Custom ...747 TBI issue. Help!

[JF in MI]

If the calibration is going very lean during deceleration then it's likely the manifold has no fuel left by the time you press the throttle. This is a tough condition to detect and to tune around with 7747 / $42. The challenge is covering the difference from very low kPa MAP to high MAP without causing flooding during transitions from idle range kPa to high MAP. What is the range of MAP transition logged by the scantool when the problem occurs? I'm thinking above 60 kPa change? You may also want to adjust the lean decel so it is not quite as lean. In a carburetor where changing the pump shot is not as easy we might richen the lean stop, change a spring under a primary rod, or change taper on primary rods if those options are available.

You may also want to think about IAC operation. IIRC $42 does not have a way to differentiate between idle and deceleration when VSS is disconnected. Closing throttle to decelerate when VSS is not present signals to ecm that engine should be at idle. ECM may consider idle speed to be above threshold and will close IAC to attempt to bring engine speed into line. Stepping on throttle (Again, IIRC, study Rob Rauscher's IAC info) trips ecm to open IAC for "throttle cracker." There are time delay values that could be manipulated to prevent or increase rpm response time in order to reduce airflow changes while throttle is changing. Are you using a manual transmission calibration? Comparison between manual and auto trans cals may suggest which direction to go.

7427 is truly much more capable. 7427 is better able to handle a wider range of conditions imo. 7427 code is built with the knowledge learned while calibration earlier ecm's.

You are probably right on the kPa change but as I mentioned earlier I'm having trouble recording on my laptop (hit and miss). DFCO is switched off in the constant table so it should not enter the picture. I'm not thinking the IAC is effecting it because this won't happen jabbing the throttle, for a downshift, much above 2500 RPM which I'm thinking would make it worse (I could be wrong).

[dave w]

Often disabling factory programming has unwelcomed and unwanted side affects. I wonder what would happen if DFCO is re-enabled?

[JF in MI]

Often disabling factory programming has unwelcomed and unwanted side affects. I wonder what would happen if DFCO is re-enabled?

To answer your question; I tried it both ways and noticed no difference.

[In-Tech]

Chances are your lower(numerically) kpa values are out of whack a bit. I usually set my BLM min/max to 127/129 so only the integrator works while tuning. Once tuned, then I use "normal" BLM min/max. The integrator histogram will show you what it is "learning". If the ecm is pulling or adding fuel during decel it can cause weird AE issues.

[JF in MI]

Chances are your lower(numerically) kpa values are out of whack a bit. I usually set my BLM min/max to 127/129 so only the integrator works while tuning. Once tuned, then I use "normal" BLM min/max. The integrator histogram will show you what it is "learning". If the ecm is pulling or adding fuel during decel it can cause weird AE issues.

I was about to test this (wasn't able to get out today and it will be tomorrow afternoon at the soonest). If these numbers are what you are using at 0-30 kPa then mine are WAY off. I tripled my values down there and they are still less than half of what you are using. I thought I was adding too much but based on what you're saying I haven't added nearly enough. At least this gives me a direction to go and try.

[In-Tech]

You are on your way to perfection
Those numbers are just what is an allowable change so the integrator does the work(it's much faster than BLM changes).
The 7747 is old and weak but can do its' job with enough effort on your part. I've been doing this since the early 90's way before we had software. It is a struggle but you are well on your way.
Joby's WinALDL is still my favorite scanner for the 160 baud stuff. https://winaldl.joby.se/

[JF in MI]

FIXED!!!
First off I want to thank all those responders for their ideas which spurred me to more testing. I tried jacking up the BL numbers at high vacuum on fuel table #1 with no results. Set them back and went back to the Accel Enrichment table. From the base program I long ago started with I had to eventually raise the pulse rate usec at 0 kPa 19 fold, 13 fold at 20 kPa and 3 fold at 40 kPa. No changes at lower vacuum levels 'cause it always ran good there. It runs like a dream now with no trace of a flat spot while doing a double clutch down shift.