Deceleration tuning struggles with manual trans swap

XR7-4.6

Administrator
Staff member
Joined
Sep 12, 2023
Messages
1,873
Location
Roselle, IL
Vehicle Details
1994 Cougar XR7 DOHC TR3650
Country flag
So I've been datalogging and tuning with a wideband on and off for almost two months, mostly dialing in the MAF transfer function which is pretty damn good now but I'm now working on other stuff and I just having zero desirable results.

Coasting fuel shutoff(CFSO) plain doesn't work, I've tried everything, even contacted SCT to expand the section in my initially limited pro racer database and can confidently say there's no magic value missing, so it's got to be something else preventing it. Even copy pasted over other people's verified working values(such as Brandon's) to see if I could get it, but no. No CFSO flag in the datalogs, no lean peg on the wideband(actually it goes a little rich to ~13:1) and injectors still pulsing. Unfortunately I cannot say for sure if this worked with the automatic or not before the swap, I *thought it did* but I can't go off of 9 year old memories.

Other issue, semi-related, When I let off the throttle to decelerate or if I'm in between gears from about 2500 rpm or above the exhaust crackles and pops in between like a god damn BMW, I don't like it. It's obviously dumping fuel into the exhaust when the throttle closes where it ignites, what I can't quite figure out is what I can do to remedy it. CFSO shutting off the injectors would help but I I know that's not the root of it. I don't understand the transient fuel section at all, there's no descriptions, very little help on the pro racer forum but I suspect this is the section I should be lookin at?

Other thing also related is there's not much engine braking, not for my liking anyway. It's not a matter of rev hang or flare when I push in the clutch but revs are really pretty slow coming down. I even have a IAC restrictor I threw in to help this some and it's still not quite right, and I'd like to actually learn this stuff the right way. I copy/pasted the idle/dashpot values for a 98 GT manual into these areas I've messed with dashpot a little from these, adding up to 20% dashpot decay and subtracting as much from dashpot clip 1 as per the descriptions but the difference has been negligible if it did anything at all. Dashpot preposition truly does absolutely nothing at all, the stock table in my database is all zero'd out and the Mustang values I used are .270 to 2.000 air #m and from 344 to 16383.7 rpm... loaded two tunes today, with that and one the stock zero'd values, and no difference whatsoever.

Anyone got any ideas? Is there something I'm overlooking? These are a few things I'm currently thinking about, feel fee to say if I'm warm or cold.

-Under MAF scalers there's a descriptionless scaler called "MAF dynamics filter for decel" that has me curious
-Manifold volume I have set to 10 which supposedly is the Mach 1 intake manifold volume according to mitch B (though I have file that says it's 14 but that sounds wrong). This has an effect on transients but it seems fairly happy based on my wideband data.
-the scalers in transient fueling that mention decel or dsfo(decel shut fuel off??? DFSO is used interchangably with CFSO, or mayby its just a type o)

1696471118480.png

-I am fairly certain there is no clutch input to the EEC like there is on SCs and other EEC IV Fords, and logically I'm not sure why that would effect these things anyway but that's about all I can come up with from that wouldn't be tune related
 
I can't contribute much as I bought a tuned chip for mine, and don't know much about the subject.

It's not a matter of rev hang or flare when I push in the clutch but revs are really pretty slow coming down.
I have noticed this same thing on mine. If the car moving, RPM is slower to drop down between shifts, clutch in when coming to stop it slowly drops to idle. If the car isn't moving the revs will drop quickly like expected. I have assumed it was some programing parameter for the automatic that is left behind.
 
I haven't fussed much with dashpot functions. If anything I think my car decelerates too much in CFSO (I noticed the Mark VIII does not enter CFSO when coasting FWIW). And no, I haven't got a clue about the transient fueling tables either. Why they let us see those and not other things (for example, TCC torque limit) I'll never know.

What does your fuel source do when you let off the throttle at speed? Might provide a clue or two as to where to look.

I'll poke around the house to see if I can find my LaSota tuning handbook to see if it mentions anything about auto-manual swaps beyond the obvious stuff. Something tells me what you're trying to tweak is seldom addressed by PRP guys - maybe even fairly limited amongst dealers.
 
Fuel source is the base table (I'm still datalogging in OL/adaptive off) When I say it goes a little rich on decel it's a brief blip not continuous, it'll get to lambda if I just let it coast but it'll still burble and pop a good bit.

There's all sorts of weirdness in what they show and don't, I had to go up a level with recertification to see the parameters you already had. Transient stuff seems extremely useful but not even one thing in that massive section has a description, nothing in help either

I can't contribute much as I bought a tuned chip for mine, and don't know much about the subject.


I have noticed this same thing on mine. If the car moving, RPM is slower to drop down between shifts, clutch in when coming to stop it slowly drops to idle. If the car isn't moving the revs will drop quickly like expected. I have assumed it was some programing parameter for the automatic that is left behind.

And I *think* that's dashpot but my experimentation has netted zilch in improving it. I'm willing to try heavy handed tweaks but I just hate doing stuff that way. It's pretty frustrating, unlike a typical manual I use the brake to modulate speed almost as much as I did the automatic
 
Engine decel is set by the dashpot function; Oscar was dinking with it wbw I first got the prp. That, and the lockup changes got me closed throttle decel, and the wideband goes full lean, like the speed limiter. (I put mine back in when the state sends me a letter, lol.)
 
I know engine decel is largely set by dashpot, but the question is is there a more scientific yet also heavy handed way of adjusting it besides add 10% to dashpot decay rate and subtract 10% from dashpot clips and guess and check a bunch of times to start seeing results? I mean these are tiny decimal point numbers going into the hundred thousandths.

I guess I'm primarily curious if there's a good "start from scratch" set of values I could input starting from the point of maximum engine decel(which I want) and adjusting from there to refine it from possibly stalling/bucking, rather than starting from stock values for a SOHC/4R70w that don't pertain to my current DOHC/manual trans powertrain.
 
Ran a few tunes messing with dashpot decay and clips, I give up! I tried the heavy handed approach just to see what happens, drivability was fine, but didn't do squat to decel and it made the exhaust popping and crackling off throttle REALLY bad. I ended up going back to the tune before I started messing with this stuff and it's a lot better. That was a good sanity check since it got so bad I started thinking something mechanical went bad(exhaust leak, injector leak).

I think maybe I'm putting the cart before the horse with dashpot anyway, transients are what I should be focusing on. Last log I did with manifold volume set to 14 I stabbed the throttle and for a very very short period saw a lean spike to 1.18 followed immediately by dead on commanded at .89, but then letting off the throttle it would go rich as much as .71. If I can get it to stay in stoich or better yet lean out on throttle tip outs I don't really give a crap whether I can get CFSO working or not.


Off topic does anyone ever get odd blink and you miss it spikes in pid values in datalogs? I don't know what the hell that is but it's done that in two different PCMs so I don't think it's corruption, but it does seem to occur in almost every other datalog I make
 

Similar threads

Back
Top