Rich Throttle Transistion

[Sawgunner]

Hello all,
Finally decided it was time to post a couple questions.

Car configuration:
1994 4door R33 GTS25T
11U0F Rb20DET ecu feature pack 1
3" turbo back exhaust with cat delete
20x12x3" bar/plate front mount intercooler with aluminum charge pipes (typlical setup)
Koyo Aluminum radiator
HPX-N2 MAFS in a custom z32 cloned aluminum housing from Performance MRP with honeycomb air straightener. Used VQ Function Generator scaled to 31% in 2.985" ID tube Has a fully stock intake/airbox assembly with apexi filter
Denso audi R8 coilpacks with new wiring specialties harness, BCPR6ES plugs at .035" Stock dwell tables
Fuel pump relay mod
Walbro 450 pump (F90000274)
Aeromotive A1000 fuel pressure regulator 43 PSI base pressure vac line off
Innovate LC2 wideband O2
New 300zx narrowband O2 sensor
93 Octane (R+M/2) Pump gas
Knock sensors are stubbed out inside ecu as I was having the typical timing retard outside of knock zone, false readings etc.
Donut style knock sensor with stereo plug and headphone amp for knock listening
Boost in controlled by a mixture of the oem boost control plumbing but with solenoid removed and a custom sized bleed orifice in the bleed line.
Runs 11.5-12.5 psi

On to the main issue.
I have the car running decently in cruise and Wot stages with a conservative timing map and afr's in the 11's. It is very drivable and generally runs well.
The problem is anytime there is throttle transition it goes very rich, into the low 10s, sometimes even into the low 9's. It happens with slow throttle changes it seems as well as everytime I do any hard acceleration. Sometimes it seems to "cut" when it does it as it may be a misfire but I'm not sure.
It has done this since the car was all stock but with nistune installed. Don't recall it on the original ecu, it never seemed to be a problem (besides the terrible timing table and rich afrs!). It really is obnoxious when in higher revs around 4-5k and getting on and off the throttle results in very rich and power cutting (it isn't "crispy").

When this occurs there is a huge injection time/Load TP spike/hump and the resulting rich mixture.

Previous manual boost controller would also cause the problem, maybe slightly more pronounced.

Some observations are that the throttle enrich "tracers" don't ever move, and changing those tables to anything but extreme amounts (like 0) seems to do anything at all. I have only changed the Throttle enrich 4 as that is the only one active in the problem zone. Is it possible there is a hardware issue?

Another side issue is the VCT seems a little finicky. It seems to work properly when it should which is good but sometimes get stuck on at idle and can only get unstuck with a blip of the throttle. Not sure if there is any ideas for that.

Ive included an old log as well as my most recent tune/log to show the problem.
In new log the timestamps to note are 41.282, 1:17.157, 1:46.453 It is pretty short though so you can probably check the whole thing

In old log look at 2:13.032, 8:31.735

Sorry for the long post just wanted to include as much info as possible.

Look forward to some feedback

Cheers, John

[Matt]

Thanks for the logs and timestamps etc. Not sure what size injectors you are using? We have not used the N2 sensor, only the N1

I can see that when you are reapplying throttle here around 41 that the MAF is still at a high point (1.4V) probably from the turbo still spinning and sucking through airflow, so when you reapply throttle it goes rich

Not much I can suggest except BOV fitted and make sure distance between the MAF sensor and turbo, or put it on the pressure side

I'm working on firmware update to limit fueling depending on throttle position. Worst case do a firmware upgrade when I've made that available

With the 1:17 time we see the fuel map go from values of 9 to values of 27 which is a massive jump, and that is why the green line (injection time) shows that spike, so smooth that part of the map out, since there is a big jump there. Similar issue at 1:46 time, with a jump from 7 to 25 cell causing the spike in fueling

Those throttle enrich tables only really make a big difference from what I have seen

[Sawgunner]

Thank you for the reply Matt,
I am using stock 370cc injectors.

From what I gather the HPX-N2 is just an HPX-N1 in nissan clothing, based on the PMAS VQ function tool. i have also used the R35 mafs with no improvement, the hpx ran better overall.

