TD_D mod journey. From stock to the 'bastardmobile'

[Td_d]

Chuckle, chuckle... yeah, I wish I had more time to actually wrench as well - I certainly have the aptitude to do the head work, and definitely not afraid of getting my hands dirty - just time. Speaking of which, what a fucking afternoon... Financial year end, and of course, our server crashes. And I mean crashes - blown hard-drive on Raid 5 array. That's been my day, bringing the damn thing back up (IT support is amongst my numerous tasks here...) - however, writing this as I leave work, success!

Man, I need a whiskey...

 

[HolyCrapItsFast]

What?! I don't know anything about a certain plan to integrate suspension and tuning into one switch to change from dd to monster in 1/2 a sec.

Also don't know anything about concealing the entire setup including a full cage with various trim pieces and insulation.

I mean, what kind of talk is that?!

Oh I was complimenting your intelligences and your unwavering ability to do things that others say you can't do, and make them work.

I was thinking who else could build a monster and still drive it on the street. If anyone can do it I think it would be you. :tup:

 

[Td_d]

Thanks George. Was hinting at some of the ideas is all. You also know I'm crazy enough that I would drive around at 800 WHP daily. :tard:

Hell yeah! Oh, no wait, didn't I just say I'm sensible? Meh...

All I can say is, man the power is addictive. The car already pulls like a freight train, and all I want is more. Gave a 335i an unpleasant surprise this afternoon.

 

[Td_d]

Damn, what a week, still have not had time to go get the wideband fitted. Been fiddling with load comp tables some more (this is becoming an obsession!) using Mickey D's spreadsheet and CAN data (Airboy's spreadsheet just falls over with all that data, and the macros keep on crashing - very frustrating, because it's great when it does work) - and lo and behold - I have a completely stumble free map! I think given that I know understand exactly what it's doing (100's of Romraider forum pages read...) I can judge what the results from the spreadsheet are.

It's more of a mission, since the tip-in correction (excluding anything over 2% and 4%) have to be done manually - but I like the approach that MickeyD initially took (even though he now uses the other spreadsheet!) - half the compensations are applied each time to the existing maps, so it's a more gradual iterative process. And the results speak for themselves, I have never had the map this smooth (even when the tuner applied his quote unquote 'secret' fix to the stumble, back in my ignorant days). I think being able to use CAN data as well gives a much broader set of numbers to use.

Must get this wideband fitted and finally scale up top. On a more annoying note - thought it might be prudent to log IDCs again, since it's gotten much colder and you can feel the car is pulling more air in. Damn... up to 95%. I wish the mechanic had left the 1000's in and let me tune it :tard:

I'll have to get bigger injectors asap - definately 1200's or 1400's since I aim to make the E100 plan work. Jordan - if you're reading - what's the damage on those five o's?

 

[Td_d]

Ok, wideband finally goes in after a rather hectic week this Thursday.

OT - Did I mention that I absolutely love this car? Had what looked like an entry level carrera fly by whilst I was doing a log drive, and I simply could not help myself... What was impressive was how much faster this car is. Had a friendly chat at the next light, he looks at me and says - dude, your car flies!

 

[Td_d]

Yeeha, finally! MTX-L is in and working beautifully... It's a pretty thing too. Was a bit of a mission to find a place to mount, given that with a integrated controller and gauge it's quite a fat bastard. So we looked, and I suggested we put in the driver's side vent. Came out really neat, I'll post a shot.

Ok! On to do some reading on how to scale open loop, but Holy - could you give me the 2 cent tour, and I'll start logging? From what I saw I'm pretty close at full tilt, slightly rich (10.7), which is a relief. But will only know for certain logging across a full WOT pull.

 

[HolyCrapItsFast]

Does romraider monitor commanded fuel or open loop target? I forget.

Also is it all set up to log the gauge?

 

[Td_d]

Targeted fuel IIRC - I'll check. It's all set up to log, will need to use the laptop this time to get wb feed.

 

[Td_d]

So - I don't know if you saw what came out of my discussion on IW regarding the stumble (COBB's comments on the map ratio) - if that's the case, the whole way of differentiating between cruise and non-cruise using the OL/CL switch = 8 or not is kinda broken!

