Bumps in MAF curve. Injection timing?

[hjtrbo]

Hey all, had my tune dialled in nicely on a loose stall and have now swapped to a much tighter converter that can handle WOT lockup. Now I can lock up at 1000rpm and mash the throttle to get a fuel tune run in. What I'm noticing at 5850Hz (1600rpm) and 7350Hz (2550rpm) is I'm having these weird bumps in my MAF curve to get the wideband reading right. I did not have this issue on the old converter as the engine rpms were much higher at those MAF Hz as the converter was flashing.

Doing some reading and have other examples and the consensus was the injection timing. I already have it adjusted 30 degrees overall but am wondering if I should utilise the RPM table (currently all 0). Also tried adjusting dynamic enable up to 4000rpm and at 1800rpm. Both settings does the same thing so it's not that.

Wondering if someone could take a look?
6.2L LSA with pulley upgrade (10-11psi)
Harrop cold air intake
Cam card attached
1-7/8 Headers into twin 3" cat delete system
DW1200 injectors

Ignore a couple of dodgy things in the log. Still have transient / manifold volume on initial WOT tip in to sort out. Also note VVE was jacked up just as a test. The tune file supplied has my old VVE table which is close to spot on.

[blindsquirrel]

The matching humps in cylinder airmass are just a result of the humps in the MAF curve, right?

[hjtrbo]

Yes.

If you plot maf Hz, injector pulse width and dynamic air flow on a separate chart it paints a picture of what I'm trying to sort.
Once the maf bumps are fixed the cylinder air and torque bumps should also disappear.

[eXo3901]

Any chance it might be the fuel system trying to keep up and stabilize?

[eficalibrator]

Proper MAF curves do NOT have bumps. Period.

You MUST build the MAF curve based on steady state data where you have clean combustion, accurate injector data, and a reliable lambda reading. Transients are handled elsewhere (dyn_air/VE, or wall film models) if you have error on tip-in or during lockup.

[Jarrod.v]

Just bought Engine Management: Advance Tuning two days ago! Almost to chapter 7 lol

[hjtrbo]

Any chance it might be the fuel system trying to keep up and stabilize?

Delta fuel pressure is solid. You can see a small bump when the 2nd pump kicks in but other than that it looks mint.

Proper MAF curves do NOT have bumps. Period.

You MUST build the MAF curve based on steady state data where you have clean combustion, accurate injector data, and a reliable lambda reading. Transients are handled elsewhere (dyn_air/VE, or wall film models) if you have error on tip-in or during lockup.

Agree 100%. Only change has been the torque converter where now I am able to hit parts of the tune that I wasn't previously reaching.

Injector data is good. Copy and paste from DW1200 spreadsheet. Currently working though my flex sensor install and am up to E20 and so far injector data is solid. Fuel trims are the same as E0 and wideband WOT data is identical.

Thought I'd include this screen shot showing some things.
Good stretch of road, no bumps, level. Smooth rpm and speed.
Smooth MAP pressure
Ignoring tip in (still yet to tune this) VE was jacked up as a test. No change. MAF and DYN are close, but you can see the jacked up air flow at those 2 Hz spots that I've had to give it to get the wideband reading lower.
Still dialling in WOT fuelling. 3 more runs I should have it flat.
Fuel pressure consistent. 1:1 Boost referenced, 50psi (closer to 52psi) base pressure, staged 2 x Walbro 460 pumps in an internal to tank surge bucket
MAF HZ nice and linear, no ringing.
And the problem child, look at all that extra fuel I am having to inject to get the wideband down. The 2 bumps coincide with my MAF curve. Even with the small correction needed that I have loaded in for the next tune iteration it will not be enough to flatten the MAF curve.

Something is very wrong.... Would love to know what is going on? Reversion, standing waves, injection timing, scavenging. Shit cam?

