleasure - I've spent a lot of time refining the low end driveability over the past months, got the fueling down pat now, and the timing sufficiently high at the lower loads to not get knock and stumbles. I've pushed the top end quite a bit, but safely - I never get top end knock. Running 1.8 bar, I've not quite finished pushing the AVCS possibilities up top - running 30 on intake, tapering right down by the early 5's, pushing exhaust of 23* in those areas now, from what I see, could probably push it right up to 30*, but have not had the time to do the required WOT runs.

I'm amazed my stock clutch is still holding - but I don't think it's long. SAE corrected for altitude, I'm sure I'm well over 400whp, probably close to mid 400's - so I wouldn't be surprised if actual torque, uncorrected, is reaching the 400 level - 20-30 over the stock torque specs. Can feel it slipping between 4000 and 5000 in 4th - catch, grab, catch grab...

 

[Td_d]

I wish E85 was available at the pumps here, would switch over in a heart beat. One of these days, I'll try doing an ethanol run and tuning for it - would there be any downsides to 100%? Cranking? Race fuel here is only 102 Ron and hard to get, so I'm not sure whether it would yield anything better than my current meth setup...

When I get the cams put it, I'll also have the intake manifold ported, and the TGV bridges deleted (butterflies were removed, but the bridge is still in). Mechanic was showing me that they're starting to do exactly what you were thinking - using NA intake manifolds on the Sti's. He's got a Forester manifold he's busy converting.

 

[Td_d]

I'll ask him to keep me in the loop, see if if figures it out - like Scott, he has no pretence to be a tuner, but mechanically gifted and not afraid of fabricating anything basically.

 

[Td_d]

So I've discovered a situation where the Subaru knock control system / process is fundamentally self-defeating :tard:

In a nutshell, if your timing is too low in the bottom ranges, the system canibalises itself, by stripping off more and more timing, thereby increasing knock, not reducing it. It results in the stumbling / hesitation in the low RPM ranges to get progressively worse following a ROM flash, as many people have reported - as FLKC is notching up and up, compounding the problem.

 

[HolyCrapItsFast]

Haaaaa???

 

[Td_d]

Yup, think about it - I'm talking real low loads here .9 to 1.3. If your timing is overly retarded, it's obviously getting too hot , heat leads to knock. I've been testing this theory for some weeks now - the original timing I had, I discovered, for whatever reason, had a lot of timing pulled all across the really low loads - from 1.5 or so to zero - god alone knows what purpose the 'tuner' thought this would serve. It's not even targeted like the spool effect - blanket lower timing. So in implementing the spool technique, I exacerbated the situation. In effect the timing is up to 10-12* less in some of these areas than stock timing, ridiculous.

Raise timing - knock and hesitation starts going away. I have a tons of logs to point this out... Bottom line is, like Gino said, couple of degrees, not too much. And not too low either. Would be interesting to have an EGT to confirm this.

 

[Td_d]

Oh, and how the system fails in this situation is simple- it only has one real tool at it's disposal - pulling timing. I guess the Subaru engineers didn't take into account a whole bunch of enthusiasts tinkering :tard:

 

[HolyCrapItsFast]

Yup, think about it - I'm talking real low loads here .9 to 1.3. If your timing is overly retarded, it's obviously getting too hot , heat leads to knock. I've been testing this theory for some weeks now - the original timing I had, I discovered, for whatever reason, had a lot of timing pulled all across the really low loads - from 1.5 or so to zero - god alone knows what purpose the 'tuner' thought this would serve. It's not even targeted like the spool effect - blanket lower timing. So in implementing the spool technique, I exacerbated the situation. In effect the timing is up to 10-12* less in some of these areas than stock timing, ridiculous.

Raise timing - knock and hesitation starts going away. I have a tons of logs to point this out... Bottom line is, like Gino said, couple of degrees, not too much. And not too low either. Would be interesting to have an EGT to confirm this.

Makes sense... You are priming the cylinder to create a good deal of knock by heating it up prior to the intake and compression stroke. I have never found a need to reduce so much timing in the low load region other than to improve spool but never to the threshold of excessive EGT. Actaully I have been leaning out the fuel instead and increasing timing to control EGT. This does pretty much the same thing and you control EGT with advanced timing but not to the point of knock. This is the strategy that EVOs use with there Lean Spool logic. For what ever reason someone would go that low in timing, the way to combat EGT at that point is to increase fuel. This strategy is used by some tuners but as it relates to the high load regions of the map. They will use high levels of boost and reduce the timing to control knock and increase fuel to control EGT. It all has to work together and maintaining the same fuel but reducing timing is going to have the effect they are seeing in the low load regions. Timing wants to be a specific range and for the appropriate fuel being commanded.

This is true the engineers never accounted for us fucking with their good work but we are changing the game completely and all bets are off at that point. So we have to figure it out on our own but it is not that hard when you understand fuel, boost and timing as one thing.

Hmm... my low end timing is rather high, so is cruise. Maybe we should make a timing thread in elite... And also put more data into the AVCS thread there too. I may create an AFR thread at some point also. We need to really get pretty much all the basic stuff knocked out in there and then get some intelligent discussion and experimentation stuff up.

