New thread for Speed Density tuning?...

[Creamo3]

Originally Posted by mrfred

Yep. Don't remember where the thread is located though.

Speed Density airflow difference

[GST Motorsports]

Originally Posted by Creamo3

Speed Density airflow difference

Yup it would be interesting to see the boost / timing / AFR graphs from those overlays.

Both of the runs in my test where carbon copies of each other in all three areas.

It also looks like the MAF to No MAF was just a open turbo with no induction tube or filter, and the third pull was with a tweaked tune. I am not discounting aarons findings whatsoever, it would just be interesting to compare the data.

- Bryan

[Slo_crx1]

Originally Posted by 03whitegsr

Didn't JohnBradley do the same test and documented ~15HP gains by ditching the MAF on a mid 400HP car?

As for tuning on SD, has anybody else noticed the AFRs to be pretty inconsistent when running in open loop under cruise conditions?

3000 RPM, 70 kPa gave me 13.2:1 AFRs on the freeway on my way to work.
I get off the freeway, do a 2nd/3rd pull on a highway, and then drop back to the same 70kPa, 3000 RPM load condition and my AFRs are 14.2-14.4. A full point leaner for the exact same operating condition.

The only time I noticed a large variance was if my trims were a decent amount off and the front o2 over-corrected/under-corrected. At the same time, I also turned lean spool back on from my initial tuning where I had it disabled and it seems a little more consistent to where I want it.

[03whitegsr]

The car is running open loop, fuel trims are ignored in this situation. I noticed the AFRs were more consistent this afternoon though, maybe it was a fluke.

[rolly1818]

sorry for the noob question, i have SD on my car for a while now and its running great with no problems except for throttle hang (fixing this now).

can someone give me an explanation how do i modify the actual SD tables? guess i have been lucky thus far since i did not need to tune them for the stock turbo.

i am specifically looking at the SD MAP sensor VE and calibration table & SD RPM VE table.

[mchuang]

So what do you guys that run speed density do about barometric pressure change? Like in Cali, for example you can be at a certain elevation in LA, then head somewhere else in the mountains or hills and be a diff elevation.

[mrfred]

Originally Posted by mchuang

So what do you guys that run speed density do about barometric pressure change? Like in Cali, for example you can be at a certain elevation in LA, then head somewhere else in the mountains or hills and be a diff elevation.

SD does not require (or need) a baro correction.

[scheides]

Originally Posted by JohnBradley

I have been finding the ratio is definitely off from 1:1 like I had posted originally. The ratio is non-linear but I have it here for those that want to try it out:

MAPVE-

41-40
61-63.5
101-103
121-124
340-430

Kpa to load will be pretty close to spot on with the 4 bar, your fuel map will look pretty close as well.

These numbers work for me on my 3586, stock turbos, stock appearing turbos, 3582s, Evo3 16Gs, etc.

Originally Posted by 03whitegsr

Those values look like you have a "Map16bit" scaling setup around a JDM 3bar.

Does "Map16bit" in your xml file equate to:
1) toexpr="x/3"
or
2)toexpr="x*0.4072"

JohnBradley, would you mind sharing how these values translate to the standard 'off the shelf' XML in the v7 roms? Yes, I took the plunge....I'm using the jdm/EvoX 3.3 bar sensor, does tha matter?

If my math is right, comes out to:

(a few values above this, then):
33.7-40
50-63.5
82.7-103
99.3-124
278.3-430

I'm just starting to dive into this, so any pointers are welcome

[scheides]

Goodbye MAF

Spent a few hours converting over my ROM and trying out a few iterations based on some stuff I had seen posted online. Initial impressions: WOW! The car is so smooth. Idle is ROCK solid (albeit a bit high, gotta work on that, 1300rpm), and any stuttering/bucking I had while on the freeway before is GONE. The car is SO smooth now its amazing.

A few notes:
-It took a few iterations to get the car to start/idle correctly. It wanted to idle in the 11's and load was way high. I got one that would work ok, made a few trips around the block, and all seemed decent.
-initially I set my WGDC all to zero. No boost! I'd hit about 9-10psi and confirmed seeing AFR in the 10's under 'boost' before trying anything more. No alky on or anything yet. Baby steps.
-Made a few more iterations to the SD maps this AM on my way into work. Tried a bunch, with mixed results. Ended up going back to almost exactly what I had the night before (some JB numbers, some starting with what others had posted, some of my own tweaking to get it to run):

-turned up boost to about 19, just keep it simple. AFR in the 10's and high 9's (I used my old afr map from my tuned rom @ DB Performance) and while it is SMOOTH as silk this new rom has lean spool completely disabled, and i don't think my previous one did (my bad).
-Got all the way to work and got CEL for fuel trims. Oops, forgot to log those, even though I had been watching AFR all the way. low/mid are -12/-12, so now I need to try and figure out what to do to fix that.
-lastly, the car wants to buck if I leave from a stop w/o revving it up to ~2000rpm. It would do this before, but I'd give it more gas and away I'd go. Now, it looks like it is leaning out as rpms drop to ~500 and then I need to feather them back to life. Not sure what's causing this, but I'm sure my messed up fuel trims have something to do with it.

Anyone have any insight for me? I'm using my injector scaling/latancies I have been using for almost 2 years, FIC 1050's, I'm using the GM IAT and the EvoX MAP sensor.

[03whitegsr]

Is there a better way to scale for different MAP sensors?

