Just strait up can't figure it out. $50 reward offered.

[Dudeland]

 

 

 

 

3

55 minutes ago, AustrianVespaGuy said:

Ok, here's what I find very suspicious, although I can't yet explain how/why it's causing your problem, just that I don't ever believe in coincidence:

At the EXACT moment that the RPM starts to drop (several times in the 0033 file), it's when your indicated AFR plot CROSSES the target AFR plot (12.6 AFR to be specific).  Now, causality is going to be trickier, but I feel this is a big clue.  So here's some experiments that you might want to try:

1.) Is closed-loop feedback on or off during these times? Is Sniper 'trying' to keep idle stable? If so, I'd start with disabling that and seeing what things do 'on their own.'

2.) RAISE and LOWER the AFR target, say to 13.0 and then 12.0, and see what happens.  Does idle still stay stable up to the 13.0 threshold before dropping, or drop sooner when if hits 12.0? If so, it's definitely in the electronics control and something is chasing its own tail.  If however, idle still drops at 12.6 AFR regardless of the target threshold, then it's truly an issue of the mixture being too lean for some reason, and next step would be to run a little richer at idle.

 

My guess is that when your AFR hits the target value, Sniper then 'does' something, or stops doing something, and whatever that is causes your idle to destabilize like that:

 

 

My only theory at this point is that unmetered fuel is entering the system via fuel pressure fluctuations.   

 

 

 

[Dudeland]

2 hours ago, flagoworld said:

So, it looks like the idle air valve ramping open is totally normal. It also looks like it is doing so at the right time. However, you can clearly see that it initially bogs down to around 300 rpm, then sits there sputtering for a second while wildly lean. It looks like the computer is trying to correct for the lean condition, and injector load and fuel delivery goes up a bit. And then it recovers and runs very rich for a couple seconds until the computer corrects again and reduces the fuel load. Am I seeing that right?

 

Almost definitely a legit lean condition in my mind. The data and the behavior points to it. Either a vacuum leak or a fuel delivery problem.... But the fuel flow, according to the data, does not seem to be lacking at all. Which leaves extra air getting sucked in somehow....

 

This only happens when pressing the brake, right? What happens if you use the handbrake instead of the pedal? Does it ever happen then?

This is my observation as well. It happens when I come to a somewhat abrupt stop.  I am not sure I can emulate it with a handbrake.   

I did check the booster, both the check valve and by sucking on the hose while someone pressed on the pedal.  There was no drop in the vacuum that I can detect (using my tongue sensor). 

 

Also pressing up and down on the pedal rapidly does not affect MAP or the idle. 

 

The only thing is that Actual fuel flow may change due to small fluctuations in fuel pressure. 

 

 

[Dudeland]

2 hours ago, jimk said:

Created my own graph and the fuel is not responding enough when the IAC opens the hold rpm.  Map begins to climb and AFR climbs at the same time, nothing unusual there that indicates a vacuum leak.

 I'd stuff about 20% more fuel at 800 and lower rpm in the fuel table and see what happens.  Since I can't bring up the fuel table, I'm only guessing the table has cells for 800 rpm and lower.

I may try fattening it up even more down low to see if that changes anything.   On a carb setup, the system returned to idle just fine at that 12.3 to 12.5 AFR range. 

[AustrianVespaGuy]

Can't remember if we already covered this, but what does the timing map in you 123 look like?  Do you have more or less advance down at say 700 RPM compared to idle at 1000 RPM?  You should have maybe 2 degrees more below idle to help keep the idle RPM stable.  If you instead have less advance down there, you indeed will have stability problems as the EFI tries to add more fuel and air to speed things up while the timing is being retarded. . .

[Dudeland]

Can't remember if we already covered this, but what does the timing map in you 123 look like?  Do you have more or less advance down at say 700 RPM compared to idle at 1000 RPM?  You should have maybe 2 degrees more below idle to help keep the idle RPM stable.  If you instead have less advance down there, you indeed will have stability problems as the EFI tries to add more fuel and air to speed things up while the timing is being retarded. . .

