A 3.4 DOHC Build then... F40 Turbo (Page 31/57)
Fierobsessed NOV 14, 03:44 PM
I think I've found my issue. The injectors MAY be too big after all. I sent L0089 to my data stream which is the LSB for the PWM sent to the Injector driver. It's the last value that gets sent to the injector driver to tell the hardware exactly what to do.

It seems at idle, I was throwing out 18 counts, which was shooting 11.5:1 But then as I was lowering the idle VE table, slowly at first, I watched the number drop in decrements of 2. It would take around 4-5 whole % VE just to drop the counts a single decrement of 2. ONLY when this output value would change, I would see it affect the AFR's. This explains why the Idle VE table seemed to be numb to minor changes, but responsive to major ones. I could get the counts down to 12, and it would be shooting right around 13:1. But, the instant the value hits 10... the engine dies. I mean just plain cuts off. This may be a limitation of the ECM's hardware apparently. It's calculating BPW correctly, but it's subject to the resolution of the hardware then, the limitations of it as well.

This also explains why it just plain wouldn't idle in closed loop. It was hitting this same hardware wall when it was trimming back the AFR's as I have been trying to do.

I may need to look into a different injector management method. There is Async, and Quasi-Async. I'm not sure how that is going to play out. I may have to break out my 42.5's for the time being if I don't find a solution to my dilemma.

Bottom line, it does NOT like the 60 LB injectors. If I can't get it to shoot the right amount of fuel by one mean or another at an idle, then I will have to resort to changing injectors.
Joseph Upson NOV 14, 04:21 PM

quote
Originally posted by Fierobsessed:

I think I've found my issue. The injectors MAY be too big after all. I sent L0089 to my data stream which is the LSB for the PWM sent to the Injector driver. It's the last value that gets sent to the injector driver to tell the hardware exactly what to do.

It seems at idle, I was throwing out 18 counts, which was shooting 11.5:1 But then as I was lowering the idle VE table, slowly at first, I watched the number drop in decrements of 2. It would take around 4-5 whole % VE just to drop the counts a single decrement of 2. ONLY when this output value would change, I would see it affect the AFR's. This explains why the Idle VE table seemed to be numb to minor changes, but responsive to major ones. I could get the counts down to 12, and it would be shooting right around 13:1. But, the instant the value hits 10... the engine dies. I mean just plain cuts off. This may be a limitation of the ECM's hardware apparently. It's calculating BPW correctly, but it's subject to the resolution of the hardware then, the limitations of it as well.

This also explains why it just plain wouldn't idle in closed loop. It was hitting this same hardware wall when it was trimming back the AFR's as I have been trying to do.

I may need to look into a different injector management method. There is Async, and Quasi-Async. I'm not sure how that is going to play out. I may have to break out my 42.5's for the time being if I don't find a solution to my dilemma.

Bottom line, it does NOT like the 60 LB injectors. If I can't get it to shoot the right amount of fuel by one mean or another at an idle, then I will have to resort to changing injectors.



Yep, all of that. You may still be okay by adjusting the two tables I mentioned as there are tune specifics for a stock motor and smaller injectors at play not to mention all the other stuff that a computer program is probably calculating the appropriate values for. There is also an injector spec sheet and I imagine yours is the same as mine which has flow specs for fuel pressure as low as the 30s. In other words you may need to drop your fuel pressure if you can. I removed the regulator along with enough outlet tubing from an 8100 V8 fuel rail which is adjustable and set my static fuel pressure to 40 psi which drops a little lower with idle vacuum.

This is why I often question members who plan to use injectors that appear to be far more than what they need as a "just in case I" approach. It's not a big deal until you decide you want to run closedloop. There is a member planning to use 80s on a 3800 SC right now with practical performance goals and that's a bit much for an intended daily driver.

I believe you can also select for single pulse at idle as well which I believe is tied to QUASI which you can disable to possibly get the desired effect.

