nAst1: Progress and Concepts Thread

[cavfan1]

Ok, that went rather well. I carried over most of my original tune to the BFUZ and it worked nicely. The goal of this was to fiddle around with the car some more, sure, but to also try and get rid of the realy annoying tip in stumble. I've tried to play with all the AE settings but just never got anything to work right. I could make a very quick shot as rich as I want, but never get it to last long enough to keep it from falling on its face. It would go from say 12:1 for a half second or less, to a spike of ~17:1 for around at second after usualy. Was almost like a tease. I know the VE tables being correct help, but even when they were nearly spot on, it would still do it. I also wanted to try and get some better mileage, and get another car running it for feed back.

Driving around now, most any tip in is met with a nice surge of throttle reponse! No falling on its face or anything. There is a deffinite switch over from single fire to double fire though. Its almost like it falls off a little, its soft, but just doens't seem to have the same umph after the switch to double fire. It takes a rather large throttle transition to hit it though, or the upper RPMs at WOT. I only did one WOT run through 3rd getting on the highway. Without a wide band, I'm not gonna push it much.

I have an idea why my original port didn't work. 2 tables, the double fire pulse width allow/dissalow scalars are way off. In the BFUZ port, they are 6.5 and 6.0, on mine they are 1.28 and .98. Other than that, I didn't see much that would cause a big issue. Watching my trims perodicaly while driving, they are around 5-8% rich for the most part, idle is nearly spot on, though the idle is still rather unsteady, with lots of stalling if I don't catch it. Not sure if its tuning yet, or an idle relearn er what. If people can check the logs and bins for insight, it would be much appreciated.

Overall I like whats been done, and I thank you Robbertissar for the work so far. Also thanks for the welcome, I've lurked here periodicaly for a while, some very good OBD1 info around.

Ok, XDLs aren't uploadable, so heres a couple:
This one is the first drive with the good bin...
Night drive 1

This is a shorter drive after playing with the follower real quick to see what would happen...
Night drive 2

As for the bins, 1.0 is the original port I tried using my tune. It might be useful to look at to figure out why it didn't run properly. 1.1 is what I was driving around on and works well.

I did have a 1.2, only changes were AE tables from my own tune. It was more curiosity on my part, to see what tip in would feel like on the new setup vs. old. It still didnt fall on its face, but didn't have the same umph feeling as the stock BFUZ settings, so I reverted back.

[robertisaar]

yeah, the 1.28/.98 values don't allow much if any room to operate in single-fire, the 6.5/6.0 covers all of the cruising range and then some for me.

since you're running a gen3, change the idle spark advance tables as follows:

idle spark vs coolant
0.839
0.918
1.000
1.000
1.000
1.000
1.000

idle spark vs MAP
1.000
1.000
1.000
1.000
1.000
1.000
1.000
0.753

idle spark retard vs idle error
0.000
1.764
3.175
3.881
4.586
4.939
5.644
5.997
7.056
8.467
9.878
11.289
11.994
11.994
11.994
11.994
11.994

idle spark advance vs idle error
0.0
2.5
3.9
6.0
8.1
9.8
9.8
9.8
9.8
9.8

this setup seems to work well with the gen3 motors.

[robertisaar]

thought about it more... for the idle fuel stuff, i will definitely be making it TPS(alpha-N) based, since making it MAP based would potentially run into the same issue. so, it's going to be 17 values wide(TPS% from 0.00 to 6.27%), but i need a reasonable estimate of the RPM range it will need to reach(upper and lower) for anyone planning on running it.

to actually get into the table, i'm going to modify the code that i used to jump into my own routines to do a qualifier check to see which table should be done before actually jumping out to do them. should save me from having to rewrite too much.

qualifiers:

most definitely a "max TPS to use idle table" is one, but should i setup a hysterisis?
"max MPH to allow"
"max RPM to allow"
any others? since it seems that the only people that will NEED this are those running rather radical cams, i'm hesitant to make a MAP based qualifier.
i imagine i should add in some IAC based correction as well, since if the IAC is at ~40 steps and all other values remaining equal, a different amount of air is flowing in compared to ~20 steps. likely another table for this.

any other thoughts/opinions?