[eXo3901]

I am speaking form little experience here. The change happens right when delta pressure changes in that screen shot. I see the same lambda pattern when my BAP kicks in at 3psi(about 7k hz), that is why I mention it. My MAF curve was trending to show a "hump" from ~7k hz to ~10.5k hz. Which is the range that my fuel pressure drops, BAP kicks on, then fuel pressure recovers/stabilizes slightly and starts dropping again around 10k hz. This uses an E40 with 2barOS, so as far as I know only the 0kpa flow rate column is used while in boost. A single flow value with an unknown(to the ECU) rail pressure drop says under-fueling to me. Adding air to the MAF may not cure it because the computer still is not aware that it needs to compensate for the rail pressure dropping. My not-perfect fix until I install a return system was to add a little air right before 7k hz and carry that change throughout the rest of the curve, curve fit in excel, rinse and repeat a few times.

In your case, and I am theorizing, with the old stall this did not presented itself because your 2nd pump was already on and rail pressure stabilized in the higher RPMs. It was all baked into the MAF.

As you said and have shown, there is not much else that is making noise in the system. So I was thinking the fueling or maybe a post-MAF airleak big enough to introduce more non-metered air but not enough to show a dip in the actual Hz reading. If the latter, I would think a SD only pull would not show this same issue.

[GHuggins]

Just something to try. Although I am a little confused - it's just a lsa running a touch more boost with a decent duration cam, cai and headers? It shouldn't have required that many changes I would think. Also where I believe your trying to be safe with certain things such as timing - that combined with the extra volume spike from the bap as previously commented on is actually what is causing a lot of your problems . That cam is going to cause some fueling headaches on the setup, but I don't know if injection timing per say is your whole issue.


EDIT - Forgot to put in what all I changed - injection timing, transients, reversion settings, lower rpm spark settings and smoothed the MAF out

[hjtrbo]

The change happens right when delta pressure changes in that screen shot. I see the same lambda pattern when my BAP kicks in at 3psi(about 7k hz), that is why I mention it.

Also where I believe your trying to be safe with certain things such as timing - that combined with the extra volume spike from the bap as previously commented on is actually what is causing a lot of your problems . That cam is going to cause some fueling headaches on the setup, but I don't know if injection timing per say is your whole issue.

Cheers lads. The hobbs switch pressure setpoint was changed during the initial shakedown from 2psi to 7psi as the 2nd pump was always clicking in and out all the time under moderate load. Damn PD blowers hey, always boosting!

As a quick test I can unplug the relay for the 2nd pump and just run it up to 3,500rpm and see what it looks like. Should that prove fruitful I can muck around with the hobbs switch pressure setpoint to see if there is a spot that the setup likes. Failing that then I will have to investigate a 'soft start' type device to ramp the 2nd pump into service over a short duration (e.g. 1 to 1.5 seconds ramp)

I do have a Radium fuel pulse damper fitted. The spring in that is quite stout. It is boost referenced so I will also remove the boost line which will help it to move the diaphragm a little more. My fuel supply setup is all PTFE braided hose and utilising the OEM hard lines (3/8"). The FPR is mounted down the back thereby retaining the OEM style dead head fuel rail system. There would also be no inherit dampening built into the fuel system. Side note, was reading today when researching this issue that a number of Subaru guys back in the day would coil up 3 feet of fuel hose in the engine bay to act as an additional fuel damper. A lot of positive reports for that hack was surprising!

Forgot to put in what all I changed - injection timing, transients, reversion settings, lower rpm spark settings and smoothed the MAF out

Thanks very much for changes. I am stuck on my round of shifts will test next week and report back.

Although I am a little confused - it's just a lsa running a touch more boost with a decent duration cam, cai and headers? It shouldn't have required that many changes I would think.

Yeah that's pretty much it, plus ported heads, lid spacer and interchiller system. Presume your referring to the tune? I like to tinker. Ignoring the MAF bumps for a moment, the car passes the car park driveability test and plays nice when loading up at low rpm I.e no bucking or horrible seat of the pants stuff going on when TCC is fully locked up lugging it at 37mph in 5th gear @ 1000rpm through town. Except for the MAF curve I am really happy with the car overall and am looking forward to getting it up to e85.

[GHuggins]

I did one really close to that build several years back. Only it had slightly less exhaust duration, little more boost and 850 injectors if memory serves. Your injectors are probably more of the issue than anything now that I'm running the numbers through my head. Your pw is probably pretty small. You might even want to lower your rail pressure to increase it out some more.

