COBB AP/ATR Beta Fuel Scaling
 

Post your scaling information/experiences here; I'm sadly still waiting for my maps. :(

Subbin. No results here since I'm currently on 93 octane.

@silvapain will have some, as well as Dustin.

I have a map to test, but won't have data until either tomorrow or the weekend.

I was just testing the maps without changing anything else. Basically to make sure the fuel trims behaved as expected, without any changes in fuel or e85 mix ratio. So far, worked like a charm.

Still waiting for Evan to get back to me, but in the meantime...

:grouphug:

Rob, If you need another car to increase the testing data shoot me a map.

@rfinkle2

I'll send my maps today. subbed.

I don't have any MAF datalogs yet, but...

I'm using a map originally tuned for 93 octane pump gas that's been adjusted 40% by David at COBB. The LTFT/STFT combinations are higher than they were on pump gas, but based on discussions in other threads my theory is that it's attributed to the 40% increase in injector pulse width. This means the MAF scale is much more sensitive to errors. A simple MAF cal will fix it.

My CL AFRs are meeting their targets no problem. I'm seeing 1.07 lambda in light cruising (my CL target), and 1.00 lambda at idle in neutral (an AFR target from a yet unexposed fueling table).

I edited my current map in ATR and checked the box for 'keep non-table data' prior to saving. Afterwards I sent it to @David@COBB and he verified the beta changes were kept.

Good to know we can still fuck with the maps and keep the scaling intact.

Edit: At least until these settings are exposed via ATR.

eff you guys in the eye...I'm jelly!

fuck it...I'm gonna go make the run out to that place and get me some E85...

Here's my initial data, somewhat cleaned up and consolidated:

http://i359.photobucket.com/albums/o...ntComparo1.jpg





I did a "maf cal log" last night on the way home, on my current 2 gal e85 mix map. Basically just getting a snap shot of the various fuel trims across the entire maf range. Then i flashed the new 20% scaled map from David this morning... put some miles on the car (commute to work is roughly 25 miles), and did another "maf cal log" on the last road to my work.

I combined both LTFT and STFT for a total trim... and included AFR so you can see that these trims are indeed resulting in the right AFR, and then tried to match the maf curves as close as possible, just to align the various LTFT break points.


Close enough to convince me though... the new "scaled" map is most certainly resulting in more fuel (more negative fuel trims). Roughly 15-17% more.

Dustin, I should thank you for the info but groan you for the huge-ass graph.

Just cause you didn't say it explicitly... but your running your pump gas maf curve on 100% e85... with the 40% scaling, and indeed it's working as intended, right?

I'm not surprised that the trims are off "slightly", i mean 40% is kinda a shot from the hip unless you used an ethanol content analyzer and got super scientific on it lol. But close to zero, and the fact that your running 100% e on a pump gas maf curve says a ton.

Great to hear :)

Hahaha, it's not that big, but i can copy and resize if you want me to.

Correct; I'm on 100% E85 but using the MAF scale I made on straight 93 octane pump gas. As a matter of fact, my entire map is from my pump gas tune, just with some adjusted timing and AFR targets to optimize for E85. No MAF or load adjustments at all.

For reference, I scaled my MAF roughly 39% when I went from pump gas to E85. It's not a linear change, though; E85 needs even more fuel than pump gas at high MAF g/s to hit AFR targets. Therefore my MAF was scaled closer to 45-50% at higher voltages.

Fucking dope.

This can probably be accounted for with seasonal blends; the ethanol content in tank is most likely markedly less now than it would be in the summer time (which may even be dead on).

Honestly, I'd leave it pulling fuel as it is now so that when you switch blends you don't run lean (not that that will hurt anything except power output on pure e).

I don't intend to derail this thread, but think your tune is still performing as well as it is because you are pure boost tuning.

Calculated loads will still need manual adjusting (correct me if I'm wrong here) if you just convert an old 93 map and still have some active load comp tables.

So we just got about 15% more fuel out of the stock injectors with this new Beta software? (Sorry I'm a noob)

Since the MAF is no longer scaled, his calculated load values should be much more reasonable than they ever were when the map was manually adjusted for the fuel.