I don't quite understand the fuel table jumps as I am basically working off of stock maps, 99% of the "left side" of the maps are unchanged, most work as you know is in the right side with fuel being taken out to get back to a reasonable afr. The 9's in the map are all from the stock map and the afrs in that region are a healthy 13.5:1, not sure how to fix that? The big jumps are due taking a couple of O2 flags out due to my load scaling with the map resulting in hanging onto closed loop a little more into boost than I was comfortable with (I was shifted 1-2 columns left with my maf settings and K constant). Does omitting that o2 flag and dropping a "similar" value in that cause a problem? Either way it's the same issue has been there regardless of those settings

Also stock BOV and intake setup are fitted. I was having this issue with the bone stock car with nistune added. I am hoping to get it resolved because I have a lot of faith in nistune and would love to keep it. No other red flags with the tune file that you could see?

I would think you would hear about it more but is it possible that the rb setup/plumbing and turbo size it can be metering far to much air than the engine can handle initially? Like it needs to catch up with the airflow coming in, but the computer is calculating the load for right now?

Not sure where to go. Am I putting too much faith in the stock map with my upgrades and new mafs? I thought i could get the fuel good before needing the dyno. I also thought that running the stock maps with a basically stock engine would run pretty well. Didnt realize skylines were so finicky! my sr with full bolts ons on the stock computer ran awesome!

Thank you for any further insight

Regards, John

[mtnickel]

I have a similar setup with similar problems:
RB25Det
Z32 ECU Feature Pack 1
Z32 Maf
650cc injectors

I have the throttle enrichment table totally disabled (255). If i stab at the throttle, AFR will dip to low 10's (sometimes high 9). It doesn't appear to be a faux spike in TP either. I will have to log it. But if i pass the same cells in a steady state, the AFR is much more normal. I wonder if there is a part of the code that adds an extra bit of fuel pulse based on throttle speed. Like a digital accelerator pump (as seen in Carbs). And it's a fixed ms amount that doesn't make nice with big injectors.

I also can't seem to take a rich warmup condition even with warmup table at 0. Once warm, it's fine. It cruises in the 11's for a while.

While I've been out of it for a while (Computer science major in 2006), I can follow assembly code to a certain extent. Any collaboration opportunities to investigate into this? Perhaps have a semi-decompiled code of the program with a few labeled variables? I will spend the time to trace which subroutines are changing the fuel pulse. Might be out of my league, but happy to help if possible. Also wonder if its better with different base codes. Don't remember having this issue with my 450cc injectors and K-constant adjusted tune...but that wasn't on a feature pack rom.

[PL]

I'm not in the office at present so I can't check the .bin. But I had an interesting tune last night with a similar issue. R32 ECU for R33. It was bogging rich on fast throttle appplication at low rpm. Finished up adjusting the TPmin table by lowering all the values below 2500rpm. Down to about 20 from memory. This limits how much fuel will be supplied in that RPM range. Without this you'll get super rich mixtures during rapid throttle application. Mine had all 255's in it (not sure why) but high values in the area below 2500 will certainly result in overly rich mixtures at throttle tip-in. Idle and full load were perfect.

Pete L

[mtnickel]

Here's a log of mine:

TipInIssue.png

(383.53 KiB) Downloaded 1398 times

If I sweep same fuel area but TPS not changing, then AFR is fine. I've also discovered it seems like its an additive fuel that decays over time.
Also, I made the last column 10 to see if lowering that column would reduce fuel. I'm wondering if my only solution here is to not use the injection multiplier and revert to changing K constant.
Log was too big as I logged all the way home:
https://www.dropbox.com/s/az7z21rdgj52h ... 1.csv?dl=0

I was going to create a new thread, but it seems relevant enough...My cold start enrichments also don't seem to be working. But in hindsight, I feel its possible that the enrichment being added above is just being exacerbated when it's cold (like it adds more or a longer Tip-in while cold). Look how long it takes the injection pulse to decay to get AFR back in line:

RichColdStart.png

(366.03 KiB) Downloaded 1397 times

[mtnickel]

one more with the cold start tip-in enrichment:

RichColdStart2.png

(492.77 KiB) Downloaded 1394 times

[Matt]

Had a look at the tune and I notice that K is stock, but scales changed from ... 96 to 115. What you could do is deduce K constant by about 20% to lower the load, increase TIM by 20% to offset fueling change, and keep the factory load scales