[This message has been edited by Joseph Upson (edited 11-14-2013).]

RobertISaar NOV 14, 04:23 PM
increments of two sounds like you're already in single-fire mode, since the BPW calc is done assuming double-fire until BPW threshold is tested for, then doubled or cleared if single-fire conditions are met....

also, d10 = .153mSec.... that seems awfully short, even for 60lb/hr.
Fierobsessed NOV 14, 07:10 PM
I guess I should really think about fuel pressure then! I'm not far from where I need to be, I just need that little bit less fuel, and of course a little breathing room. Realistically, I need something like 50 lb injectors for what I am doing. So a little less fuel pressure looks like my best option. I'm going to have to machine myself an adjustable, vacuum regulator. This sounds like a fun challenge!

[This message has been edited by Fierobsessed (edited 11-14-2013).]

ttt123 NOV 15, 12:49 PM
I just read through this entire thread and i hope I'm not the only one who feels like they were just neutered.... yup i'm not a man after seeing this work wow man just wow I'm going to go outside and kick my fiero now lol

needless to say you did an amazing job and cudos on your talents sir.

take care
brian
Will NOV 19, 09:56 AM

quote
Originally posted by Joseph Upson:

I had trouble at idle with the 60s so I can imagine what you're probably going through with even less displacement, especially if you are running a base fuel pressure in the stock 50 psi range. It's very strange but yes there is something about attempting to run stoich AFR at idle that the motors do not like and usually it manifests with a rough idle although the rpm maybe stable. You basically need to give it what it wants until you find the balance as there appears to be some fine tuning fuel delivery that I don't believe the wideband can pick up but you can feel the difference.



This is due to mismatched dead times among your injectors. Most injector suppliers only match gross flow rate, but don't try to match dead times. Thus, the minimum fuel delivery per pulse can vary by 100% of the group average across the set.

If you're operating at minimum pulse width, one injector might be delivering 50% of the group average and the next might be delivering 150%... so one cylinder's always going to be lean (and misfiring) and one cylinder's always going to be rich, even though the average is stoich. To get the lean cylinder up to the point that it no longer misfires, the overall average has to be noticeably rich.

Injector Dynamics supplies injectors with matched dead times, but they're not cheap.
Will NOV 19, 10:04 AM

quote
Originally posted by Fierobsessed:
It seems at idle, I was throwing out 18 counts, which was shooting 11.5:1 But then as I was lowering the idle VE table, slowly at first, I watched the number drop in decrements of 2. It would take around 4-5 whole % VE just to drop the counts a single decrement of 2. ONLY when this output value would change, I would see it affect the AFR's. This explains why the Idle VE table seemed to be numb to minor changes, but responsive to major ones. I could get the counts down to 12, and it would be shooting right around 13:1. But, the instant the value hits 10... the engine dies. I mean just plain cuts off. This may be a limitation of the ECM's hardware apparently. It's calculating BPW correctly, but it's subject to the resolution of the hardware then, the limitations of it as well.

This also explains why it just plain wouldn't idle in closed loop. It was hitting this same hardware wall when it was trimming back the AFR's as I have been trying to do.

I may need to look into a different injector management method. There is Async, and Quasi-Async. I'm not sure how that is going to play out. I may have to break out my 42.5's for the time being if I don't find a solution to my dilemma.

Bottom line, it does NOT like the 60 LB injectors. If I can't get it to shoot the right amount of fuel by one mean or another at an idle, then I will have to resort to changing injectors.



Definitely sounds like you're hitting your minimum injector on time.

Don't the batch fire computers fire all the injectors every revolution?
That's a horrible way to run large injectors.
First, that cuts the pulse width in half, which makes the problems with short pulse widths MUCH worse
Second--and compounding the first--there are two fuel delivery events per valve event, which means that the greater error in fuel delivery acts TWICE for every combustion event.
Joseph Upson NOV 19, 10:14 AM

quote
Originally posted by Will:
This is due to mismatched dead times among your injectors. Most injector suppliers only match gross flow rate, but don't try to match dead times. Thus, the minimum fuel delivery per pulse can vary by 100% of the group average across the set.