Logs?

I don't think we have any injector adjustment stuff in these maps. Correct me if I'm wrong, but I'm pretty sure we are just talking about scaling fuel requirements via scalar instead of MAF (for use eth guys).

Anyone?

Calculated load is based on MAF g/s and RPM. Load is just another term for Volumetric Efficiency (VE), which is the percentage of actual airflow through the engine at a given RPM compared to the theoretical ideal.

My load values on E85 are all jacked up due to tricking the ECU by scaling the MAF. On the COBB beta map, my MAF isn't falsely scaled, so my load values are 'actual'. Therefore, I don't need to adjust any load-based tables.

Now, at WOT my car will hit higher load values on E85 now than it did with pump gas because I'm making more power. I may need to up my load targets for that reason to hit the boost I want, but I don't have to change it if I don't want.

I wonder if this will resolve the issues seen over in Dano's boost tuning thread regarding some sort of artificial load limit capping output of the engine no matter what tables are modified.

Correct. This beta map from COBB let's those of use running Ethanol blends make one scalar value change and not have to use the MAF scale to 'trick' the ECU into running the correct amount of fuel to hit our AFR targets. This won't gain use any more injector headroom.

Crappers.. mystyped when I said "adjust calculated load values"... intended to say load axis tables will need tweaking FML.

A big step foreward in ease of use, nonetheless.

Gotcha, thanks!

I certainly hope so. I had no problems hybrid tuning on pump gas before going to E85 and scaling my MAF. The caveat is that I haven't load tuned on pump gas with the latest firmware revision; if the issues we're seeing now are due to a firmware change this won't fix it.

One way to find out.

This is essentially an injector "adjustment". Meaning it is fooling the ecu into thinking we are taking in more air than what the MAF curve says, somehow. Of course we don't know the mechanics [yet], but it is the same thing we are doing with MAF scaling for E85. Because the fueling calc is based on the MAF curve, there has to be a multiplier in there somewhere (or actaully a multiplier scale, since the scaling is NOT linear on E85, as silvapain said).

So no actual injector head room was gained, we are just injecting more air per revolution, but without the scaled maf curve.

Unless David has been making tweaks to inj latency, or something along those lines. I have my fingers crossed that he's found a way to actually make the injector open earlier than normal. That means we won't necessarily be "bound" between the intake valve opening and spark ignition events. We would theoretically be able to spray earlier, even with a closed intake valve, and significantly increase or fueling window.

Shit, honestly, i'd be willing to spray fuel at the tail end of the exhaust stroke if i could lol.

IIRC Evan is adjusting injector latency.

I think david is adjusting tables on fuel mass and latency or something along those lines.

I plan on attempting to load tune on this beta map, or at least hybrid load/boost.

Thanks finkle for your ninja forums skillz on finding this link:

http://www.mazdaspeedforums.org/foru...tml#post335113



That post shows the injector opening event hugging the intake valve opening event. So lower VVT and higher spark advance = much shorter fueling window.

There has to be something else involved if the trims/AFRs are staying consistent with only the scalar change. The MAF curve isn't a linear one, so there's no way it could just be a simple scalar value IMO.

+1, WTB moar injector

On these maps? You sure about that?

Whatever it is, as long as it works and we don't have crazy load values anymore, I'm down :)

That's what I was thinking too as I DID hear that they were working on getting us an additional 15%~ out of our stock injectors. (I'm assuming this is being worked on as we speak)

So we can use simple math to discern effective (current) max spray window based on map values across multiple tables...right?

Also, has anyone dyno proven that there is a benefit/detriment to running VVT values past spoolup?

This would be awesome and a HUGE step forward for our platform as a whole. You guys think I should update the thread title/OP with information regarding this fueling update as well? Basically make this thread about both fuel scaling and improved injector flow changes?

No. LOL.... "changes to your injector dwell time and mass of fuel scalar" are being changed.

If dwell time = latency than yes, if dwell time does not = latency than no, I'm not sure.

Bolded type is what is being tweaked.