As per your advice ( I think) I put two more degrees in down low to create a bit of a through to help catch idle overuns


Sent from my iPhone using Tapatalk

[AustrianVespaGuy]

One thought along the lines of fuel pressure: MOST EFI fuel pressure regulators have a vacuum line going to them, so that the fuel pressure is always adjusted to compensate for manifold vacuum; is that the case on the Holley? It's not clear to me from their pictures that this is necessarily the case.  What this means is: if the FPR is set at 60psi-gauge, then there is 60psi drop across the injectors when your at WOT and the pressure in the manifold is a full 1 atmosphere.  But when you're idling at 33kpa (only about 1/3 atmosphere), if the FPR remains unchanged then the pressure drop across the injectors is now closer to 70psi (60psi + 2/3atm).  Clearly the Sniper system has some means of compensating for this, but if we can figure out how they do it, it might help clear up that part of the mystery.

Also, I see that you are indeed in closed-loop during idle; can you turn that off and test it out in open loop instead to see what happens (issue goes away or maybe engine just dies)?

Third, what is your Basic Idle target set at?

 

[jimk]

31 minutes ago, AustrianVespaGuy said:

if the FPR remains unchanged then the pressure drop across the injectors is now closer to 70psi (60psi + 2/3atm)

That tends to over fuel at low manifold pressure which is clearly not the case from the log.  The injection time is increasing (although slowly) during the excursion and the AFR is increasing and is contrary to over fueling.

From previous threads I am understanding the fuel tuning process is primarily the auto learn method.  Low rpm sites are rarely accessed and auto learning needs time in a cell to create a correction.  Also at low rpm the exhaust gas flow relatively slow so there will be a delay for the O2 control loop to provide feedback. It's hard to mathematically determine the correction under those operating conditions. Those cells just need some manual manipulation and I suggested above to stuff some fuel in those and see what the response is. 

Edited May 22, 2019 by jimk

[Dudeland]

2 hours ago, AustrianVespaGuy said:

One thought along the lines of fuel pressure: MOST EFI fuel pressure regulators have a vacuum line going to them, so that the fuel pressure is always adjusted to compensate for manifold vacuum; is that the case on the Holley? It's not clear to me from their pictures that this is necessarily the case.  What this means is: if the FPR is set at 60psi-gauge, then there is 60psi drop across the injectors when your at WOT and the pressure in the manifold is a full 1 atmosphere.  But when you're idling at 33kpa (only about 1/3 atmosphere), if the FPR remains unchanged then the pressure drop across the injectors is now closer to 70psi (60psi + 2/3atm).  Clearly the Sniper system has some means of compensating for this, but if we can figure out how they do it, it might help clear up that part of the mystery.

Also, I see that you are indeed in closed-loop during idle; can you turn that off and test it out in open loop instead to see what happens (issue goes away or maybe engine just dies)?

Third, what is your Basic Idle target set at?

 

 

1) The injectors on the TBI are on the atmospheric side (top side)  of the throttle plate, not the vacuum side (bottom side) of the throttle plate, so I am not sure if a vacuum referenced FPR is necessary in that case. 

 

2) The system will handle boost, so I suppose that internally in the TBI the FPR may be vacuum referenced.  I haven't taken a look at boost options,  it may require an additional vac referenced FPR to make it work under boost. 

 

3) I have tried open and closed loop to no effect, the same issue is happening. 

 

4) My target AFR at idle is 12.4-12.8 ( I have tried a range).  As per your suggestion, I have created a pocket around idle so that small +- 200 rpm changes in idle don't result in variances in the idle target  AFR.  My rpm is set at 950.   I do have a lightened flywheel (nothing crazy).  I think Rick said it was about 13.5-14 lb-ish, although he didn't weigh it specifically. 

 

[Dudeland]

2 hours ago, jimk said:

That tends to over fuel at low manifold pressure which is clearly not the case from the log.  The injection time is increasing (although slowly) during the excursion and the AFR is increasing and is contrary to over fueling.

From previous threads I am understanding the fuel tuning process is primarily the auto learn method.  Low rpm sites are rarely accessed and auto learning needs time in a cell to create a correction.  Also at low rpm the exhaust gas flow relatively slow so there will be a delay for the O2 control loop to provide feedback. It's hard to mathematically determine the correction under those operating conditions. Those cells just need some manual manipulation and I suggested above to stuff some fuel in those and see what the response is. 

