Hi folks.

First off, don't worry... this isn't another one of those "why don't my fuel map AFR numbers match my actual wideband readings" threads.


What I would like to know is, why does my map seem to act so differently to examples I've seen from other peeps' vehicles? Why is there such a sudden AFR variation in a small rpm range around 4000rpm? Time to explain... background first...


I booked some track time the other day so I could start getting into AFR tuning. My plan was to:

- Disable lean-spool
- Block-set all map cells over 120 load to 11.5
- Put down some test runs along the track straight, WOT from 2000rpm to redline 3rd gear (4th too, distance permitting!).
- Feed the data back into the map, based on the well-known "desired, actual, map" ratio calc.
- Reflash, rinse, repeat.

Aim was to get 11.6-11.7 from peak torque to 5500/6000, enriched to 11.2 at redline.


This all went like clockwork - my last runs nailed the required AFRs.

However, I was pretty surprised to find such a massive variation in required map settings. All the real-world tuned examples I'd seen for this motor had nice, progressive changes in their High Octane Fuel Map... but my tuned map cells looked like a cliff face around 4000rpm to 4500rpm.

Here's some data based on the block-set-11.5 "calibration" run...

I can't really get my brain around why, in the space of 500rpm (3750 to 4250), my AFRs go rich by nearly a full point based on a flat calibration map?



Here is my final run tuning... taking the above data and reworking the map AFR to hit actual targets...

This is a Ralliart... but don't hold that against me.


So... why the cliff-face AFR map?

I've read about the fuel pump getting switched from "lower voltage" to "full" at some point (the good ol' relay circuit thing). Is this relevant? When does that occur? What is responsible for fuel pressure stability? My fuel trims all hover around zero, so there's no apparent lean running with the fuel pump at "low voltage".

Why don't I see evidence of the same shape in the various other example fuel maps I've seen posted up? No other "lean spool off" fuel maps from X or RA owners seem to have severely "scooped-out" values around the 3500/high-load region.


My next job is to flesh out a full map from the collected data, which is a biggie. Still got lots to learn. But before I spend countless hours on that, I'd just like some opinions from those far more experienced at this stuff. Is this all just down to what is right "for my specific vehicle", or something more?

Rich


PS. Lean-spool was definitely off. I set the relevant ECU options bit to 0... and I also set the Lean Spool AFR v FuelMap AFR table to like-for-like mappings, eg. 11.8->11.8 etc... because I'm a paranoid kind of guy!

 

Because the Volumetric Efficiency map is not linear... and adjust the fuel injectors to spray more fuel in those areas of the map.

 

Do you have a different intake? Did you scale your MAF to that intake?

Have you looked at the "Calibration Fuel Map"?


I'm with Clipse on this, it's something to do with your MAF or MAP tables. In theory, you could have tuned those first with your blocked out map, then you would be running a flat 11.5 across the board. Then you could easily just set your fuel map to the values you want and call it a day. The method you chose is easier and quicker.

 

Yep... Fuel Calibration Map / Unknown Fuel map.

 

No to each of those.

It's the stock intake.

And I don't think my XML def currently has a Calibration Fuel Map. Still, as long as the fuel map changes I make are applied as consistently in the "steep" range as they are everywhere else, I'm totally happy.

If I start getting accuracy issues because of the rapid changes around 4000, maybe I'll do more digging and look to find/rescale the above map!

Thanks very much...

Rich

 

It's basically because the VE of your tune doesn't match the VE of the original tune. That may sound strange, but MIVEC plays a massive part in VE..

I suspect if you took the time to fiddle with the 3x MAP -> Load tables you could get the fuel map to look more "correct".

 

ROM id?

I can find it if I have a copy of your rom id.

 

its probably 53610010.

but the Calibration Fuel Map probably isn't the best thing to change

 

Yes indeed, Bryan's GST tune does include significant MIVEC reworking.

And yes, my Oz RA ROM is 53610010. But please don't spend any time hunting down maps that I'd really be better off avoiding atm. Currently, my poor brain is at its limit taking the GST "factory style lean spool" fuelling and adjusting it for Merlin-style lean-spool, plus my measured AFR amendments. Revised VE tables would add another migraine to the two I have already. Maybe later!

However, if you guru disassembler lads want a bit of a mystery to sink your teeth into, I've got a juicy one to do with ECU spark map choice during SST upshifts. I won't go spinning with it here. If you're interested, can I pm/email you?

Thanks for the fuelling input, in any case. Much appreciated.

Rich

 

If this is one of the "3xMAP/MAF tables", it's (now) in my XML def...



The tables weren't in my xml, but were present in the file on your central storage site, Golden.

Additionally, I was missing MAF Scaling parts 1-3 - also sorted now.

Thanks a million for putting that resource together.

Rich

 

yeh, so look at the 21.53 column, you can see how Mitsu designed this table with various VE's of the stock motor/tune in mind.

change MIVEC and it all goes out the window - however luckily for us its not a big deal, we can just add/subtract more fuel from the table..

its also why timing doesn't seem to go up in the right amount, but rather taper off the higher the RPM..

 

So it looks like this says based on rpm and psi (boost) look to load cell "x" of that same rpm on the Fuel Table, as well as Timing.
How can I also get this table,
Rom ID: 55590006

Thanks,
Steve

 

I'll add them to my site later tonight.

 

UPDATE:


Pretty sure my supposed "skewing" is in fact coming from a significant lag between the ECU data and wideband data - which comes via serial link.

So my block-set 11.5 map results were not actually as follows:

The results above is just what you get when your stupid serial interface feeds you data 0.5-0.7 seconds after it happens!

This is what it looks like when manually corrected...

That makes sooo much more sense! The supposed "leaner" parts of the map weren't "leaner" at all - the WBO2 interface was giving me a delayed telecast of spool-up AFRs.


Note: I added a load column to the above. Given all the talk of VE differences, and the 3xMAP tables, I thought it was good value to show the variation between 3500rpm and 3750rpm.



To really illustrate how bad this data sync issue is, here is my throttle-off data from the same log...

The timestamps show a 0.7 difference between those two highlighted points. That kind of throws a hand-grenade into the map tracer.



I'm well aware of the "rear o2 sensor delete and splice" trick. I didn't want to do that unless there was a very good reason. Hello, very good reason.


So, in summary, I'm not going within a bull's roar of those 3xMAP/PSI/Load tables - because my underlying issue is with the data interface itself!


Once that area is clean and reliable, I'll see how my fuel tables turn out! I bet they will look less like a rollercoaster once they are based on time-synchronised data!


Rich

 

log IPW as well. then you can really correlate

I wouldn't be so sure its WB lag