7730 ecm, 8D code.

What is the main purpose of them moving between cells? From what I've read on here, having them move arround is mostly for gas milage. I think that there is more to it than tho..

Kat

 

The BLM cells are to provide some aspect of "self-tuning" to the calibration. The VE tables specify how things "should" be, and the O2 sensor gives feedback on what's actually happening. The difference between what "should" be and what "is", is the BLM and is stored in the learn cells. Lets call it a correction factor.

this is useful in a lot of ways. cold startup for example: the O2 sensor is still useless, but if the calibration (VE tables) are off, then there is no active feedback system for a correction and the car will be lean or rich. With the block learn cells recording the last known correction factor, the ecm can use these values to correct itself without feedback from the o2. not perfect, but closer than no correction at all in most cases.

on some calibrations, they are also used at WOT.. if the calibration is lean (BLM > 128) before going WOT, then it's used to richen up the WOT. makes sure that a car running signifigantly lean is less likely to burn down a motor at WOT due to lean problems.

 

I've heard this with regard to the $8D calibration too. However, even though I have pesonally witnessed >128 BLMs at WOT I cannot locate a spot in the code that actually allows that greater BLM to apply. See attached source. I have shown both the ANHT and the AUJP. Both are the same. Of course, I fully admit that I might be looking at the wrong part in the code ... but, AFAIK, I am not. In the major fuel routine it pulls the BLM from L00E0 - which will be 128 if in Power Enrichment. I am wondering if the scan tool display just isn't updated for the current BLM? Kind of like in WOT it still shows Closed Loop even though learn mode is off and for all potential purposes we're not in closed loop. Hmmmmmm.

Tim

 

I thought that the ecm goes to 128blms at WOT if they are richer than 128. If the car is at 129+ blms the ecm will "freeze" the blms at the lean blm number. Is this what you are talking about?

 

That's what 91L98Z28 and others have said. That's also what I have personally seen. But, I presented some actual code because I can't find the actual code that allows for BLM fuel adjustment at WOT.

Here's a situation where an ECM bench would be really nice.

Tim

 

I should note that I have never verified personally the "if BLM > 128, then enrich WOT" thing. I had remembered reading it and thought it had been accepted as true. i'm more inclined to trust trax's first hand experience than "something i remember someone saying"

 

I have seen something going on with the blms and WOT. But I am not a code reader, maybe rbob can find it.!?

 

I personally have seen BLM other than 128 @ wot. And rob found how to disable WOT influence from BLM in code for me.. If I can find the thread I will post it.

 

But what is the ecm doing with the blms and wot

 

It was making PW correction..
I just can't figure out when and where it gets that correction value. meaning what cell.

 

87_TA,

As mentioned, I have also seen this phenomena at WOT. I'm just trying to track down exactly how it is done in the code. Basically, there needs to be a place in the code that looks for PE Mode L0046, bit 5 (#$20) and does a comparison of the BLM against 128. If greater than 128 then it applies the correction ... if less than 128 then it does not apply the correction. I just can't find that code OTHER than what I have already posted. I have searched the HAC and the BLM is set in two, and only 2, specific locations. One location is the one presented above. The other location is during the 50ms Air/Fuel Minor loop (LCCE1). However, in the 50ms loop there is nothing (that I can see) where the BLM will be forced to 128 when the current BLM is less than 128.

Given that I can't find anything - I reversed my search and began looking at the INT since the INT is always forced to 128. There is a specific routine that does force the INT to 128 (LCA85) ... and I can track the code to where the INT is forced to 128 when in Power Enrichment.

I have come to the following conclusion so far ...
1) When moving into PE Mode the ECM retains the current Open Loop vs. Closed Loop flag. If in Open Loop before PE then Open Loop will be retained. If in Closed Loop then Closed Loop will be retained and Learn Control will be disabled (because learn control fails based on shooting for an AFR not equal to stoich).
2) When in PE Mode the ECM forces a 128 INT.
3) When greater than 128 BLMs exist then they are applied. Otherwise, the BLM is limited to 128 according to the first code that I indicated.