I've raised it in the Romraider forums so that the developers can look at it, quite exciting I think if it is done!

 

[HolyCrapItsFast]

Yes I suscribed to it. :tup: Can you send me a link to the thread in RR forums? I would like to follow that also.

So for open loop I don't use any special spread sheets to derive my final MAF. It is a combination of comparing and smoothing.

I usually start out with a log of CL/OL switch, MAF Voltage, Calculated Load, RPM, Target Fuel, Gauge AFR.

I then compare the AFR at a specified MAF voltage and I compare that to the target AFR. If it is close I will leave it and move on. If it is not I will either make and adjustment to the MAF calibration or the Primary Fuet table for that load and RPM. It depends on what I see or how I feel about the readings. You have to rely on allot of intuition.

I usually consentrate on ranges of MAF Voltage instead of one MAF Voltage individualy and I apply some interpolation smoothing between that range.

My process goes like this...

1. I start with a log for a rolling 1st gear pull and observe the AFR from the onset of boost.
2. If needed, make an adjustment to the MAF voltage regions as observed in the log.
3. Take a log for a rolling 2nd gear pull and observe the AFR from the onset of boost.
4. If needed, make an adjustment to the MAF voltage regions as observed in the log.
5. Continue on through 4th gear with this process.
6. Finally take a log with a pull from 1st gear all the way to 4th gear and confirm the ARF's against target and make minor adjustments if needed.
7. Apply smoothing to the final MAF calibration and you are done. (3 cell smoothing only)

 

[Td_d]

Thanks Holy - I actually started already using the open loop scaling spreadsheets on the open ECU forums, and all I can say is thank god I have very conservative fueling at the top... scale was off (lean!) by about 4% mid load rising to 8% high load, shudder. Luckily - and I've also done some calcs on this - at the worst areas (i.e. hi load high RPM) I'm at an effective 11.5 to 11.8 AFR (and logged confirmation of this from the MTX-L), which is quite common with meth tuning from the reading I've done.
On the Map Ratio being discussed, I started a thread (very tentatively!) in the developer section of the openecu forums (big brains in there) - and it looks like the parameter that COBB is referring to has indeed been reverse engineered in RomRaider - its called, predictably, Map Ratio :tard: What I'm not clear about is that COBB seems to intimate that it applies directly to the load comp tables as well, in which case I'm not sure. But it's likely it is this very parameter. Now what I'm trying to get my head around is how to incorporate this in terms of filtering, especially for the load comp tables. One can log the parameter - and if it's on either end - 1 or 0 - it's easy CL or OL. However - how to handle if it is in between...
Here's the pic of the long awaited wideband - I think it came out really neat!



View attachment 1100The link to the thread, btw is http://www.romraider.com/forum/viewtopic.php?f=14&t=7241&p=67927#p67927 (in it there's a link to the load compensation thread as well).

 

[HolyCrapItsFast]

Yeah maybe I should start using that spreadsheet for OL instead of relying solely on me :lol:

Let me ask you... To derive the AFR error do you first calculate the % error by comparing target with the gauge and do you use lambda or AFR? I would imaging you can't derive AFR error from short and long term trims like we do with CL since they only reference the O2 sensor.

 

[HolyCrapItsFast]

BTW your dialog at IW and RR is impressive. The student has surpassed the master. You are now a full fledged Jedi night!

 

[Td_d]

Hey Holy - thanks for the compliment, but I've still got plenty to learn! I'm curious to see what comes of this - I really don't have the wherewithal to disassemble the ROMs, but there are people on there who definitely can do it. What I just figured out :facepalm: is that Bill from Cobb is Merchgod! Let's see what comes out of it!

The spreadsheet uses the open loop fueling target and compares that to the actual value that is picked up by the wideband - it's quite cleanly done - you paste your open fueling table straight in from Romraider, and it pulls it 'automagically' in the area used to compare against the actual AFR. I've been using AFR - am I correct in assuming that the AFR that the wideband is reading (using the gas ratio) is the actual AFR, or do I need to correct it for the meth ratio? If so - how?

 

[Td_d]

Hmm - ok, just some cursory browsing lets me see that lambda is far preferable (assuming that the wideband can - which I think it can), as it takes the changing stoichometric value out of the equation. That basically means that I've like made it a little too rich at the top, so no harm done.