I would think we should put it all in one thread because timing and AFR are directly linked. There are strategies for adjusting both together and you would probably want to discuss those strategies rather than each parameter individually.

 

[Td_d]

Yeah, what can we say, it's one giant experiment! That's half the fun :;

I have no idea why he dropped the timing so low in those regions - in the higher areas, as you say, a good balance between timing, fuel and boost is critical. On the whole, I have to say that leaving the lower areas as stock as possible maintains best driveability (expect for issues like the atmospheric pressure compensations and load comps for example). Who knows what he was thinking... all bets are off there too! I actually think the 'base map' that he worked off was from a different year or region completely - the reason I say this is if you recall, some 40 pages back I asked why my very low load timing was 40* as opposed to 49* on stock. Subsequently I saw very similar cruise timing, on a completely different open source tune.

I should probably just revert to pure stock figures down low, but I'm loath to make any sweeping changes, it's taken months of tinkering bit by bit to get it just right. If it aint broke...

 

[HolyCrapItsFast]

don't fix!

 

[Td_d]

Got her running nice and smooth nice - upped the lower end AVCS again, and instantly got low load, low RPM knock and hesitations, so backed it off a little. So yeah... now what do I do Cam time! It's like a disease, I tell you...

On a more annoying note, I found another error in Airboy's spreadsheet - well not error, but since the sheets were only set up for SSM logging, all the limits are too low. This one is as problematic as the one I found at the top end of the MAF scale - there is a 'clear' function that removes all the previous calcs in the results page... .... up to 200 cells With CAN logs, you get up to 10,000 lines sometimes, easy. I could not figure out why the numbers were so far off the Matlab version - now I see the Matlab calcs were right, and the spreadsheet wrong.

So George - if you're still tuning for people, change the code in Visual Basic (or I can do it for you, no worries). Can I ask you another favour... if you've tuned a GR anytime in the past and still have the Roms... could you mail me a version with changes to the fuel overrun tables, and one without, so I can narrow down the code location? Still struggling to find it...

 

[HolyCrapItsFast]

Ever since RR implemented the fast poll feature I noticed the same limitations and have since modified my spreadsheets. :tup:

Unfortunately I have no recent GR tune that I can find... or any for that matter

 

[Td_d]

Talk about synchronicity - Airboy just updated some of his spreadsheets today for the same reason.

 

[Td_d]

So, decided to switch to pure Toluene to boost octane, for a very practical reason - I can now get 210l drum delivered directly to my house. Winner. Was previosuly using a 50% Toluene, 25% Xylene, and 25% Methanol mix, very effective, but I also suspect that the meth was causing some of my fuel trims havoc. So yeah, guess what, have to rejig the MAF scale again, it's gone a little rich ... fuck's sake, I really need to put a stake in the ground, and stick to it :tard:

One good thing that came out of it, in buying a siphon hose from a big hardware chain here, I discovered that they now stock the full Sonax range. Bonus!

Also after a long day out of town for business, on the way back from the airport, got to absolutely spank another porsche. That was fun

 

[Td_d]

Finally decided to start tuning timing per gear. The only reason why I haven't done so yet, is two fold - firstly, I needed to research how to approach it, and secondly, bear with me, unlike ATR, the multiplier being used on the per gear timing tables was hardcoded as a negative value in open source, which basically means that the tables could only be used to subtract timing - which is counter to what I would like to do, safely.

In a nutshell, I can run significantly higher timing in the earlier gears, and effectively reduce the compensations to 0 in fifth and sixth gear. Basically, the taller gears with the potential for extremely high loads rely on the base and advance tables only, whilst the other gears get increasingly high timing (from what I've seen, 3* to 5* more is possible leading to 1st gear).

So, out came the disassembly toolkit, found the multiplier, and defined it in ECUflash, and away we go!

I'll try and post results, could be interesting.

 

[Td_d]

Holy torque, batman!

Bumped up timing by 3* in 1st, 2* in 2nd and 1* in third - given that most of the tuning was done in 3rd and 4th gear. No FLKC at all - one knock at medium loads in 3rd that I paid attention to, other than that nothing more than the occasional tip-in knock I always get (need to still clean tip in up). I also find that you tend to get knock on the drive immediately after flashing a Rom, usually.

Car feels notably peppier in lower gears - I think I can probably kick up 1st and second another 2*.

 

[Td_d]

I see we have a number of new vendors, some tuners on here - I'd be interested in chatting to them about their approach to per gear compensations (without giving trade secrets away of course ).

I'm liking what it's done for the low end so far - muchos.

 

[HolyCrapItsFast]

These are some of the things I would like to put in my tuning guide... I have to get back to that soon. It's turning into a novel. I will give you props if you let me reference your findings!

 

[Td_d]

For sure, George - you were my mentor here!

 

[Td_d]

So, like a deer caught in headlights, I'm seriously keen on upgrading turbo and cams now. Early 400whp and the dom3xtr is getting tired

I know which cams I want (the Kelford 272's), but not sure on the turbo. Options are dom5xtr, FP black (although I've heard fitment can be a ballache), or I've heard great things about the Precision turbos. Budget is not too big a constraint (within reason).

Ideas?