If anyone has any evidence please correct me as I am definitely wanting to be proved wrong at this point if I am wrong. Or, maybe I am right? Seems like Rob knows the answer if he changed code for Tom (87_TA).

The following post is the one 87_TA mentions where RBob showed that >128 BLMs are utilized but less than 128 BLMs are not...
https://www.thirdgen.org/techbb2/sho...hreadid=175688

Tim

 

You have the correct code location. If you want to lock the BLM at 128 whenever in PE mode change these two locations to $01:

$49B3 and $49B4

A $01 is a NOP and will eliminate the BMI instruction. The 128 will always be loaded and used when in PE mode.

Note that those locations are for AUJP, other BCC's may vary.

The BMI instruction at $49B3 is what is checking if the BLM value is great then or equal to 128. BMI is BRANCH MINUS and checks the sign bit, a value of $80 (128d) or higher has the sign bit set (b7).

RBob.

 

does that mean there is some wot adjustments made by the ecm if the blms are above 128? What I have seen is the blms will lock at the valuethat its at just before wot(if the value is over 128). If I am richer than 128 then it goes to 128 at wot. I know I'm repeating myself but I don't understand why the blms do that

 

its a safety recaution, gm does not care if you are to rich @ wot..
Lean is the hazard.

 

I thought it was known that the gm ecm made no wot adjustments?? That is what's confusing, compounded by my desire to understand why/how the ecm is using the blms at wot.

 

I'm not going to comment on the BLM/WOT thing (don't have the knowledge to do so), but the basic thing about no adjustment at WOT is that, at WOT, the system goes open loop, and does not monitor the O2 sensor at all. It just sets the inj pulse width based on the internal tables (and possibly the BLm/WOT thing)...

 

All my log data, the car stays in closed loop durring wot. At least thats what datamaster reports to me.....

 

yeah, the data logs show that you are closed loop even at WOT.

an easy way to tell that you are "open loop", that is, the ECM is not monitoring the O2 sensor, is to look at the AFR field. If it's anything but 14.7:1 then you know that the ECM is not looking at the o2 sensor. The o2 sensor is only useful for maintaining stoicometric (sp?) fuel ratio which is 14.7:1. anything else and you have no feedback loop. so typically at WOT, my aujp used to show somewhere between 12:1 and 13:1 and that's when you know that the ecm is ignoring the o2 sensor.

 

The true method for knowing that the ECM is not looking at the O2 sensor is that Learn Control will be off when in Closed Loop. I posted about how to view the Learn Control Bit in Datamaster. Diacom already displays it.

The ECM can definitely make adjustments to WOT fueling. The proof is in the Code that RBob posted awhile ago ... along with me duplicating his effort and posting the code again above (doh!). The proof also is in all the datalogs that a bunch of us have seen where the BLMs are > 128 at WOT.

I do want to comment that the BLMs are not locked at the value right before WOT. Rather, they are truly the BLMs that are for that area of the fuel map. As an example, I have a datalog where for the first 2 frames of WOT my BLMs are 132 (cell 14). Then, when the cell switches to 15 (based off of my new boundaries, RPM window) the BLM drops back to 130.

This is new to me too. It was a general 'knowledge' that the ECM was in Open Loop at WOT and made no fueling adjustments. That's just not true ... and the actual scan data proves it. The ECM can be in Closed Loop at WOT and make lean adjustments to the fueling based off of BLM. The INT is locked at 128 and Learn Control is turned off.

Tim

 

That is kind what I thought. It happens so fast that I guess one really need to pay attention to notice what is exactly going on. It would be nice to know why/how the ecm is doing this. It raises the general question of how does the ecm know it's rich or lean at WOT w/o using a wideband. I think this is proof that there is something funky going on that the public hacks have missed, that people are not aware of...

Who the hell has an ecm bench? That would help alot

 

Actually, we do know how and we do know why. The how is the easier of the two - it is doing it according the to code that both RBob and I posted. The why is because of the exact reason that 87_TA posted. It's more dangerous to run lean at WOT than to run rich so GM took that into account to ensure less warranty claims.