From John Banks original thread,
SD - first test success

He says he is using a multiplier (r2) with a value of 0x73 so that when RPM VE and MAP VE are set to 100, and a 1:1 ratio, respectively, that the Load = MAP [kPa]. This value is setup based on the JDM MAP sensor.

When using the Omni 4-bar (or any other MAP sensor for that matter), you either have to set up your table like JohnBradley has:

Originally Posted by JohnBradley

I have been finding the ratio is definitely off from 1:1 like I had posted originally. The ratio is non-linear but I have it here for those that want to try it out:

MAPVE-

41-40
61-63.5
101-103
121-124
340-430

Kpa to load will be pretty close to spot on with the 4 bar, your fuel map will look pretty close as well.

These numbers work for me on my 3586, stock turbos, stock appearing turbos, 3582s, Evo3 16Gs, etc.

Or you can change your MAP VE scaling to reflect the different scaling of the MAP sensor. This essential makes the raw values in the table the same between the two methods, but it forces the user to edit the ECUflash definition for their MAP sensor.

Would it be possible to just setup a scaling for the r2 multiplier so that the MAP VE table is more of a VE table and less of a MAP sensor scaling + MAP VE table?

Based on the 0x73 value and the 3bar JDM sensor, the scaling fromECU ->ECUFlash would be 2.609*x and the input value for a 4 bar with zero offset would be 400/2.609 = 153 -> 0x99.

This would also potentially make it easy to create an offset value as well for MAP sensors with offsets.

[mrfred]

There is a scaling for the OMNI 4-bar in the second post of this thread:

Links to jcsbanks' Speed Density patches

Once the correct MAP scaling for the OMNI 4-bar is in the ROM.xml file, then the typical way to populate the values in the left hand column is to use values that cover the range of the sensor, e.g., 10 - 400 kPa for the OMNI 4-bar. The right hand side of the table is then the load values that correspond to a given MAP reading.

If the final value in the left column is less than the max value of the sensor, then the MAP-load correlation value for all MAP readings above the max value MAP in the table will be based on the final value in the table. For example, if the final values in the table are 340 kPa and 430 load, then the scaled ratio is 430/340 = 1.26 for all MAP values greater than or equal to 340 kPa. In this example, MAP VE tunability at kPa values above 340 is limited. One reason for not covering the full kPa range of the sensor might be because of a need for greater resolution. If that is the case, then I can revise the SD MAP cal+VE table to have more cells.

[03whitegsr]

Mrfred, looking at 96530006, I don't see this 0x73 scaling value. Any idea where it went?

The table doesn't actually scale 3:4 since the JDM 3bar reach 300kPa at a different voltage then the Omni 4-bar reaches 400kPa.

[mrfred]

Originally Posted by 03whitegsr

Mrfred, looking at 96530006, I don't see this 0x73 scaling value. Any idea where it went?

The table doesn't actually scale 3:4 since the JDM 3bar reach 300kPa at a different voltage then the Omni 4-bar reaches 400kPa.

The 0x73 was for handling RPM VE scaling within the code. In the current SD rendition, It has been replaced with equivalent math and a change in the ECUFlash RPM VE scaling definition. Have a look at some of the posts on this page:

SD - first test success

[scheides]

So much for 'waiting till this weekend' LOL!

The big improvement since yesterday is in the fuel trims, I have them dialed into right around 0% for LTFT mid. Also leaving from a stop the car would want to stumble and die, some logging showed that load jumped into the 80 range and AFR leaned out to 18:1. Tweaked the SD MAP VE map a bit and fixed that, stays in teh 14's now as I leave from a stop and no stumbling!

Here's where I'm at:


My LTFT Low is still at -12, but I haven't let the car idle a ton. While giving my buddy a ride the other day he was semi critical of the idle, which I had left around 1200-1250 rpm. I was super happy because of how it would not dip below that at all, and he was saying he thought it really should be lower for me to say its 'solid'...anyways, I made one tweak to idle rpms (haven't touched any of the ISC Stepper tables or anything yet!) and bam, the car idles at about 900rpm, maybe 925. Logs show 900-950-1000 but the tach needle just sits dead inbetween the 800rpm and 1000rpm tick marks. Kelford cams, ported head, ported everything else, not a single dip or stall, its absolutely mindblowing!

Going to review some of my WOT logs and see where timing is at, but boost is set at 19psi now and afr is in the high 10's on my wideband. Ready to get back up to 24psi and try and get dialed into my old tune. I have yet to touch AFR, timing, injector scaling or latency values aside from translating or copying them over from my old map.

Ooh, also, the car seems jittery while its warming up, and I've noticed that at 2000-2500 rpm while cruising sometimes it feels goofy to. AFRs are around 14.5-15 so need to see what else is going on to cause this. It is very subtle once warmed up.

[Slo_crx1]

Cold idle and trying to cruise when not fully warmed up does seem to run a bit edgy, a lot of breakup I've noticed until it warms up. I initially had a small amount of breaking up in the cruise areas that was really annoying to me, since it always seemed to be around the 25mph spot through my town lol. I majorly re-adjusted the Asynch vs TPSDelta table and it smoothed all of the minor breakup during warm cruising, AND seemed to smooth out most of the cold sputtering as well. Another plus was the initial throttle tip-in...I used to see afr's jump to around 16-17 when trying to take off, now I see a steady 12.9-13.6 afr on average for initial tip-in.




Next up I think I'm going to get rid of my mbc and try out the psi-based boost control since my mbc's starting to flutter with the cold weather and every time it flutters my bov pops open. Not fun at 28 pounds.


I'm really happy with how this threads starting to turn out...great job with the info guys, keep up the good work!