No luck in the DFI/ECM forume so I thought I'd try it here?

After runnign Diacom on 4 different cars (and making sure it's in "normal" mode, not a diagnostic mode of some kind) I have noticed something I didn't think was supposed to happen- all the Block Learns go back to 128 after every time you shut the car off and restart. Shut it down, remove key, immediately restart and everything is back to 128 in all cells from idle on up. As soon as the ECM goes back into closed loop/learn mode ON you can watch them start moving up and down until they stabilize wherever they normally do (little above or below 128 on all cars tested- all are stock, unmodified engines in good running order).

I thought the ECM was supposed to "learn" over time and remember but apparently it doesn't from what I've seen. Unless Diacom always blows away the block learns every time it links up to the ECM again.

Anyone know whether I'm interpreting this correctly?

 

From what I've seen, Open Loop is always, 128s.
So when you start the engine, and until it goes closed loop then it would be 128s. Then when it went to closed loop it would start with the learned values, and update from there. Depending on the BL update rate, and closeness of the cal it might just seem to be 128s all the time.

 

Then it's one more reason to make sure your chip programming is real close to begin with! The "learning" doesn't seem to carry over to places you think it should or might. Open loop, WOT, etc.

I always thought that the ECM would arrive at a block learn value for a certain cell and then apply that block learn to that cell under all conditions from that point on unless it had a reason to change it. Clearly, it doesn't work like that.

And I'm still not sure the ECM doesn't re-learn everything after each restart. You're right that the numbers mighta been so close that it looked like learning, not just immediately applying the old values, but I'm not so sure. I'd have to play around a while longer. I gotta go find a car that has normal block learns around 120 and see if I can watch it learn down or if it just "jumps" down to them as soon as it goes into closed loop.

 

It really depends upon which ECM/mask is in use. As an example the '747 (a TBI ECM) has two sets of BLM min/max limits. The first set is the closed loop learn limits. Can be along the lines of 108 min & 172 max.

Then at power up all of the cells are checked against the min/max init values.

These can be along the lines of 125 min & 135 max. So, at powerup any cell that exceeds one of these two values will be set to that value.

The current BLM value is typically used during open loop fueling. During PE mode fuel will be added but not removed in accordance to the BLM value.

RBob.

 

Bob- I can beleive that. It FEELS like something GM would do, even though I'm not ECM expert by any stretch. So maybe it doesn't blow away the BLMs completely but if they're outside a tight range it does- to give the ECM/sensors something of a fresh crack at relearning stuff that looks too far out of whack from the last restart.

The truest part about your answer: "it depends." Damned straight! That little phrase seems to creep into a lot of answers related to the ECM! If only there were simple answers, but there aren't many times.