Maybe then it wont have such an effect on your temperature enrichment, which I've noticed you've 0 from 30degC upwards compared to factory

I was interested in what you said with the TPS and tried it a few times. It only enriches about 2ms with about a 2 second decay. See second screenshot

Keep in mind that O2 trims might be causing adjustment to your logs. Check the STFT/LTFT percentages. I tried to keep mine stable on the bench

Having TPS go low for a long time would cause a fuel cut

Temperature enrichment only normally works under 80degC. I've changed cells 20-40 degC and logged this. You can see the changes in purple in the log

[mtnickel]

Had a look at the tune and I notice that K is stock, but scales changed from ... 96 to 115. What you could do is deduce K constant by about 20% to lower the load, increase TIM by 20% to offset fueling change, and keep the factory load scales

I opted to keep K stock so that I wouldn't have to adjust all the TP dependant values (load cut, minTTP, maxTTP, VTC dependant values, etc.). I figured that was the point of the TIM. I'm actually not even using load columns up to 115 and could change it back to stock scale (running low boost for a while). In that case, it wouldn't have any effect I gather. Does the code use information about what column you are in? like reference column 6...i figured it used the Load value only. I'm happy to drop K const and really hope it also affects the enrichment time (and the enrichment isn't hard coded as a MS value).

Maybe then it wont have such an effect on your temperature enrichment, which I've noticed you've 0 from 30degC upwards compared to factory

I dropped it to 0 in an effort to lower the fuel.

I was interested in what you said with the TPS and tried it a few times. It only enriches about 2ms with about a 2 second decay. See second screenshot

I wonder if this decay and amount is LONGER when coolant temp is low...that might explain really rich enrichment.

Keep in mind that O2 trims might be causing adjustment to your logs. Check the STFT/LTFT percentages. I tried to keep mine stable on the bench

I worked a while to get STFT/LTFT bang on. Made sure inj latency was correct, etc. LTFT are less than 4 everywhere, and STFT vary a bit, but don't exceed 7 or so.

Temperature enrichment only normally works under 80degC. I've changed cells 20-40 degC and logged this. You can see the changes in purple in the log

I'll run this test at a steady state and make sure it's working appropriately.

I had one other thought. I don't have a speed sensor connected. I wondered if the ECU had a different tip in routine for when in park/neutral or at stop. Sometimes enrichment times are also revolution dependant (like it will enrich for 300 engine cycles). If I am in gear, it could be adding much more enrichment since the RPM aren't changing. Just thought it's possible the ECU has special enrichment for idle condition to make sure throttle is crisp off the line...but less when in gear since risk of bogging isn't there.
I've got a GPS unit I am going to connect shortly (car is 260z and doesn't have electronic speedo). GPS is scaled wrong however and the IDLE valve would keep high idle. Soon I'll have it rescaled through an arduino (which also runs a sequential shift light and exhaust sport valve modes).
Be interesting to see if that makes a difference.

I've got IDA pro running and just getting feet wet into tracing the code. not great with scripts and creating Symbol files, so just adding stuff manually. nistune ADR based values (constants)
CPU addresses (entry points etc)
Have basic Ram address (port 6 etc), but trying to figure out where all the variables are (ADC inputs, RPM, etc).

[Matt]

I can give you the consult table if you email me

[Matt]

Does the code use information about what column you are in

The code just looks at the current load and reads out the value of the matching cell for each type of table

Includes:
- Fuel
- Timing
- Knock feedback
- Throttle enrich
- VCT
- O2 feedback

Among other things. So if you stay in the factory load ranges, the reference points for the above dont get changed. If you get more load (increase boost) or less load reading (larger MAF) then it affects the above. We lower K to get back in those ranges and if it affects fueling, adjust TIM to get fueling back to stoich at cruise

I wonder if this decay and amount is LONGER when coolant temp is low...that might explain really rich enrichment

Would need to check again on the bench, but I would doubt it?

I had one other thought. I don't have a speed sensor connected. I wondered if the ECU had a different tip in routine for when in park/neutral or at stop.

