AE pulsing in DFCO

16197427, $0D, AMC/Jeep 401 MPFI.

Ok, so lately I have been getting some random backfiring out the exhaust when decelerating (DFCO) It started out a few months ago and has been getting worse. I can not remember exactly when it started and what/if any changes I made to the bin when it started. I dont remember because it started out weakly, just every once in a while, then it would only do it when cooler, and stop once it was fully warmed up. Now its doing it all the time.
I couldnt figure it out, why it was doing this in DFCO. Only thing I could think was fuel had to be getting into the engine in DFCO. I was fearing a leaky injector. But before that I decided to do a data log, and low and behold, AE was pulsing rapidly when in DFCO!?!?! You can see it going on-off-on-off and the injector BPW going from 0 to 1.69 and back to 0 as the AE pulses. I am at a loss of what could be causing this. Should it not be disabled in DFCO?
Only thing I have been making any change to is TPS AE and a little to the MAP AE. With it doing it more when not fully warmed up, I thought maybe it was a MAT/AE correction issue, so I disabled it in the bin and set all the MAT parameters back to what it was stock in the bin (no IAT from the factory in this bin) This did not make a difference. I have tried changing the 0 TPS cell in the TPS AE table to 0 and lowering the lower TPS values a bit. This did not help either.

I have changed the AE settings in this bin as it has never been enough. For the life of me I can not get the AE to be right. It seems like the more I raise it, the better it does and no more lean spike if I hit the gas, but if I blip the throttle at idle it almost floods it. The backfiring seems to get worse too. AE has been the one thing since day one on this conversion that I have not been able to sort out. Its driving me nuts.
Anyone have any idea why it would be pulsing in DFCO?

Attached is a copy of my bin and a short data log showing the issue.

I was playing with DE trying to get rid of that same issue when I blip the throttle. the higher % of fuel I pulled out with DE it would pop. when I blip the throttle I only get a little DE no DFCO though. looking at your log I find it weird that you are getting DE and DFCO at the same time.

Have been analyzing AE in AUJP $8d, so this is of some interest. Admittedly, not overly knowledgeable of $0d, but some fundamentals are common across all code masks. Some observations and thoughts:

• In $8d and other masks, the following occurs with Decel Enleamnemt (DE) and DFCO. When TPS% goes to 0 from some significantly higher value, DE is first invoked as PW begins to drop to zero. This occurs rapidly (as it does in your log extract below). But then DFCO begins when PW = 0. Your log shows both DE and DFCO occurring simultaneously. That may be the norm with $0d, but unknown without tracing the code hack.

			TPS	IAC			BPW			MAP
Sample #	Secs	RPM	%	Steps	MPH	AE	mSec	DE	DFCO	kPa
327	54.955	2450	13	150	58		2.73			46.88
328	55.119	2450	13	150	59		2.81			46.15
329	55.283	2400	12	150	58		2.61			45.41
330	55.446	2450	10	150	58		2.37			41.35
331	55.631	2475	9	150	59		2.20			39.14
332	55.795	2450	9	150	58		2.12			39.14
333	55.959	2425	9	150	58		2.08			38.03
334	56.123	2450	6	150	58		1.72			34.71
335	56.287	2425	3	150	58		2.59			28.07
336	56.451	2450		150	58		2.01			23.27
337	56.615	2425		150	58		2.08			23.27
338	56.779	2400		150	57		0.00	YES	YES	23.27
339	56.943	2425		150	57		1.69	YES	YES	24.74
340	57.128	2375		150	87		0.00	YES	YES	24.38
341	57.292	2425		150	64	YES	1.69	YES	YES	24.74
342	57.456	2350		150	58		0.00	YES	YES	24.38
343	57.62	2375		150	57	YES	1.69	YES	YES	25.11
344	57.784	2375		150	56	YES	0.00	YES	YES	25.48
345	57.948	2375		150	56		0.00	YES	YES	24.01
346	58.112	2325		150	56	YES	1.69	YES	YES	25.11
347	58.276	2350		150	55		0.00	YES	YES	23.27
348	58.44	2300		150	56	YES	1.69	YES	YES	25.11
349	58.625	2300		150	55		0.00	YES	YES	24.38
350	58.789	2300		150	55	YES	1.69	YES	YES	25.48
351	58.953	2300		150	55		0.00	YES	YES	24.74
352	59.117	2300		150	55		1.69	YES	YES	24.74
353	59.281	2300		150	55	YES	0.00	YES	YES	25.85
354	59.445	2275		150	54		0.00	YES	YES	22.53
355	59.608	2300		150	54	YES	1.69	YES	YES	25.85
356	59.772	2225		150	54	YES	0.00	YES	YES	24.74
357	59.957	2250		150	53	YES	1.69	YES	YES	25.48
358	60.121	2250		150	54		0.00	YES	YES	24.74
359	60.285	2250		150	54	YES	1.69	YES	YES	25.11

