"Block Learn" vs. "Intergrator Value"?

[Paul Workman]

I've much more experience with scanning the LT1 (my previous toy), so recent Qs re LT5 O2s got me curious about something, i.e., "block learn" and "intergrator" values for the LT5...

A really useful tool for observing the ECM's reaction to O2 signals is the injector dwell time milliseconds. Dwell time is a direct measurement of the ECM's reaction to O2 signals (available on the (95) LT1 ECMs). But, apparently, the LT5's ECM doesn't have that data available; or at least not on my AutoXray scanner. So, check me out here:

I'm left with-
-the O2 millivolt value averages over 1 sec. time
-left and right block learn values
-left and right O2 cross counts
-left and right intergrator values.

I feel pretty clear on the O2 voltage readings and cross counts interpritations, but I'm a little fuzzy on "block learn" and "intergrator" - just how and what, exactly, are the derivitives of these values?

My understanding is the block learn values are indicative of fuel flow values and intergrator values are indicative of air/fuel mix. Is this the correct way to read these values? (How I miss "injector dwell"!!).

P.

[XfireZ51]

BLM is the Block Learn Multtiplier and also known as Long Term Fuel Trim, whereas the the INT is the Short Term Fuel Trim. The INT will change several times before the BLM is updated. When I first started tuning, the old C3 ECM like the 7747 would update a frame every 1 sec. The BLM would hardly move. So I would lock the BLM to 128 and the ECM would use the INT only which updated more frequently. Once the INT would steady out to around 128, I would unlock the BLM and fine tune it from there.
By Injector Dwell, do you mean Injector Bias?

[Paul Workman]

Originally Posted by XfireZ51

BLM is the Block Learn Multtiplier and also known as Long Term Fuel Trim, whereas the the INT is the Short Term Fuel Trim. The INT will change several times before the BLM is updated. When I first started tuning, the old C3 ECM like the 7747 would update a frame every 1 sec. The BLM would hardly move. So I would lock the BLM to 128 and the ECM would use the INT only which updated more frequently. Once the INT would steady out to around 128, I would unlock the BLM and fine tune it from there.
By Injector Dwell, do you mean Injector Bias?

Thanks for that. That is what I figured.

Well, actually the term is "injector pulse", and the value is in milliseconds.

I had 3 O2s go bad on my '95 LT1; one was bad when I bought it, and like a dummy I filled the wire port on the new sensors with RTV, thinking I'd be protecting it from anything getting in. Live and learn.

Anyway, one O2 sensor would apparently short to ground at about 3000 rpm - reading zero volts. The injector pulse time (dwell was my word...sorry 'bout that) for that side of the engine was more than double the pulse width for the side with the working O2. Lots of black smoke on that side and stumbling, etc. to indicate too rich on that affected side. Once the funky O2 was replaced, the injector pulse width compared almost exactly with the other (good) side.

Unfortunately, I do not see "pulse width" in the 90 version of the ECM software. Too bad, cuz it was very useful, IMO.

But, getting back to the bad O2 sensor question, if the same thing happened in an LT5, I would expect the intergrator readings to be significantly higher on the shorted O2 side, compared to the normal sensor side, yes??

P.

P.

[XfireZ51]

Paul,

Injector Bias is part of the overall Pulse Width. Its there so you can fine tune the actual amount of time the injector is open. Sort of like Spark Advance. The larger the Injector Bias the more advance time allowed for injector to open. In the case of TBI, the fact that there were onyl 2 injectors made it critical to get every bit of effciency from the injectors.
Increasing the Injector Bias gave you more "effective" pulse width because less of the overall pulse width was devoted to opening the injector. PW is measured in msec in the LT5 also, just like TBI.
Regarding the bad O2 sensor scenario, I'm not sure that if the O2 is dead whether that bank would be running open loop, in which case the INT would be at 128. If the O2 is reading 0, that would indicate a dead lean condition requiring gobs of fuel. BLM should be high and INT dropping like a rock.

[Paul Workman]

Originally Posted by XfireZ51

Paul,

Injector Bias is part of the overall Pulse Width. Its there so you can fine tune the actual amount of time the injector is open. Sort of like Spark Advance. The larger the Injector Bias the more advance time allowed for injector to open. In the case of TBI, the fact that there were onyl 2 injectors made it critical to get every bit of effciency from the injectors.
Increasing the Injector Bias gave you more "effective" pulse width because less of the overall pulse width was devoted to opening the injector. PW is measured in msec in the LT5 also, just like TBI.
Regarding the bad O2 sensor scenario, I'm not sure that if the O2 is dead whether that bank would be running open loop, in which case the INT would be at 128. If the O2 is reading 0, that would indicate a dead lean condition requiring gobs of fuel. BLM should be high and INT dropping like a rock.

Me bein' edumuhkated in E-lectronix enjuneerin, I totally understand bias and the reduction of hysteresis biasing affords; makes perfect sense to bias the injectors (biased voltage selected at a point just below cutoff, i.e. class "C"). However, biasing raises a question: It may be splitting hairs, but I wonder if the cutoff for the various injectors is a standard value, or should a different bias be applied according to injector type? (I'm guessing the ECM can accomodate various injector characterisitcs, provided they are "in the window". But, it could make another case for not mixing primary and secondary injector types, huh?)

Well, (the AutoXray 1000) scanner s/w for the LT1 did divulge the injector pulsewidth, and I gather from what you said, there is pulse width data available (no big surprise) for the LT5....Dayahm! I really wish the "upgrade" I downloaded to my scanner would have included pulse width!!

Maybe someday I'll make the plunge and get into the ECM code - just "for fun". In the mean time, I'll have to muddle through by using indirect (BLM and INT) to evaluate sensor or other possible issues, when/if they arrise.


Check me out:

• If an O2 shows lean, and the fuel delivery indicates lean, that would cast a shadow on the ECM, methinks. Swapping the location of the O2s to see if the problem follows might be worth doing, and if the O2/delivery does not change sides, the O2 proves OK, and there and the focus would be on the ECM, no?

• If an O2 shows lean and the fuel delivery indicates rich, then the O2 itself is in my crosshairs first, especially if the instant fuel mileage is dips into the mid teens and the exhaust smells bad (assuming cats are installed). However, it could be a fuel pressure issue, and even tho it isn't likely to affect just one side of the engine, it is a lot easier to check and eliminate pressure issues as a possible cause than it is to jack the car and swap those pesky O2s!

Well, I'm rambling. Can you tell I'm intrigued? Apparently I have a hunger to get into the meat of these LT5s, and am feeling frustrated by the limits (pointed out by you) of my current scanner.. I want more!

Thanks for the response. I'm enjoying the journey upt the LT5's learning curve. It is a sweet piece of engineering, to be sure, and the expertise present on this board is top drawer, "fer sher"!!

P.

[tomtom72]

May I climb that curve with you? 2nd gear w/key on, right...to get up this hill?

I just think that ya also need to realize which side of the flow chart each read out is comming from....ya know is it data from a feed back leg or a control leg...

MHO: O2 = feed back side
Int # & BLM # = control side........

What do you think?


Tom

[LT5-Lee]

I also have the AutoXray 1000. I need ot become better acquainted with it and reading the values. It seems slow and limited on the amount of time for it's dialogging feature, but it is better than nothing. I also have the CD for EZ-PC connection and having problems getting a printout that matches the screen graph, does anybody else have this problem?