Another variable I would need to check

[mtnickel]

Few tests ran:
1) Tried all the Sync acceleration tables and none seemed to do anything to tip in (Temp based, RPM based...all had no effect).
2) I was also going to try the "warm-up enrichment" as you did where you changed mid log and observed changes. I however was unable to disable closed-loop. I changed the Temperature activation threshold , as well as lowering the RPM activation to 1000 max; neither turned off closed loop, so I couldn't check if warmup enrichment was working. Wondering if I corrupted my bin or something? Base image bad? It does work to disable closed loop using the flag settings, but my map has a few negative values which can't be toggled. Might have to just raise entire map 3-4 points to get the negatives out. Is the above (not actually disabling closed loop) something to be concerned about? I may pull out my ostrich which i used to use and try an older or different base image (though then can't use Feature Pack).
3) Tried running much lower K const, and high injection multiplier. Had no effect on enrichment or driving.
4) Overall, driveability is perfectly fine. My only concerns are 1) really rich warmup driving - 10AFR during light acceleration and any throttle application, which may be washing cylinder walls with too much fuel? and a very very slight hesitation when stabbing below 2000rpm when warm (AFR dips to high 9 for very short time). Everywhere else it's fine (even though AFR does dip to 10's momentarily, feels fine and crisp). If this is fine from an operational standpoint, I'd probably leave it. Might result in fowled plugs more often?
5) Finally, I did not some of my long term Trims were 7-9 ish. In the event that the learning also changes tip-in amount, I have increased TIM to get trims down maybe slightly negative. will see if this helps. Counterintuitive to make it richer, but if it learns to lean out, it may do so for other systems.

[Sawgunner]

Hello all again! Good to see some action to discuss these issues in here!

Mtnickel, you and I had exchanged emails a couple of months ago, I imagine you remember? Good to have you in the thread!

Anyway a lot of what you said is pretty similar to what I've been dealing with; mine runs very well, trims in cruise run in very low single digits, wot of course needs a tune but it is generally in the 11's. Pretty impressive for a huge maf and injectors (hpXn1 and id1050s!). Still having major issues with chasing closed loop into boost and the throttle enrich stumbling issue. I tried decreasing K 30% (310-217) and increasing TIM 30-35% (210-299) as per Matt's recommendation. Inserted stock fuel and ign load scale values. Had to move my map values over a little bit to the left but it was running pretty good, albeit with the exact same issues. Ive gone to extreme values with the throttle enrich 4 table to no avail. Interesting note; I can induce the medium rpm (4000) high load stumble, by going WOT, backing off to 75%~ and right back on quickly and it stumbles. And I believe we all know that at high throttle angles on a turbocharged engine that that does not equal 75% load, probably closer to 90% load that the engine sees which tells me it probably isn't an issue with the engine gulping too much air at that time, but an issue of the throttle movement telling the ecu to do something to the engine. Hope that makes some sort of sense. If the engine had a lean stumble, would it show up on the wideband?

Still at at loss with the engine holding closed loop. I know one thing about it, I can almost always induce it from a lazy shift from 4th to 5th gear at 3000-3500 rpm, with immediate full throttle application, watch the inj ms drop and the afr's climb! I have logged both STFT/LTFT during these events at 12%~. I even disabled my neutral switch to make sure it wasn't an issue with seeing neutral mid gear change.

Looking forward to more collaboration on the subjects.

Thanks for all the input from everyone.

-John

[mtnickel]

Post your current tune, and a log. Also let me know what TP you hit for just atmospheric pressure.

[mtnickel]

If your VTC is erratic, you could try the Z32 handling of it.

Set Recover RPM to be ABOVE the Cut point (in z32 it's 6100, but it doesn't matter...could be 7000 or whatever). In this way, it will operate based on load > load enable AND RPM less than Cut point.

Any logs of your ECU holding onto Closed loop? I'm not sure about you, but on my tune, closed loop operation at low load and high rpm is annoying...sorta surges. This could easily be because I don't have speed sensor connected. But I cut closed loop above 3600RPM...just drives smoother up there. Can't really tell what's going on without a log though.

Which injectors are you running? You first mentioned 370s, then 1050s! If 1050's, your latency looks awfully low. Did you purposely drop it in an effort to reduce tip in amount?

In other news, I'm considering attempting a hardware "hack" to see if I can remedy this issue. Wondering about putting a capacitor as a filtering device to really slow the TPS signal into the ECU...gotta draw it up...something that really smooths the rise of voltage but allows it to fall quite quickly (to allow responsive return to idle or overrun). will let you know what I come up with and if it's effective at all.