Choppy AFRs - What's causing this?

[phuz]

Trying to get the VE table smoothed out in SD mode, but I've been noticing these "blocky" steps of the AFR even through very smooth regions of the map.
I've been fighting issues of going back and forth on cells that sometimes read rich and other times lean, and I really honed in on this log today. At this one particular spot, you can see the AFR is 18.2, dips to 16.2, and then comes back to 18.4 all within 400-500ms, but the RPMs and MAP are in the exact same cell of the VE table. Zooming out, you can see this type of "blocky" AFR behavior all over the log.
Any idea what's causing this? It's probably what's making it take so long to get this VE table dialed in.

Untitled.png
16-apr-2023-vetuning-2.hpl
2001 Corvette_with_Procharger_2bar_VE_tuning.hpt

[RDF1]

any possible exhaust leaks?
Spark plugs all look good?

[kingtal0n]

I advise a referenced regulator
Something is wrong with the fuel injector size scale
The VE Table is way off, the numbers should not be over 100 usually
The step size in the VE table doesn't make sense. Numbers should flow smooth for the most part without a big step.
That is a sign of injector data being incorrect. The delay is wrong or something with a pulse adder is wrong.
Transient mg may need to be reduced.
Pulse adder may need reduced.
Try to limit the pulse adder type of features for low pulses.



I advise first fix the injector size, acquire a referenced regulator.

The lean spikes could be any number of issues but there is so much to fix first I didn't look too deeply.
MEchanically there may be some issue with fuel pressure or misfires to look into.

Use new plugs, cheap coppers, don't gap them too much. Stay out of boost. Make sure they all look clean.
Tune to 15.1:1 air fuel ratio around that for clean plugs and cruise idle.

[phuz]

any possible exhaust leaks?
Spark plugs all look good?

I just put it up on the lift the other day with it running to listen and tighten down the clamps on the exhaust, but even those were AFTER the wideband. I need to double check the header on the wideband side again, but if there is one, it certainly isn't audible. Haven't checked the plugs, but they're all brand new with about 500 miles on them.

I advise a referenced regulator
Something is wrong with the fuel injector size scale
The VE Table is way off, the numbers should not be over 100 usually
The step size in the VE table doesn't make sense. Numbers should flow smooth for the most part without a big step.
That is a sign of injector data being incorrect. The delay is wrong or something with a pulse adder is wrong.
Transient mg may need to be reduced.
Pulse adder may need reduced.
Try to limit the pulse adder type of features for low pulses.

I advise first fix the injector size, acquire a referenced regulator.

The lean spikes could be any number of issues but there is so much to fix first I didn't look too deeply.
MEchanically there may be some issue with fuel pressure or misfires to look into.

Use new plugs, cheap coppers, don't gap them too much. Stay out of boost. Make sure they all look clean.
Tune to 15.1:1 air fuel ratio around that for clean plugs and cruise idle.

The injector data came right from FIC for these 60# Siemens injectors.
Ignoring the section of the map with the VE over 100 right now. I set them high and haven't tuned that part yet.
I don't think there's an issue with pulse adder, at least with this scenario. The pulse width (logged PID) doesn't jump around in this "cruise" condition where the AFR is acting funny.

I'm thinking I should fix the wideband orientation, too. The previous owner installed it on an angle vs perpendicular to the pipe.

I also wonder how much of this is the cam overlap, which I'm seeing can play trouble with AFR readings at idle and light cruise.

[Alvin]

Trying to get the VE table smoothed out in SD mode, but I've been noticing these "blocky" steps of the AFR even through very smooth regions of the map.
I've been fighting issues of going back and forth on cells that sometimes read rich and other times lean, and I really honed in on this log today. At this one particular spot, you can see the AFR is 18.2, dips to 16.2, and then comes back to 18.4 all within 400-500ms, but the RPMs and MAP are in the exact same cell of the VE table. Zooming out, you can see this type of "blocky" AFR behavior all over the log.
Any idea what's causing this? It's probably what's making it take so long to get this VE table dialed in.