also, i've tried 4 different code revisions to try and get single-fire to double-fire transitions working correctly, but i've run into a couple of different situations now, some of which appear in even the same revision.... double-fuel delivery than intended, 0 fuel delivery for 2 full engine revolutions and then picking up into double-fire(this one is rough), and an inconsistent chance of these happening.

so, suffice to say, the ECM has won for now... i'm currently dialing down the BPW required to enter double-fire in an attempt to see if i can get it low enough to where the factory code doesn't produce bad drivability. otherwise, i may have to dial them all the way down to idle usage only, which would still provide benefit for those running ridiculous injectors compared to engine size, but near idle operation could be flaky.



clearly, there is something i don't understand about the code in the single/double fire transitions... sometimes it works beautifully and i don't notice the transition at all..... other times, it feels like i dropped all 6 coils instantly for a revolution or two(which is REALLY rough during a 3rd gear TCC lockup) and i start fearing for the life of the vehicle....

[pocket-rocket]

Now I don't want to go back through every page of this thread (lol), but did the factory use double and single fire mode switching? If so, wouldn't it be easy enough to snip out the portion of code that controlled that and put it in yours so it does it like factory? I'm going to assume here that factory didn't do it, otherwise I imagine you would have already.

[robertisaar]

well, there already is code to do it from the factory.... problem is that it does it at VERY low pulse widths and there is also a MPH constraint. the factory didn't want it done at speeds less than 3 MPH or at double-fire pulse widths of .98mS or greater. for reference, my LH0 with 16.7 lb/hr injectors idles at roughly ~1.82mS in double fire. i would need to cut fuel flow roughly in half to trigger single-fire in a factory BIN.

i imagine GM may have more or less disabled their code(via the BPW limits) for the reasons i'm running into now: couldn't get the transition to work correctly. by only allowing the change to happen at very low pulsewdiths, the interruption in torque is much less significant and therefore less noticable. i SHOULD be able to get a bit higher than the .98mS value that GM used.

other than just the ability to run significantly larger injectors at lower flow rates, i haven't really noticed any kind of differences in anything that i can measure... supposedly, the airflow into the head can be disturbed by the injectors firing(plausible enough), so reduce the number of firing sequences in half to reduce airflow disruptions.... but again, nothing i can measure really. it also has the potential to cool/clean intake valves better since a larger amount of fuel should sit against the valves longer compared to double-fire... but being batch fired(bank fire is effected slightly less if setup to do so), each valve will be effected differently since no two intake valves open at the same time or for the same duration after the injectors have fired.

so, i guess it's a wash.

at the very least, it was an interesting expiriment if i can't get it working at higher PW.

[robertisaar]

blah.....

went and tested code with the single-fire enabled only in near idle conditions.... still causes random engine speed dips/flares when transitioning into double-fire.... less of a kick now, but it is still concerning... feels like the torque converter locks and unlocks instantly.... not to mention it makes AE fueling impossible to correct.

so, it seems i may need to setup a bit in the calibration to allow between single and double-fire. well, after i can figure out how to get single-fire to work with 100% predictability at all times anyway... i may have a solution to that thanks to a suggestion on TGO..... calc the BPW required and all that jazz, then add in the async pulse width, then store to the async register, completely bypassing the normal injector register completely, since from what i'm told, storing a value to the async register causes a pulse width of whatever value is stored to be instantly injected... i just need to find out if a 0 BPW getting stored to the normal injector register will cancel out the async register....

that process MAY be the key to full time single-fire.

otherwise, Brad, Dave, anyone else considering running this on a radical engine, i need idle speed input for the idle fuel table.

[34blazer]

*subscribing*


@superdave, looks like im in the same pickle as you with the optical VSS. only my application is the mulletstang with a 7730 and harness. my speedo is cable driven as well and it seems the only VSS i could find is the optical that piggybacks the speedo cable. *edit* i saw your post about S-AUJP with the optical sensor, but id be interested to see if it can work with nAst1.

@Robert, id be interested in using nAst1 but my current setup will be using a Ford/TFI dizzy with the HEI modification. basically it uses the duraspark dizzy's pickup and reluctor and i mounted a ICM from a SD TPI car in place of the TFI module. just finished it today actually, lol. what all would i need to change to be able to use nAst1?

small cap max timing?
ref angle?
spark latency for small cap?

thats all i can think of for now. nice work btw

*edit*
after i finish the poopstain(stang) im going to convert my blazer to 7730 and use this def. the blazer also has a pretty aggressive solid FT cam and my idle sits right around 850 at ~58-60 kpa. sometimes i can get it idle lower but ill have even less vacuum, like 65-70 kpa. not as aggressive as superdaves 3500 lol.