I think that you might be misunderstanding something too. The ECM is not in Learn Mode when in closed loop at WOT. It is not trying to figure out the proper AFR. The O2 sensor is not being used for anything. Learn Control is off. RATHER, it is just applying any lean BLMs that it has figured out through part throttle when operating in those areas. Quite simple - yet quite ingenious from a warranty perspective.

I've learned a lot in the past 2 days by pouring over the code for hours on end. If you haven't done it yet, then definitely open up the ANHT hac and spend hours just trying to figure out some of the routines. You'll need a copy of the Motorola 68HC11 Reference Manual to understand the assembly commands ... I was lucky enough to have taken an assembly language course in college so it doesn't look greek to me - I don't understand it all (as evidenced by my misinterpretation of BMI which RBob corrected me on) but I do understand enough to be dangerous.

I definitely have to disagree there. I don't think the public hacs have missed anything. It is the users ... us ... that have failed to utilize the hacs and interpret them correctly. The information is all there. You just have to know how to read it.

Tim

 

I guess you are right. I don't know enough about assembly lang to say the public hacks missed something. You did clarify a few things for me,i think. So the ecm is technically in closed loop at WOT but learn control is off....right???

The main sticking point is how can the ecm know's it's lean at wot. From what I understand is the ecm can sense a lean condition at wot. So it'll do something(richen) if the blms are higher than 128 while in the WOT cell? If so then could 128blm in wot cell mean somthing other than 14.7:1? I'm speculating at this point. So I'll stop myself and hear what you guys have to say.

 

How is it possible to make wot fueling adjustments w/o figuring out a/f ratio. Even if it's based on part throttle(u sure ?), then is is still calculating a projected wot a/f ratio, or something.!?

 

Not necessarily. It is as I explained in one of my posts above. If the car was in open loop before WOT then it is still in Open Loop during WOT. If it was in closed loop before WOT then it remains in closed loop but learn control is turned off.

No. I think that you just don't understand. I'll try to explain this as best I can. The ECM senses nothing with regard to AFR at WOT. Please forget that. Drop it Don't think about it again. Learn control is off. It senses nothing. Ok? Ok. During normal part throttle operation when the car moves from one BLM Cell to another then it automatically begins using the learned values for that BLM cell. It doesn't need to sense anything to do this. It just begins using what had already learned. At WOT the ECM takes a similar approach. As you move from one BLM Cell to another it just uses whatever BLMs have already been learned during part throttle operation ... but applies them to WOT. No sensing is needed. It just applies previously learned BLMs to the RPM/MAP ranges you are in at WOT.

You know what this all stems from? Well - this goes back to airflow and EVERYTHING that I have previously said about MAF and SD systems. In a SD system the ECM assumes that the VE tables yield a 14.7:1 ratio. The PE tables modify this ratio to get to 12.5:1 (or whatever). If the ECM needed a BLM value in part throttle to reach that 14.7:1 ... then it uses it at WOT in order to ensure that the VE Values for 100kPa really produce at least a 14.7:1 before also applying the PE adders.

Please read everything I wrote and think about it at least 3 times before replying.

Tim

 

makes sense...I understand and thank you..just a few questions.

How does the ecm determine what the blms are at in the wot cell? The ecm will never seen that cell durring part throttle. That is why I though that the ecm just goes to 128 in the wot cell(usually).

Earlier I mentioned, you replied later you said That confuses me because the ecm uses the part throttle for wot calculation or not? If so then how because the wot cell is never used at part throttle. And the learned blms are in part throttle cells...

Maybe you could define how the blm functions. Because I might have that wrong. I'll attach this thing I did with th ve table. I applied the stock cell boundaries to both ve tables. From what I understand is that the ecm uses the ve entries inside the cell to come up with an averaged blm number. How can the ecm average the blms the cells when they are only touched at wot?

Thanks for helping out, I'm a little slow sometimes:lala:

 

The BLM cell matrix only covers a portion of the VE table. However, its influence extends beyond the outer cell borders. The following is a general discussion of how the BLM cells matrix and learn control operate (which means that I won't be throwing bin locations out ).