Untitled.png
16-apr-2023-vetuning-2.hpl
2001 Corvette_with_Procharger_2bar_VE_tuning.hpt

Looking at the log.. this wideband doesn't look like its behaving very normally at all. Have you verified what it reads in HPtuners is the same as what is displayed on the gauge? Add narrow band O2's to the log just so we can do a sanity check.


Looking at the tune file... The VE table looks like it needs work. It could be that you are just that far off.

[phuz]

Looking at the log.. this wideband doesn't look like its behaving very normally at all. Have you verified what it reads in HPtuners is the same as what is displayed on the gauge? Add narrow band O2's to the log just so we can do a sanity check.


Looking at the tune file... The VE table looks like it needs work. It could be that you are just that far off.

The VE table was much smoother, but I keep adjusting based on the wideband feedback. The signal coming into HP tuners matches the gauge identically. I'll add the narrow bands and should be able to get out for some logging in a few hours when the roads dry up.

[phuz]

@Alvin, I've added the narrowband voltages and they don't quite seem to match what I'm seeing on the wideband. Look specifically at the area of the log from 1:22 to 1:36. Narrowband seems extremely steady while the wideband is dancing quite a bit.

But then look between 0:28 and 0:47 and you can see more inconsistency. I think the strangest part (in that same region) is while I'm at 2000rpm, 34mph, 18% throttle, both narrowbands are showing almost 0 volts while the wideband climbs up to over 18, and then drifts back down to 14.6 in that same region.

17-apr-2023-vetuning-4.hpl

[Alvin]

I'm leaning towards the VE table being off.

The NB and WB correlate to me decently well. 920 mv + is typically 11.0 and lower and 50 mv is usually leaner than 15.5:1 This matches up.


On top of that.. As someone who does this for a living. That VE table does not have the shape I would expect at all for the mods.

[phuz]

@Alvin
I agree the table needs a lot of work, but when I had it in a much "smoother" state, it would still behave like this. I saw a behavior yesterday that I was unfortunately able to capture because VCM Scanner crashed and the log wasn't saved, but the AFRs went rich-lean-rich-lean-rich-lean, etc. very rapidly while I was in boost. I glanced down at the graph and saw a giant zig-zag pattern under both wideband and narrowband, yet that part of the map was fairly smooth.

Fuel Injector Connection told me to call them this morning because I want to thoroughly confirm all my injector settings with them. The data sheet provided by A&A Corvette was extremely different from what FIC provided me last week at my request. FIC said they had an updated version and I expected minor tweaks here and there, but it was substantially different under the voltage/MAP section.
I'll plug those in later and go out for some logs.

[Alvin]

I don't mean how smooth the VE table it is but rather it's shape.

There will be dips, peaks, valleys, etc.

[phuz]

I don't mean how smooth the VE table it is but rather it's shape.

There will be dips, peaks, valleys, etc.

Since you do this for a living, do you have an example of a typical VE profile for a supercharged & cammed engine? I'm curious what I should really be targetting because I'm making small adjustments in cells that have a ton of hits and then all of the sudden I'll see the AFR swing the other way. Definitely shouldn't be doing that.

[Alvin]

Since you do this for a living, do you have an example of a typical VE profile for a supercharged & cammed engine? I'm curious what I should really be targetting because I'm making small adjustments in cells that have a ton of hits and then all of the sudden I'll see the AFR swing the other way. Definitely shouldn't be doing that.

Sure.. Here is one..

There is still a lot of error when you plot your AFR error / Fuel trim error. You will get there.. just keep plugging away at it until the trims/error is less than +/- 5% if not much better (3%)

Capture.PNG

[phuz]

Sure.. Here is one..

There is still a lot of error when you plot your AFR error / Fuel trim error. You will get there.. just keep plugging away at it until the trims/error is less than +/- 5% if not much better (3%)

Capture.PNG

Thanks!
I'm doing all of this without trims enabled. I will re-enable them when I'm done.