Yup, theoretically simple math. But I think adding VVT in higher rpms (esp on k04) result in drastically reduced maf g/s (e.g. decreased performance). And right now, all this injector control stuff is pure speculation. Nothing has been proven yet, so i'd leave the title as is.



.... until we prove it ;)

Yessir. From the little testing I've done, we only saw IDC decrease because the air mass decreased as vvt increased in the higher rpms. On my most recent dyno session (not my car), I played with VVT on the bottom end and it didn't have any sort of effect on power (surprising).

A very basic primer on how the ECU determines injector pulse width:

1. The ECU gets a voltage signal from the MAF, and using the MAF scale, turns that into mass airflow (g/s)
2. The ECU determines RPM and calculates load
3. Based on RPM and load, the ECU determines spark advance, VVT, and AFR target
4. Using mass airflow and AFR target, the ECU calculates fuel requirement in mass rate (g/s, lb/hr, whatever)
5. Using a scalar that represents injector size (lb/hr; on DI this would also incorporate fuel rail pressure), the ECU determines required injector pulse width (in milliseconds)

It's the scalar in item 5 that COBB is adjusting I believe. It's a static percentage.

That still doesn't quite add up; step 5 amounts to a linear change, which we know doesn't work based on MAF scaling alone. What Finkle said about fuel mass makes more sense; more e = more oxygen getting into the cylinder (aside from the expected airflow) = higher scaled MAF value to equal the same AFR (if that makes sense).

If this is the way it is being done, it's pretty fucking smart. If not, please correct my understanding.

To clarify, what I'm looking to do is change the flow rate table of the injectors in the ECU - what some people call its scalar. Now what that does in turn is tell the ECU the injectors are either smaller or larger (in flow). From that, it calculates IPW to hit the AFR target you ask of it. In the case of E85, you say the flow rate is LESS so it INCREASES open inj. time to accommodate the additional mass needed to meet stoic. If you have bigger injectors, you'll say it has more flow rate so it'll cut down on the IPW.

This is the "correct" way of changing the injector routines from a fundamental level - or else it'll be correct for only certain RPM/Loads and won't scale with your car's different pressure/RPM/load.

And as I said before, still in development. For those of you that have maps, give me the final LTFT's off (and be sure to include +/- on those LTFT) so I can remake them with a final adjustment.

-David@COBB

Also, if any of you guys are still waiting for us to get you your maps, please send me an e-mail or PM. I was dealing with a pretty decent group of you yesterday and may have missed a couple.

The 'tweaked' maps that David is sending out are just estimates as to what "scalar" is needed for your fuel blend. I've been testing these for awhile now but we don't have it down to a science yet. We're still trying to figure out a good formula. The goal of having you guys see how it reacts on your cars is to try to figure out a formula that will work for any blend you have.

If you know your MAF scale is dead on for 93 and we sent you a scalar but now your trims are coming at -5 or so, that's because the scalar we gave you is really just our best guess as to what you need and may need some tweaking. Once these tables are released in ATR, you'll be able to go crazy with them to get yourself dialed in perfectly.

David, have you seen any tables or logic that look like it may control the injector opening position? Basically a way for us to start spraying fuel earlier?

even if I can't directly benefit from the E85 scaling, it appears that the fueling discussion here will have wider implications for everyone.

The benefit of this is several fold:

1. MAF will no longer need to be scaled artificially for different fuel blends meaning that going E85 etc will just be a scalar away - easier to tune.

2. Larger injector fuel potential again without messing with the MAF when and if injectors are available.

3. People hitting load caps because of MAF scaling have this issue resolved. **Here's something interesting ** Just like you can scale up for E85 you can also scale DOWN. If we have an unkown load cap/table at say ... 2.0 load - scaled your MAF DOWN so that you are at stockish loads and use this multiplier to get the correct AFR. This is called tricking the trick :) And it's a hack in the opposite direction to deal with an uncovered table or cap.

