Custom Build , self tuned ...weird data.

[dominic_pderoy]

Hello speed people ,
lets get straight to the point.
The car :
Mazdaspeed3 2012 tech pack :
Forged motor (piston, rods, supertech valve(ex) and valve springs, ported head, MLX gasket )
Ford ST intake manifold with adaptor for port injection (not plugged in yet)
3.5 Map sensor (bosch)
4 ways boost solenoid
custom 4 inch intake
Custom garrett front end intercooler and piping
Custom turbo manifold twin scroll with a Borg warner S200SX-E 7670 (twin scroll turbine AR 0.82)
Dual 44mm wastegate. ((screamer pipe
Custom 3.5 inch downpipe wit a 5 inch race cat (100 cell)
custom 3 inch exhaust to the back

I had the WMI from AEM and while testing it the push lock failed... so i'm waiting for new stuff with 3an line and 2 race valve(meth injector).

I'm running 91 octane and want to get as much as I can from that and then engage the Meth stuff and port injection.

So the weird data .... I have been raising the load bit by bit but after a certain tune that slipped .... , I got a very good punch ((spike) of boost(map reads 43.7psi) at around 5000 RPM and then it comes back at 31psi at 6000 RPM. While that boost spike is happening, my AFR goes way lean (12.7) the requested is 10.5. the injector are at 87% (in the tune I put the max at 100% so not because of that), rail pressure is at 12MPA at the lowest spot while its lean.

A load limit table makes the boost bounce back , but that should not influence the AFR...

I did try to get those difference between AFR and change the maf Cal , but still does the same thing.

FYI , my tune now is stable, I was just wondering what happend to the last one.

I want to tune it myself, so no comment about I should get an etune please.

I did not save that tune , but i have a scratch version of it with only the load lower...

log with lean spot[4356].jpg (229.13 KiB) Viewed 2884 times

91 Octane tune NO METH for dominic v1.01.vtune

(88.7 KiB) Downloaded 80 times

[attachment=1]

Thanks

[mituc]

Hey Dominic, would you mind also posting the log? It's easier to look at the actual log rather than the picture. You can attach the .vtlog file and not the CSV.

[mituc]

At a first glance the load tables are way off and in contradiction with each other and the calibration for the EBCS us stock. Also the MAF calibration is way off for a 4". Looks like a 6" or more.
This tune can't work properly.

[dominic_pderoy]

Hi,
Thanks for the reply , this was an early tune an I'm learning, I do know lots of tables was off and as I said in the first post "I did not save that tune" the one I shared is similar with different loads... Also, the fact that my maf cal is off is because I calibrated it where there was a "lean spot" in the tune , so it made it "bigger" . I have a 4 inch custom intake, so it will be different from any "off the shelf " stuff.



I'll try to post the log in the next few days.

My tune now is good , I began to tune with meth and port.
I was just worried about that lean spot magically appearing again and wanted to know why . So far , it never occurred again.

[dominic_pderoy]

I just put the log (CSV) in this reply .

Thanks

[mituc]

No, something is very off with either the car or the new tune. One degree of timing at 6000rpm will not get you anywhere. It actually got from negative to 0 at like 5700rpm.
Also, the AFRs went pretty lean at 4300 up to like 5300rpm when the boost spiked to like 33PSI (!!!), and even though during that event the MAF voltage increased with 0.5V (which however in your tune this equates for an increase from 300-ish g/s to like 570) this reflected in the fuelling very slowly. I guess it's because the load remained stuck at 2.239 the whole time while it actually went well above 3 (33PSI with no aux fuelling at <5000rpm will hurt even built engines).

The tune you have is definitely wrong and I noticed that from the beginning and now looking at the log it's pretty clear, but what's not clear is if the car has some other issues as well (leaks, improperly connected or inverted lines, and so on). I guess this is something we'll find out once you will progress with the tune and clean it up.

Import this logging preset to your dashboard and use it to log from now on. The parameters you log don't tell the whole story:
https://www.dropbox.com/s/i0ndre5umzanz ... vtpre?dl=0

Before you go WOT again you have to sort out your boost control, load control, load calculations, MAF calibration, fuelling. Well, pretty much the whole tune.

P.S.: Since you're using the Bosch 3.5BAR sensor, did you also calibrate it correctly on your dashboard?

[dominic_pderoy]

Thanks for the answer ,
For the timing I put it very conservative until I reached a stable load then I began to play with and add very little at the time.
Lean AFR has been corrected after that pull with maf cal and new tune. It was actually one of the first time I could get to read higher g per sec.