With the engine running and up to temperature along with an active O2 sensor the ECM goes into closed loop. Once in closed loop the INTegrator will change in order to adjust the PW and therefore the AFR. The feedback device being the O2 sensor. Learn does not need to be active for this to happen.

Learn control only affects the BLM cell matrix. It is entirely possible and it happens that the INT will change without the ECM learn control enabled.

If either the RPM or the MAP is outside of the BLM cell matrix outer borders then learn is disabled. No BLM cell learning can take place as the engine parameters are outside of the matrix. The INT can still make adjustments.

So two things are happening here, the INT will change whenever closed loop is enabled, the BLM cell value can only be learned when the engine RPM/MAP is within the cell matrix borders (we are excluding PE operation).

Now comes the fun part, the BLM cell matrix outer borders become transparent once the RPM or MAP is outside of those borders. Learning will not take place, but the edge cell(s) will be used to adjust fueling.

The BLM cell matrix is a 4x4 for 16 cells. There are 12 cells along the outer borders (4 borders), then 4 more cells on the inside. The borders extend to the edge of the VE table, if just for use, not learning.

An example, the upper RPM limit of the BLM cell matrix is set to 3600 RPM. MAP is at 100, 80, 60, and 40 being the lower limit.

When the engine is at 3400 RPM and 70 kpa it will be in learn mode using the cell bounded by 80 & 60 MAP. This cell will be learned in. As the RPM increases above 3600 RPM (70 Kpa) learn control will be turned off.

However, since the cells transparently extend past the outer borders, that same cell will still be used to adjust the fueling. But not in a learn mode.

If the go-pedal is pressed a tad more and the MAP goes to 90 Kpa and is still above 3600 RPM, then the next cell up will be used: the one bounded by 100 and 80 MAP Kpa.

This using of these cells will be the same even in PE mode. PE mode is a special mode as learn is always turned off. Closed loop will stay active but the INTegrator will typically be locked at 128 (not all masks do this, some will lock the INT at whatever value it is when PE mode is entered).

And, when in PE mode the BLM is ignored if it is removing fuel. It will only be used when adding fuel.

Confused yet? The more I run open loop mode (except idle) the more I like it.

RBob.

 

Posted here to make the discussion easier

 

Could you lend a hint as to what you mean? If the BLM cell matrix only covers a portion of the VE table then can you give an example of a portion that it doesn't cover. When I look at the 16 cells in the pic I posted above ... I see that every portion of the VE table is fully covered. Basically, I see that the center 4 cells are bounded. However, how can the MAP or RPM be outside the matrix? As an example, lets talk about 90kPA and 4000rpms. That's cell 15. How is that 'outside' the matrix?

Are you indicating that learn control is disabled whenever the ECM is in an outer cell ... i.e. one of the 12 outer cells?

Tim

 

So how does the ecm do that? You make it sound like the ecm wonders over to the wot cells and that is how it "learns" to richen if a lean condition exist.? The nagging question is how does the ecm detrmine that WOT is/could be lean? The thing is that even if the ecm is at XXXblm it doesn't mean that WOT is rich or lean. So how can it conclude a lean condition exist to correct for? Is is just a fixed percentage from 14.7:1(if over) that the ecm will compensate? What is the calulation for wot fueling corrections that the ecm uses durring so called lean conditions?

Thanks!

 

11's , its merely a safety device....
If it senses a lean condition @ part throttle (that it assumes is already supposed to be 14.7:1 because ve should not be wrong)
So if the ECM senses a lean condition @ part throttleit assumes there must be a calibration or possible mechanical problem and or disadvantage - so if part throttle was supposed to be 14.7:1 and was leaner it will assume worse case senario that WOT could be in danger of being lean.. so just in case it ads fuel...
When in doubt, fuel it up
Hope that made sense , it did to me but I have been drinking.

 

quote:
--------------------------------------------------------------------------------
Originally posted by RBob
The BLM cell matrix only covers a portion of the VE table. However, its influence extends beyond the outer cell borders.
...
If either the RPM or the MAP is outside of the BLM cell matrix outer borders then learn is disabled.
--------------------------------------------------------------------------------


As each mask is different maybe I should have concentrated on just $8D, I was speaking in general terms.