In terms of VVT. Increasing VVT even on the K04 results in higher flow but lower power at high RPMs. However, increasing it on a big boy turbo should allow more fuel and more power. Increasing timing also decreases the amount of fuel available so pick your battles :)

Ansolutely. What David is doing will be needed if/when we have larger injectors available to us. The ECU will need to adjust the dwell time for the higher-flowing injectors.


I noticed there was a discussion about injector latency and dwell times earlier. They are not the same:

Dwell time = how long the injector is 100% open, usually defined in milliseconds. If you take dwell time and divide it by the total available time that the injector can be open, you get duty cycle, measured in %.

Latency = how long it takes the injector to go from one position to the other; either from closed to open when the ECU commands it to open, or from open to closed when the ECU commands it to close.

Actually, you can benefit, Fobio. Get a bigger MAF housing.

To latch onto one of the topics - Injector duty cycle is calculated by a lot of things:

1) Crank Angle\RPM\VVT
2) Spark timing/advance
3) Load and airflow
4) Injector startup/duration/close time
5) HPFP pressure
6) Injector Flow rate size

This is why you can get 140-150% when using the old way of MAF scaling to run E85. Remember that it is an approximation.

As for opening injectors earlier, we need to find out if it actually moves the injection earlier for more time or just starts in the same place. This will ultimately determine if the stock injectors can run 100% E85 correctly.

-David@COBB

I am already running a bigger maf. But there's BlueStreak with the forged motor and aiming for 26psi+. He will likely benefit from the refined fueling control. This is exciting stuff guys!

I love to read the nitty gritty from all of you guys. I'm glad to have everyone contributing, and to have Cobb so involved in what was once an otherwise overlooked platform (@ least compared to other, more popular aftermarket platforms).

Thanks guys for the technical info, and David and Evan for working out some of the technical "stuff". (I'm particularly glad Evan has an ms3 and he and Phate are all about e85 LOL).

With companies like Deutchwerkes, Cobb and EFR working on shit for our cars, it opens up a whole new can of worms, imo.

:fingersx:

Can of worms for the other platforms on the road. :147:

If we can get larger injectors, control the injector open/close, and get fuel and injector scalars in the next few months, I see even more breakthroughs for 2012 than we saw in 2011. This is exciting stuff. 500-600whp without pouring in meth by the gallon may be closer than we think.

this thread is win, someday when im not so cheap i can buy alot of go fast parts, all this work from you guys would make tuning easier. Sure am starting to wish i had some e 85 around here.

Subd.

I hope all you fuckers have tiny dicks to make up for your super huge brains-----I'm so confused-----so much to learn yet. And I didnt even get a big dick.:ugh:

If you really want to know, pictures of my meat sword are conveniently located in the cock thread in VIP.










You're welcome.

ehh dont really wanna know, just jelly of the knowledge. Hard to catch up with you guys.

also sorry for fuckin up a good thread Ill stfu now

@David@COBB @Evan@COBB

I got DTC P2096 - Target A/F feedback system too lean today. I cleared the code, and after a couple key cycles I got it again. I flashed a new map to clear the LTFT's and did a datalog; it appears my STFT's are maxing out at +25 at very light throttle in 5th and 6th. See attached datalog.

If it was fine just after flashing the map, the issue is either due to a vacuum leak or an injector sticking shut.

The log is from just after flashing the map. It was also happening before flashing the map.

It only started after going to the COBB beta map.

Silva throw in injector PW in the log.

I am suspecting this is due to the poor resolution of the MAF at that low airflow. The MAF reading is almost the same as at idle.

For shits and giggles multiply the 1.37 to 1.60 V region by 1.1 in the MAF table and smooth the 1.60 to 1.76V and 1.29 to 1.37V regions.

Will do for the next log.

I wanted to get input from COBB before I did any MAF corrections, to see if there was any more information they needed from me.

Have you loaded up your 45% fuel map yet?

You need more fuel - the STFT's are at their max allowed error correction, so we need to up the amount.

Try the 45% and see if it helps. We might need to go higher if your STFT's are still too far.

-David@COBB

I know the max trims allowed are +/- 25%. My question was why is it only at this specific point? My LTFTs are within +/- 8% otherwise; not ideal by any means, but well within the CL adjustment range if the ECU.