What lambda do you normally use at highload WOT - .8?

 

[Td_d]

Ok - but then how do I calculate stoich for a 50/50 meth water mix?

 

[Td_d]

Well - that's effectively the approach that I've been taking now, following the fueling maps that the tuner originally put on, for meth, which targeted 11.76 AFR. What's interesting is that the revision to the scaling that I made were resulting in AFR up to 11.8, not knocking and delivering immense power.

Can I monitor EGT with the stock ECU, or will I need a separate sensor?

 

[Td_d]

I'm really beginning to get myself into a knot here. The LC-1 (basically the same sensor as mine, older controller) reads leaner on meth - so effectively, if I'm seeing AFRs of 11.8, the true AFR is actually lower. Now from what I'm reading, one can tune for slightly higher AFRs with meth and water mix, given the lower temps - around 11.5-11.7.

What's hurting my brain is, am I correct now in assuming that if I've tuned the OL using corrections based on the wideband reading (and basically got the errors down to very small percentages, sub 1%) - I have effectively tuned it for a much richer AFR? Furthermore - from what I've read as well, given the different stoich ratio, achieving the same lambda (i.e. on gas, AFR 11 is lambda .75, and hence meth would be lower given a stoich much lower than 14.7), one should tune for lower AFRs using meth?

I'm getting myself thoroughly confused here, and I need to have at least a safe working OL tune without using an EGT (which many people seem to manage?).

 

[Td_d]

Ok, I think I have a path, chatted to the mechanic this morning - I had a niggling feeling that I was throwing too many variables into the mix at once.

I thought about this last night, but wasn't sure - but he recommended that I initially scale the open loop MAF without meth, which makes sense - I have a known stoich AFR. Once I've got that buttoned down - then start dialing in the meth, in terms of AFR and timing (for which I actually have a good base line from the initial tune - whilst the closed loop was horrible due to the latencies being completely off, the open loop had been tuned well for power, and the scaling for that set of flow and latencies, so the timing was pushed up to the safe limit without knock). He's more comfortable with a less aggressive AFR up top of 11.5 for meth (tuner had it dialed in for 11.76_ - although I do know that the car was liking the 11.7 (but I don't know whether EGTs are acceptable).

So, now I need to put on a much more conservative non-meth fueling and timing map, and scale off that. Any suggestions - should I set WOT for 10.88 through the entire WOT section, and then drop the current timing by 4-5 degrees (which I gather is what meth will allow you to push)?

This is where I'm getting a little confused - lean = high egts and likely knock. Too low timing = high egts and likely knock. So would the best option then be to drop AFRs right down to say 10, leave the timing, switch off meth and tune open scale.

Help!

 

[HolyCrapItsFast]

I have to read through that all a few more times. But if I get you so far...

Essentially you want your final AFR at the wide band to match what is shown in the primary fuel table plus any compensations and when all is said and done. You can make your primary fuel table leaner than with just straight gas and then try to target that value with your wide band while you introduce meth. You will have to find the best values that work for you. Don't concern your self with stoich of meth versus gas because it is all the same when talking lambda. Plus you are mixing it with the gas charge so you create a whole other animal.

Remember that if you add meth on top of gas you will richen the mixture because meth is a fuel. So as you tune and increase the meth, you need to compensate by reducing the maf scaling to command less fuel. You have to trust your wide band and tune as if you were tuning for straight fuel and based it off the primary fuel table. All of your adjustment will be done with MAF scale.

Now for a 50/50 mix of meth and water, this will have a much less effect on AFR but you will still notice a change when you begin to introduce it. This is because water is inert and is there simply to cool the charge and increase compression.

If you go 50/50 or 100% meth you need to have a tune for each respectively.

I would start with a known good map for gasoline. Have your timing adjusted for gasoline. Then begin tuning for meth and once complete move on to timing. Don't try to do them both at the same time. I always keep things mildly rich and I would never go beyond 11.2:1 AFR no matter the fuel, primarily because I don't tune using EGT. Keeping AFR's low pretty much guarantees a good EGT and I'm willing to sacrifice some power for a margin of safety