• In your bin, the following tables are both all zero:

Decel enlean PW reduction .vs. delta MAP

Decel enlean PW reduction .vs. delta TPS

The code hack I have shows those tables to be non-zero ("OD PFI update.zip" here: http://www.gearhead-efi.com/Fuel-Inj...Information-OD). Perhaps that is somehow contributing to PW not being zero when DFCO is enabled if in fact DE and DFCO do occur together in $0d

• Most importantly, the bit that indicates AE is active is (L003D b6 = Byte Offset 19 in the DDS as reported in the TPro Log). That bit is set only when L0005 b7 is set. L0005 b7 is set any time L0261=AE Fuel Mass is > $05 as shown in the code segment below. That would indicate some sensor is telling the code to calcualte an AE Fuel Mass (L0261) believed to be needed, that fuel is in excess of the $05 threshold and therefore AE is flagged as active. Can it be that the MAF sensor is not reporting correctly? Here's the $0d code:

Code:

---

LSAE08 STAA L0261              ;Save it, AE fuelmass
            CMPA L48B5            ; $05 = Fuelmass threshold to flag AE as being active
            BLO  LSAE09          ;Bra if fuelmass < threshold to flag AE as being active [Set AE NOT Active]
            BSET L0005,#$80  ;Status word, set b7, AE in effect
            BRA  LSAE10          ; Bra to continue
LSAE09 BCLR L0005,#$80    ;Clear b7, AE not active

---

and then later, if L0005 b7 is set, L003D b6 is set:

Code:

---

;-Check for sync. AE
          BRSET L0005,#$80,L822F ;Status word, bra if b6==1, AE in effect 
;
;-No AE needed here
;
          BRA    L8247  ;Bra to continue
; 
;-Flag AE and lock INT for AE active
;         
L822F      BSET    L006E,#$08  ;Status word, set b3, AE in effect
          BSET    L003D,#$40  ;Status word, set b6, AE active

---

Finally, the 1.69ms PW consistently shown indicates a constant anomaly. With all the factors determining PW, it would be rare indeed to get a repeatable PW.

Two other PW observations from the log indicating something is not right, possibly with a sensor:

• 1.69ms PW is lower than the lowest PW encountered during idle (not shown)
• When DFCO is completed at sample 422 below, RPM increases from 950 to 2250 whereupon DFCO is again entered -- but all this happens at 0 TPS% but increasing IAC counts. Something clearly not right.

			TPS	IAC			BPW			MAP
Sample #	Secs	RPM	%	Steps	MPH	AE	mSec	DE	DFCO	kPa
419	70.342	1625	0	111	53		0	YES	YES	25.85
420	70.527	1375	0	104	47	YES	1.69	YES	YES	28.07
421	70.691	1150	0	99	41	YES	1.69	YES	YES	31.76
422	70.855	950	0	96	40	YES	2.11			38.4
423	71.019	1125	0	96	40	YES	1.89			36.55
424	71.204	1150	0	94	40		1.68			32.86
425	71.368	1100	0	94	40		1.62			33.23
426	71.532	1075	0	91	39		1.65			33.6
427	71.695	1025	0	90	39	YES	1.74			35.45
428	71.882	1025	0	89	39	YES	1.8			36.18
429	72.045	1175	0	86	39		1.59			33.6
430	72.209	1400	0	86	39		2.62			28.07
431	72.373	1600	0	84	39		2.27			24.74
432	72.537	1825	0	84	38		1.89			22.53
433	72.722	2050	0	81	37		1.69	YES	YES	21.79
434	72.886	2225	0	106	36		0	YES	YES	22.9
435	73.05	2425	0	132	35	YES	1.69	YES	YES	24.01
436	73.214	2525	0	138	35		0	YES	YES	23.27

Guess I should have mentioned I have a manual transmission. So I think the point where you see rpm go from 950 to 2200 with no TPS increase is when I downshifted.