Tapadatass

Dave, take a look at my comment a few lines above. His STFTs are only high at very low airflow values.

For reference, here's an Excel file of my pump gas MAF scale and my E85-adjusted MAF scale to show the percentage difference.

My averaged percent difference is 41%; hence the reason I'm using the 40% scalar

Never seen that before. Your car seems to be the only one that is having this issue.

are the LTFT's you are having typically positive or negative?

EDIT: One of our calibrator's looked at your logs, and said that it is most likely a MAF scaling issue - so try to change it and give it a shot.

-David@COBB

Could also be a slight air/vacuum leak past the MAF but less likely because it wasn't there prior to the calibration change.

Is your fuel pressure still all fucked up below 3k rpm?

I have both positive and negative LTFT's. That's the primary reason I don't think I need to go to the 45% scaled map.

Dustin, My HPFP issues have been resolved (at least temporarily) by cleaning the spill valve thoroughly. You can see in the datalog I posted earlier that I'm maintaining fuel pressure throughout the log.

the one time my LTFT kept worsening, was when I had a leak at the manifold/turbo gasket.

To test this parameter why even jump directly to 40%? Why not use a smaller step size that you can monitor directly without even having to use E85 in the car - just use the same fuel (same tank) and make small changes that can be tracked through trims and commanded PW changes as well as AFRs under WOT. This way there are less variables to consider.

Dustin also made a very good suggestion of making smaller increments such as 10, 20% and then scaling the MAF back by those same percentages. Then observing trims and WOT AFRs. No need to ACTUALLY use E85 and many map variations to test a fuel scalar.

If I was going from pump gas to E85 just now, that's how I would do it.

I don't know if the data table COBB uncovered is 0-D or 1-D, but if it's 1-D I'd get my MAF dead-on with pump gas, then incrementally increase E85 concentration and use the new table to adjust the MAF accordingly. That way I'd know my mass airflow readings are 'actual', and not skewed due to recalibration for E85.

I could easily fix any LTFT errors I'm having now by simply rescaling the MAF; but then I might be back where I started with mass airflow and load values that don't mean anything because they've been skewed too much.

Just change one variable at a time and don't let it be the gas since that's hard to control.

Slowly increment the scalar and scale down the MAF to maintain the same AFR under WOT and trims.

What I'm saying is that if the point is to maintain legitimate mass airflow values in the MAF table when going from pump gas to E85, you would want to adjust the fuel scalar table instead of the MAF table to hit AFR targets.

I think right now David's just looking for "proof of concept", and trying to establish how much of a tweak is required to get X percentage increase. From that data, he'll be able to wrap it all up nice and cute for ATR, and all we'll have to do is plug in a simple scalar.

Look at the correlation between boost (vacuum) and the STFT's. You have very high STFT's as vacuum increases. That looks like a very small vacuum leak, to me.

This is obvious, but: To check, just flash back to your other map and see if it does the same thing. If STFT's go crazy, it's a vac leak. If all is peachy, then it's the scalar.

I'm getting a steady 21 inHg vacuum at idle and my idle trims are stable, so I don't think it's a leak. I can do a vacuum and boost test anyways to verify.

I'll flash back to my MAF-scaled E85 tune tonight and see.

This is the exact same thing I was seeing. I pulled my inlet, patched any rough/scraped/scratched areas with RTV, lubed up the connections to ensure a positive fit, then retightened the easily accessible bolts on my intake mani and so far that seems to have fixed my high vacuum trim climb.

I have nothing to contribute at this time except a round of

at a boys! great work from all.

To clarify my above statement:
My trims would only climb ABOVE 12 inches of vacuum. Idle and WOT was pretty much dead on.

So if you're boost based, what is the difference between using these scalars I haven't seen (on Cobb's site somewhere?) for E85, and just re-calibrating your maf? I run a D010 w/m w/ 50/50 and just calibrating the maf for it.

Sorry for the noob q, I just came across this stuff tonight, and learning the E85 ropes for the big switch (well, small switch... 2 gals).

Uh, changing one cell vs 100?