Yeah that tune was wrong , No leaks and all the connections are right, the only things I could think about is maybe at that exhaust flow, the position of my o2 sensor could have cause a lean reading.(probably not, because since with MEth I got it to a much higher power and not lean reading).

Tune wise : since that happened I made a new tune from scratch going very slow to raise it bit by bit( I learned) , and got it to a 288whp (VDyno), with no boost spike , no lean spot ,no knock, not perfect obviously, but im getting there.

My 3.5 Map was updated in the dashbord since day 1 .

So with the answer you gave me (thanks BTW) that lean spot could be caused by the boost bouncing back and forth (not being well controlled) trying to reach the load, which causes boost spikes and because the raise in boost happens so quickly , the injectors react slow and cause a lean spot?

Any other theories ?

Thanks

[mituc]

So that lean spot was definitely because the boost (thus airflow) raised too fast way beyond the factory DI + your WMI to keep up. Since you confirmed (and I was almost sure that was the case) that you had everything calibrated properly on the dashboard it means that the actual readings were correct.
So either the WMI nozzles are to small or depending on where they are installed it's possible that the boost simply cornered that WMI flow on the intake manifold or hoses walls, temporarily.

I've also seen boost spikes caused by the WMI itself, if the temperatures are above a certain threshold sometimes the water and methanol evaporate (even under boost) which makes the volume of gas to increase in the intake manifold. Or, under high boost, the WMI droplets condense even more and then suddenly when they accumulate more energy they break up again and produce that increase in actual gas volume. However at least on my car I see this happening at relatively high boost (32-33PSI or more), when sustained for longer periods of time, during warm days. Otherwise if outside is not hot enough not even in 6th gear pulls from 130kmph (3000rpm) to 230+Kmph (this pull takes a while and the boost is almost all the time above 33PSI or even above 34PSI) I would not see this happening.

So since you say you corrected these issues you will have to see how things evolve.
My suggestion is to tune the car and response WITHOUT WMI at first, and then when everything is perfect on pump gas with the factory DI start working on the MAF calibration and timing from that point on.

[dominic_pderoy]

Thanks For your answer ,
My 91 tune is good now, making 250 whp with no knock and no lean spot. I'm quite happy for that .

I began to tune with methanol and obviously, got weird data again , probably cause I'm better at building things than tuning them ...

So as my first post says , my meth injection is controlled by a microsquirt, I have 2 650cc injectors.
I began to change the AFR target and compensate with the meth to reach the desired AFR of around 10.6 (I think thats OK for an AFR on meth...).

Then I read that the proper way to do it is to reduce MAF from around 3V to the end and compensate with the Aux fuel witch I did ...
I did many pulls , raising the RPM by little each time to try to dial the Aux fuel table in the minisquirt. In the last couple of pulls I got something that I hope ill get an answer for, I command lambda of 0.73(10.59 AFR) and suddenly, the command AFR goes to 14.7 during the pulls ... this shouldn't be affected by the maf as the AFR as WOT is dependent of the LOAD and RPM right ?
My MAP is a 3.5 bar (bosch) which can read up to 43 psi and since I got to 47.7 psi in that pull, could this cause to throw the desired AFR to 14.7 ?

P.S. my actual AFR is lean, and I know it, I'm raising bit by bit and raising or lowering my aux fuel after each pull.
P.S.2 I got wheel spin on the last pulls , I would be surprised if that would make my command AFR to go lean... but I thought I should mention it.

91US_95EU +METH v1.00.vtune

(88.7 KiB) Downloaded 74 times

Full-3MPSBL-Logging - 2023-06-18 15.54.48.csv

(6.37 KiB) Downloaded 91 times

[mituc]

With WMI you should command like 11.2AFR for now (until you are sure things are good, then lean it even more up to 11.9-12.1).
To make it work you will have to adjust your MAF calibration down. That's the reason why the desired AFR will go to stoich, because the ECU kind of shits itself when it sees AFRs that rich because of the extra fuel (WMI) which it doesn't expect.
Your MAF cal is already a total loss and way off for a 4", I would send you a MAF calibration for a 4" but with so many things off on your tune and car I have no idea where that will actually land, and being the one to blame for a broken engine is the last thing I want.

Anyway, from how it looks I'd multiply your MAF cal by 0.9 from 180g/s upwards, and until there from like 160g/s I'd gradually multiply it by a smoothed vector from 1.0 - 0.92 or something.
However this is not the right way to tune. You don't jump straight to WMI tuning when the MAF calibration is totally off even on pump gas, because the pump gas MAF cal will be the base of the WMI maf cal (as indicated above).

P.S.: your load tables are still out of sync, expect juddering and boost spikes.