Most of the time there are additional terms defined for the BLM cell matrix outer boundries. I just looked and found this one for the $8D mask:

In this case there isn't an upper or lower RPM or an upper MAP boundry. There is the lower MAP boundry though. With that the lower portion of the VE table can be cut off from learn.

For the '747 and '8746 there are upper & lower MAP boundries along with an upper RPM boundry:

Once outside of these thresholds learn will be turned off.

No, only when outside of one of the outer boundries. In regards to the $8D mask the example I gave wasn't the best, as there isn't an upper RPM boundry in that mask.

RBob.

 

This discussion has been very interesting....but. you are missing a key piece of information in explaining how and why the BLM's work. This key piece of information is something I call 'knowing the basics", and without knowing the basics, this is how myth and misinformation get started. Kinda like a Native American feeling the wind blow and say it's due to the Wind ***, when in reality it's more like a difference in air pressure, or if you're from Kansas Like I am, it could also be uneven heating of the air.

So, what is this 'key piece of information'? Someone needs to rustle up an injector flow graph. And all it takes is just one look and everything becomes abuntantly clear as to the why and how of BLM's. An injector flow graph is a must, I mean, if you want to design an ECM to control an injector, the *first* thing you do is characterize what you want to control so you can design in what it takes to overcome any idiosynrocies (where's that spel chekker?). In my many years of reading the automotive BB's myth and misinformation about fuel injectors has been rampant, even to this day and even to this BB.

I hope this doesn't sound like a flame or anything I'm just trying to help, but sometimes I get a little frustrated when the same converstations take place on the different BB's I visit. And .....I guess that's enough.

Dave

BTW, I have got the info needed and I spent about $500 and a lot of time to get it. So if you post something, I'll know if it's right or not. So why don't I just post it? Well, maybe to prove a point, and I would also like to recoup some of my investment, so I'm reluctant to give it away.

 

As far as I can tell, learn in ANHT / AUJP is only disabled by the following parameters...

1) If not in closed loop
2) Coolant temp is lower than the minimum temp needed for BLM update
3) Coolant temp is higher than the maximum temp for BLM update
4) Shooting for an AFR not equal to stoichiometric (PE Mode, Hwy Mode)
5) If the MAP is less than minimum needed for learn (this is 0kPa unless changed)

So - to me the code really indicates that learning will only be disabled whenever the ECM is Open Loop, PE Mode, Hwy Mode, and other routines such as DFCO, etc. Almost all the time you will be within the coolant boundaries and never less than 0kPa.

I went further and searched the entire hac for LD31A (which is the first line for disabling the learn bit). This yields nothing except for the main code in the Learn routing (which I commented on above in layman's tems).

To prove my points I have presented the code below.

RBob, if you have read this far into my post .... it seems that maybe a lot of what you said doesn't necessarily apply to the $8D?

Tim

 

Dave,

Nice to see you. Thanks for posting. Guess you don't visit ETO much anymore? I think that site is DOA. I see that you have the info needed to explain things. It would be cool if you shared. I understand that you spent $500 and alot of time. But if you share,then you'll help everyone alot. Besides, there is one thing that I have come to know about you. You do not like misinformation about these ecms. If you shared the data and what you know, then it'll do alot in dispelling the myths.

 

Tom, I know this is just some safety measure. But I still am trying to figure out whats going on. I did understand your drunk ***, btw.

I'm looking for the calculations that the ecm makes to compensate for a potential lean condition, how it works...

 

Abundantly clear?, maybe to you.
So what is the reasoning for your post?.

If there is so much myth out there, why not do your part and try to correct it, rather then going nenner, nenner, I know but ain't telling.

If the $500 dollar figure is supposed to be impressive, you missed. I've spent alot more then that over the years trying to forward the knowledge available about ecms.

In other words, I'm calling your bluff. ie, feel free to contribute. If your worried about your money, just think if we all behaved like you do, we'd still be where we were 10 years ago.
No flame intended.
And since pretty much everything I own is open loop, the BL issue is no more then a curiosity for me. So it's not that it really makes any difference to me what you post or don't.