[robertisaar]

i'm about the last person to ask about dizzys...

max timing values should be close to the limits of a small cap already.
reference angle will likely need to be changed, might have to peek at the $88 F-body 3.1 BINs for an idea on what it should be set to
spark latency, no idea, but again, the older F-body stuff might be a good place to start.

[34blazer]

i was going to peek at an $8D since i have iron heads. only thing i need to figure out is where to set the ref angle at since my pickup is in a completley different spot than the usual. ill search some on the TGO about setting ref angles so i dont hijack this awesome thread. thanks!!

[robertisaar]

well, since it's a topic that can apply to anyone running a distributor, regardless of the engine, any accurate information is welcome.

[Superdave]

for idle i'd say 700-1400 would cover just about anything. I've seen non-cammed autos idle around 500 but that's pretty low for a '660 in MPFI mode.

[robertisaar]

hmm.... i know the wife's GP can tolerate a 600 RPM idle, but only without any kind of steering input, that causes too much of a dip and risks stalls.... but it's also dead stock, so yeah, 700 does seem more than reasonable. i'll wait a few days to see if anyone else might need a higher range, but i guess that guves me time to work on the single-fire stuff...

[robertisaar]

been a bit busy with the 3100 stuff, but i am testing a new single-fire revision today.... one that transitions from dual/single as necessary but without the async register.... should be interesting.

but, i made a little tool for anyone wanting to "upgrade" from an existing A1/DF or any other mask that uses the same main spark table style(particularly if you have heavily tweaked tables), you copy in your existing table and it will spit out all 3 main spark tables for use in nAst1. saves a LOT of time. it can't account for the >4800RPM spark adder stuff, but from the looks of it, not very much used it anyways. but it would be simple enough to correct for it using the TP toolbox, just use the offset function.

also included are some 93-95 3100 and 3.4 main spark table starting points(ripped from their original PCMs), in case anybody might need them.... they already have the >4800 RPM adder stuff factored in as well(which BTW, only the L and N bodies use it(identical main spark tables too) and add 6.4* of advance between 4800 and 5600 RPM and hits it's breakpoint).

this package will be integrated into the next major release.

[LZeppelin513]

I have a stock 3500 in my Fiero. I used one Roberts nAst1 '92 3.1 bin he uploaded on post #116 and the car fired up.

[robertisaar]

somewhat of an update:

learning how to do PID stuff, assuming i succeed at a little test i have planned very soon, i'll have a guide to write and files to upload...

what prompted this was an intermittent oscillating idle between 0-10 MPH with no throttle or brakes... eventually, traced it down to the IAC PID tables being not optimal for even a stock LH0. so, now i get to learn how PID control really works, which will be helpful in the future when i integrate the option of full-on 0-5V wideband AFR as the primary fuel correction sensor.

otherwise, i'm drafting the code for idle VE table now(sitting for a month in case anyone had suggestions other than Dave's suggested idle range), 0-6.27% TPS, 600-1400 RPM, table will show values in %VE, though if they're accurate or not as actual VE numbers has yet to be seen.... it can basically be thought of as add fuel or remove fuel, in .39% increments.

in addition, the following requirements will be used to determine idle VE or the main VE tables:

Idle VE table option bit
Maximum TPS to use Idle VE table
Maximum MPH to use Idle VE table
Maximum RPM to use Idle VE table
(likely setup a bit in the datastream to indicate if idle VE or main VE is being used)

if/when idle VE is being used, an IAC scaler will also be used to skew the VE number from the 3D lookup to account for that added air that otherwise would be unmeasured. for anyone not running an IAC for whatever reason, you would just need to disable the scaler(make it 0) and then fuel won't be added based on assumed IAC position.

should you want the ultimate in fueling accuracy while in the idle VE table and while using an IAC, disable the scaler, disconnect your IAC(not just the plug, the entire assembly and plug the port), do your inital tunes that way, then when your fueling has been dialed in, connect the IAC again, and then adjust the scaler to get back to the AFR you were at before and IAC should be perfectly accounted for.



shouldn't take too long...