If that's all it does, I can see it being very useful if you want to try multiple mixtures of E85 if there are set scalars for different % mixtures. I thought there was more to it than that, but it's late and I'm tired.

I saw a few people mention FTs... so these scalars just for your fuel trims and not wot afrs? I'm still playing w/ my meth, and boost, and will likely run a couple gallons of E85 in the near future to gain some timing. I don't mind recaling my maf twice, lol. I still think I'm missing something here :/

I think the goal of the project is to enable people with a decent 93 octane maf cal to easily convert a map to a particular e85 mixture through the scaling done "back office".

I think David may have mentioned (in one of the threads) that if necessary, tweaks can still be made to your maf cal, but the idea is to get you pretty close to running the afr's you were on pump both open and closed loop.

Good to see you around Douge.

Right now to run 3+ gallons of e85 we have to scale the the shit out of the maf sensor, which causes an overly inflated maf reading. Doing this causes all of the load tables to also read very high, and you literally have to adjust a shit load of tables to keep things happy. They are trying to get it so we can basically change one fuel table and be pretty much ready to go. To get an idea 4gallons of e needs about 10% more fuel.

From what I've played with though is 2 gallons needs very little adjustment, so if that's all you want to run. Toss it in at your next fill up and start playing. Lol.

Tappin

roger that...if your petrol maf cal is good there is no need to change it with 2-2.5 gallons of E. There should be no AF shift with that little concentration. I had to hit 4 gallons before I saw any shift and I went from 2.5 to 4 so somewhere in between there is when the shift occurred.

also between the two I had to add a shit ton of timing to make up for the slower burn :)

i had to adjust my trims after my first 2 gallons. MS6 tank is bigger too.

My wot afr's were good through 3 gallons, but @ 3 gallons, my trims were pretty bad.

Can't explain it, but it happened.

Ahhh shit. I may have made a table change and resaved the map without the "keep non-table data" option checked. Can anyone confirm this error would effectively take the e85 tweaks out the guys are working on for us?

Making ANY changes in ATR to that map is pretty risky, as you would need to have that box checked or it wipes my change to your maps out.

-David@COBB

Understood. (one of the reasons I wanted to be sure before flashing)

I tried to e-mail you yesterday so I could run the map, but you must be swamped with work.

Changes made remain in the map per our e-mail.

Thanks fellas, that helps a lot. I didn't realize that after a certain %a big change can occur. I only ran 2 gals of E85 once (to rule out false kr... it wasn't false, lol), and it didn't affect my trims much, maybe 4-5%. I'm leaning towards 3 gallons now, so it sounds like this will def be helpful.

whoa, how did I miss this thread. will have to catch up tonight

I just went from 2 gallons to 3ish gallons per tank and my trims skyrocketed to anywhere between +4 to+9. I was slightly worried but everyone's comments seem to confirm the same thing.

Did I mention how much I love E85, it makes Cali 91 octane cars actually fun to drive without constantly worrying about BAT's and KR all day.

that does seem to be the tipping point for LTFT to shift but OL can go higher before AF shift is seen. @cld12pk2go first ran into this back in the spring with ramping up his E85 concentration. He saw LTFT shift way before he saw AF shift at WOT. I too experienced the same an in fact I am not scaled yet on E35. IIRC, my LTFT is +10ish but I am not too worried about CL trims ATM. My WOT AF is about .1 leaner so I have seen no reason to rescale my entire map for that. I am glad cobb is on this scalar before I make my jump to 5050 so I won't need to scale all my load related tables in the spring when I make the switch.



Perhaps @David@COBB would care to speculate on the lack of AF shift at WOT?

On my third day using my original MAF-scaled E85 tune and no DTCs.

I did notice the same +25 STFTs at extremely low throttle though. Perhaps it was always that way; never really payed close attention to STFTs before.

I'll go back to the COBB beta scaled map, do a MAF cal, and see of that helps.

EDIT: I think instead I'll take my E85 adjusted MAF scale, reduce all values by 40%, and then put it into my COBB beta map. My trims should theoretically be identical to what they are now.

Tapadatass