This one has more "correct" tables for you.

Increased transient areas going into boost - due to larger injectors / timing is closer to desired in lower rpms / injection timing is better for the injectors and your later desire to run E85 - yes injection timing is better to be slightly advanced for e85 even boosted / also apparently reversion settings didn't get applied in the previous like I thought - so fixed in this one

[hjtrbo]

I did one really close to that build several years back. Only it had slightly less exhaust duration, little more boost and 850 injectors if memory serves. Your injectors are probably more of the issue than anything now that I'm running the numbers through my head. Your pw is probably pretty small. You might even want to lower your rail pressure to increase it out some more.

This one has more "correct" tables for you.

Increased transient areas going into boost - due to larger injectors / timing is closer to desired in lower rpms / injection timing is better for the injectors and your later desire to run E85 - yes injection timing is better to be slightly advanced for e85 even boosted / also apparently reversion settings didn't get applied in the previous like I thought - so fixed in this one

Thanks again. Certainly not be a problem to go down to 43.5 psi base. Easy done.

I had 1000cc in there before the fuel system upgrade, but I had junk injector data for them and didn't want the headache of chasing my tail for 3 months while I got the offset and adder tables sorted out. The Injector Dynamics fuel calculator had the PW for 1000cc injectors getting up there at e85 so thought I would up to the next size to be safe. As it turns out the ID calculator looks like it has a little bit of fudge built into it...

[GHuggins]

Honestly 850's with a boost ref reg would have probably been a touch on the small side, but should have worked for your setup. Just wouldn't have worked when pushed much further with boost - 1050's would have been perfect, but wouldn't account for future engine builds and so on. I don't recall if the ID calculator accounts for a boost ref reg - don't remember it doing that anyway. AND trust me, I understand going in circles due to injectors. I've gotten pissed more than once at a cheapo individual not wanting to fork out another $200 for good injectors vs some cheapos and then having to spend days jacking up a tune to get all temps decently dialed in, but these individuals will spend a butt load on non beneficial power adders just because so and so said they needed it

[hjtrbo]

Argh, the weather might stifle my plan

[hjtrbo]

Ok, weather info is a bit deceiving, bloody sun came out. Got the old girl out for a quick run before school pick up.

GHiggins, I have made no changes that you have suggested yet, have only tweaked VE and MAF. I just wanted to demonstrate that fuelling 'overall' is moving closer before I start introducing your changes, which I am looking forward to BTW...

I'm definitely leaning towards to a reversion issue.

[GHuggins]

Those two bumps in your MAF curve are evident as overly rich spots and injector pw bumps in your pull pics.

[hjtrbo]

Yeah, it's looking that way isn't it. Leave it with me for another couple of runs. VVE correlation is getting very close now.

[hjtrbo]

MAF bumps are genuine. Chasing run to run variance to get fuelling any closer than it is now.

I tried VE reversion 0 and OEM. No change to the bumps but different effects for tip in. Thanks for sharing GHuggins.

Haven't tried injector timing, just incremental changes 1 at a time to see what's up. Only changes so far are fuel.

Should be able to get out Friday hopefully for another run to try injector timing changes.

[hjtrbo]

Things are getting better. It just seems to be a case of getting back to basics and get the VVE right and also recognising there is some transport delay that I need to take into account when making adjustments to the problem areas. My wideband is around 1.5m downstream from the heads exhaust outlet.

Here is the latest progress on the MAF curve. Still have work to do but it's definitely looking better and moving in the right direction. Not worried about rich on tip in yet so let me get away with that part for now please . It was a bump in my VVE table that I have since removed.

[GHuggins]

I honestly didn't consider the transport delay of the wideband, but that's why I like to use the eddy brake to maintain rpm points while increasing and decreasing load slowly to dial in MAF curves and VVE models. I understand not being able to do that on the road. SO are you just shifting the fuel corrections to the left a cell? Also since it will require re fuel tuning, if your going to use that injection timing, you might want to go ahead with it.