If you're operating at minimum pulse width, one injector might be delivering 50% of the group average and the next might be delivering 150%... so one cylinder's always going to be lean (and misfiring) and one cylinder's always going to be rich, even though the average is stoich. To get the lean cylinder up to the point that it no longer misfires, the overall average has to be noticeably rich.

Injector Dynamics supplies injectors with matched dead times, but they're not cheap.



That's why I threw the idea of trying to relative-match the values on the injector spec sheet out the window, although my problem was with idle being too rich such as the case here. On the other hand, the two tables I mentioned when used together did help produce a more practical idle AFR without roughness and there is a gray area where the inj BPW offset table makes an audible difference when changed that can't be seen in the AFR reading. Both tables are a bit ratchetty regarding resolution but seem to work together well. RobertSaar mentioned the proportional gains tables but I found better response from the latter two. I do recall mention of the prop tables being of more significant help with radical camshafts on the Thirdgen forum.


You can look at the inj spec sheet and other specific data here in the thumbnails just below the main picture. Note the offset curve as system voltage drops. My fuel pump voltage runs about 1 volt below system/battery voltage and I was able to follow it and adjust the offset table accordingly by watching the AFR along with voltage drop as the alternator heated up.
http://www.siemensdeka.com/...114961/#!prettyPhoto

[This message has been edited by Joseph Upson (edited 11-19-2013).]

Will NOV 19, 11:58 AM
Yes, you DEFINITELY need to program dead time versus voltage into your computer in order to have the ECM trying to pulse the injectors correctly.

If you have full data logging capability of an OBDII computer, figuring out which injectors are doing what could be easier, as you can watch the misfire counters. If you try to idle at stoich and you're seeing #2 and #5 misfire a lot, then you move all the injectors to the next higher cylinder number (and 6 -> 1), then see #3 and #6 misfire counters go up... then you know which injectors are well outside the group average.

You can also log injector pulse width per cylinder on a sequential engine and back out the intake manifold correction if your samples are close enough in time.

[This message has been edited by Will (edited 11-19-2013).]

Fierobsessed NOV 19, 09:42 PM
I'm not OBDII so that's not really something I can do much about.

In the meantime, I made an adjustable fuel pressure regulator... This was a fun challenge.
First, I found a spare fuel rail kicking around


This is the stock regulator


The obvious thing to do was to cut the top off, unconventional method, but effective!


Ok, so there is the spring, I need to make that adjustable.


Design time!


Chunk of stock


First I did the internal boring cuts


Then... I got so busy I didn't take many pictures.

Then This...




Assembled, I use a large snap ring to hold the two parts together.


Old vs new


Then, it was time to tear the upper intake manifold and the fuel rail off and put the custom regulator in


Regulator is in. Its quite a tight fit, a litte grinding was necessary to clear the lower intake manifold. But I knew it was going to be very tight.
I adjusted the unreferenced pressure to 34 PSI, put the intake manifold back on and went back to tuning. The pressure under idle vacuum was 28 PSI. That's pretty low. Theoretically it brings me to around 53 lb injectors.


I tried retuning it, and unfortunately its not enough. It will idle at 14.7:1 now which is great! but... it's still right at minimal pulse width. If I give it more timing like it wants, I have to pull some more fuel to idle, and it still dies. It looks like I'm going to have to break out the 42.5's weather or not I like it. I'm stuck with it for now. It's a bit disappointing. I'll up the fuel pressure back to 46, which was what the stock regulator was set at, run the 42.5's and run them to their maximum at whatever boost it will let me run. The only upside, is that I know this engine idles great at 14.7:1, but timing is a very conservative 16 degrees, the 3.4 DOHC wants more timing, and less fuel then this.