 

I am somewhat cornfused by this statement (yes, I did say CORN-fused - lol). What don't you understand? The ECM simply changes the Pulse Width based on the BLM as it would do during normal part throttle stuff. It works the same way. I have included the necessary code for your review. So, to answer your question - the calculations that the ECM makes to compensate for a lean condition are located in the code below.

Tim

 

Of course I've read your whole post When it comes to calibration terms we all need to be careful. GM can set them to anything. This applies to all of the above parameters including #5.

Set that parameter to 40 Kpa and everything changes. Set it to 50 Kpa and things are different again.

As for a lot of what I said not applying to $8D, it all applies. Other then the example I gave using an upper RPM limit, which doesn't exist in $8D, what I stated is how it operates.

RBob.

 

let me correct myself I mean potental lean condition at WOT. It could'nt use the same calculation to correct for a lean wot condition. The blm correction (part throttle) seems to be small corrections. But if wot could be lean, it'll have to make a larger correction than a few blms.

I'm starting to understand whats going on. Let me know if I'm right,to this point. The ecm notices a lean part throttle condition(128+blms). So it ASSumes that wot will also be lean. It really has no clue if it's lean or not at WOT. But it makes corrections based on part throttle blms(blind programmed wot corrections),if blms are over 128.

What is these programmed wot corrections? What is the % that the ecm modifies for a potential wot lean condition.

What blm cell does the ecm base the potential lean condition on? I suggested it was the cell just before wot. But Traxion said that was wrong. If it's not basing the corrections on that, then what?


I hope this makes sense to you guys...

 

The math is:

NewPW = (CurrentPW * BLM) / 128

So a BLM of 135 will increase the PW by 135 / 128 = 1.055 or 5.5%

The cell used is the one that matches the RPM & MAP boundry criteria.

RBob.

 

That is the WOT correction? If so then what cell is the ecm using for that wot correction.

I thought that calulation is general blm correction. So that 5.5% will net 14.7. Are you sure that's the calculation for a potential wot lean condition? Or is that 5.5% added to wot pe adder?

 

... but it does. The code I posted illustrates this.

1) Why do you find that hard to believe?
2) Why do you think that BLM correction is small corrections?


What RBob stated is exactly how we calculate new VEs for part throttle too.

I think that maybe you just don't understand how fueling is calculated in the BIN. PLEASE download the hac and begin learning from it because it will bring a lot of things to light ... even though it obviously makes you dangerous at the same time (I'm the example on that one - lol).

The code uses a target AFR and modifies from there. The VE tables are assumed to produce a 14.7:1 AFR. PE Mode modifies this AFR via applying the coolant and RPM modifiers. The BLM is applied to the VE value to ensure it really is 14.7:1 and then the normal PE adders are applied.

Tim

 

I understand whats going on, I believe you guys. I'm just not articulating what I am wanting to know. I might be just wanting to learn too much. This WOT fueling correction is just for protection. It doesn't matter in the world of tuning,cause the idea is to run 128 or lower, in most cases. I guess we'll let this one go...

 

wait a minute, so you are saying that the example 5.5% is not JUST applied to part throttle? The 5.5% would be added to the PE also? If so then that is what I was wanting to know

 

Yes, that is correct, applied to both part throttle (non-PE) and when in PE. Of course if the BLM is less then 128 then it is not used during PE, but will be in part throttle.

RBob.

 

I have to admit that this discussion has been one of the most productive for me in a long time. It forced me to refamiliarize myself with the ANHT hac ... and to remember that the ANHT is not the AUJP, etc. Thanks a bunch to RBob for participating and helping out. Really appreciate it.

Tim

 

What is ANHT ? Is it a speed density bin?

 

ANHT refers to the 4 letter Broadcast Code (BCC) of the code version on the actual EPROM. In this case, ANHT refers to the $8D calibration used in the Vettes.

Tim

 

This is what I have been wondering. thanks!

 

And to think this is all about getting to 14.7 which may or may not be what the engine actually wants or likes.

I'll stick to using a WB and full time open loop, thank you