I have often thought of exactly this.   Holley states that it needs at least 6" past the collector for the 02 to work correctly, and that is about what my I.E try y gives me. 

 

I have played with the fuel at those low RPM's and found that reducing the V.E to 70ish ( very lean)  causes the idle to bounce up and down rapidly.   Not sure why, but through trial and error I have gotten to the point where the idle is stable once achieved.  This occurs naturally through the self-tune process or I can manipulate it manually (around 92, to 98 VE ratio it seems happy)   

[tzei]

14 hours ago, Dudeland said:

PW refers to the pulse width of the injector, how long it stays open.  Duty cycle refers to how often it opens.   Yes, you can change the minimum pulse width, some people have played with this, but I have kept the stock settings for now. 

This affects the base fuel map correct? Is the main fuel map based on TPS or MAP values?

What parameters there are that affects idle & you can change?

[AustrianVespaGuy]

7 hours ago, Dudeland said:

The injectors on the TBI are on the atmospheric side (top side)  of the throttle plate

Ah, right, I'd forgotten that these were TB injectors, sorry!

 

10 hours ago, jimk said:

Those cells just need some manual manipulation and I suggested above to stuff some fuel in those and see what the response is.

I think Jimk's on the right track here, some more experimentation with idle fueling is probably the best next step.

 

7 hours ago, Dudeland said:

around 92, to 98 VE ratio it seems happy

This is very odd, VE should not be anywhere near that high at idle.  Considering your MAP at idle is around 33kPa, I would guess that your VE should be somewhere in the 30-45% range.  98% would mean that you're basically getting a completely full 500cc/cylinder intake charge of air/fuel, which with a NA motor at idle is absolutely not the case.  Are we sure we're talking about the same Volumetric Efficiency % parameter here? Not that this would explain your problem, just that it stands out to me as odd. 

[Dudeland]

8 hours ago, tzei said:

This affects the base fuel map correct? Is the main fuel map based on TPS or MAP values?

What parameters there are that affects idle & you can change?

The base fuel is adjusted to it meet the Target AFR.   The two things that it uses as a reference is TPS and MAP.  I can change the scale of both to create a more appropriate table for my application.  

 

I can really change anything you can imagine.   There is advanced functionality that I can build my own rules based on any imaginable parameter(s) or event(s) .   I will try to send a screen.   So, for instance, I can apply table x be it fuel or timing) when values A and/or B and/or C and/or D  between or above or below a certain value and/or condition. 

 

I have used it to apply more fuel immediately after I lift off the throttle (when TPSless than 2%), when the revs above 3000 rpm until the RPM reaches 1800 rpm (with the TPS still less than 2%), then return to the standard Target AFR table.  This did help to address lean condition right after liftoff but didn't address my problem of overrunning the idle.

 

 

This is something I may try to help once we can get a clear understanding of the root cause of my problem. 

 

 

 

Edited May 22, 2019 by Dudeland

[Dudeland]

 

 

1

4 hours ago, AustrianVespaGuy said:

This is very odd, VE should not be anywhere near that high at idle.  Considering your MAP at idle is around 33kPa, I would guess that your VE should be somewhere in the 30-45% range.  98% would mean that you're basically getting a completely full 500cc/cylinder intake charge of air/fuel, which with a NA motor at idle is absolutely not the case.  Are we sure we're talking about the same Volumetric Efficiency % parameter here? Not that this would explain your problem, just that it stands out to me as odd. 

I agree I don't really understand how Holley is using VE  ratio.   There is another map that you can use rather than VE, which I think refers to flow in LBS/hr based on RPM and MAP.   I will have to check and get back to you. 

 

 

[flagoworld]

So how does the system know how much fuel is flowing? Is there a fuel pressure sensor or is it just based on rated injector flow?

[jimk]

These VE numbers are derived numbers back calculated from what the software determined the pulse should be for the given displacement/rpm/manifold pressure and temperature/injector rated flow @ specified fuel pressure and injector dead time.

if say, the fuel pressure is lower than what the software expects it should be, then the calculated injector pulse width for whatever VE is in the table will be insufficient.  So the user (or the self learning feature) inputs a higher VE in the table (or the self learning correction table) so that enough fuel is injected to satisfy the A/F table setting as read by the O2 feedback loop.