I also see you mentioned the MAF. This is speed density system, I do not have a MAF sensor.

I do not think (do not remember anyways) ever changing any DE settings in the bin. Only thing I have really fooled around with in the last few months is AE. So it has to be something with the AE settings I changed. I am about to change all the AE settings back to stock and start over.

Understand SD, and manual trans makes sense. Sorry for assumptions otherwise. But SD or MAF, there is clearly something causing the calculation of the apparent incorrect 1.69ms PW. I'd also check the DE tables against a stock $0d because your's appear to be zero.

That was some good code lookup work 84Elky! :thumbsup:

I wonder if the unusual PW number of 1.69 is from "DFCO Additional AE" = "Tips: Maximum Time after exiting DFCO to sustain additional AE specified at L4935 "DFCO Additional AE". We've talked about the additional shot when exiting DFCO before and it looks to be tied into AE in this mask.

Looks like enough changes were made somewhere that it is on the verge or in the Hysteresis of exiting DFCO when this is happening?

Quote:

---

Originally Posted by EagleMark

That was some good code lookup work 84Elky! :thumbsup:

I wonder if the unusual PW number of 1.69 is from "DFCO Additional AE" = "Tips: Maximum Time after exiting DFCO to sustain additional AE specified at L4935 "DFCO Additional AE". We've talked about the additional shot when exiting DFCO before and it looks to be tied into AE in this mask.

Looks like enough changes were made somewhere that it is on the verge or in the Hysteresis of exiting DFCO when this is happening?

---

Thanks Mark:happy:. In the hack I have (re: link in Post #3), L4935 is marked "Not used". And that seems correct as it is declared = 0 and is not found in the code:
L4935 FDB 0 ;Not used

PW is defined in the code in a few spots:

Code:

---

            CPD    #$0100          ;3907 usecs
LAIF09      CPD    #256            ;3.9 msec
            LDAB    #32            ;488 usec bias for lookup

---

So with the above, can calculate a PW factor for 1.69ms as110d = $6e as follows:
Given 256d * X = 3.90ms or 32d * X = 0.488ms
X = 3.90ms / 256 = 0.01523 or 0.488 / 32 = 0.01525

And:
1.69ms / 0.01523 = 110d $6e

Found nothing even close in a PW related declaration or table, so the 1.69ms seems to be the result of a calculation as you also eluded to. It has to be something weird though becuase again, getting the same value everytime is just not possible in the real world.

It should be well within the range for DFCO to be active. I have it set to not exit untill 1000 rpm, and must be going faster than 15mph.
I'll be honost, once you start getting into code, it gets a little above my head.

you might have a corrupt bin, it is reporting the same map[kpa] at 2200 rpms as 1600. or a map sensor that has a little oil in it.

Ok, so I fooled around with it some yesterday. I decided to start new with all AE settings.
In the past, I have tried stock 5.7 settings (which were too little) and copied in stock 7.4 settings (wich were almost too much) But this was also back when I was running 10* of inital advance. I have since changed it to 0* like the bin had stock and so many things changed when I did that. Most all for the better, except AE suddenly became too little. I was running a slightly modded 5.7 settings. Ever since the change to 0* base timing I have been trying to get AE working good, but not much luck, thats also when this issue started popping up.
So I decide to try again with the 7.4 settings and start fresh. So I copied in every single AE settings from a 7.4 bin and gave it a try. No more backfiring in DFCO and/or AE pulsing. But it was still too little AE. I have tried tweeking a few parameters and tables and got it a little better, but still not perfect. As I raise AE and get better running, it gets to where if I touch the throttle at idle it go super rich.
AE has been the one thing with this swap I have not been able to sort out. It drives me nuts!

IIRC I started with 4.3L CPI settings and then just added some AE and it was like normal tuning...

There's a lot of little tuning glitches with this TBI to MPFI conversion... when your picky and want it right, like cold start and hot start too!

Yea, a lot of it probably is the tbi to mpfi switch, since the pcm and bin were never ment to run mpfi, so things are going to be off, and probably will never be perfect.