Just thought I'd share,

I used some settings from the V6 "Apab 42.bin", mainly the values for the proportional and intagrator gains, durations, and delays, also the map and tps filters. The values were drastically different than the "close to stock" V8 ones I was prevously using. Anyways now my engine is very responsive, and no more boging when its cold. And very smooth accel. This makes sense to me because V6's are fast responding motors, unlike the stock V8's.

 

hmmm this is a neat thing to find out. any idea on what exactly has been changed or is it to much stuff to go over ?? does sound interesting

 

Async Fuel Mult vs. RPM
INT Delay vs. Air Flow
Prop Gain vs. Slow O2 Err (BIN val to adjust INT)
Prop Term Duration O2 Error (time to allow O2 adjust to take effect)
Prop Duration Offset vs. Air Flow
Prop Multiplier vs. Air Flow
INT Delay vs. Slow O2 Err

I used all the exact V6 values expect for Prop Multiplier vs. Air Flow which I gave a little more gain @ 0 gms/sec and then I decreased the rest of the high flow gains, because I had a slight stumble of idle and some mid rpm surging. I'm using V8 astro's ecu file "7747-mine.ecu". I just can't believe how smooth and powerfull my part throttle is, also my exhuast sound is alot quiter.

I'm messing with some other tables right now, it seems to me that this .bin file is set up pretty good.

 

glad to hear it. maybe you should write up something along the lines of the $32 $32b article i did for 165 users. would be a greta thing to have around. im still sorting through the $32 $32b i think i found some things of interest. gotta get some well commented code to back up my findings then its off to the test. changes between the auto and manual iac gain setting etc etc etc. good stuff to know.

 

Could you post the APAB bin here so I can get it? I cannot get to the FTP site with my firewall...

Aslo, did you START with the APAB bin file, or start with a 350cid bin file and transfer the APAB data into it???

 

There is a paper over on the diy-efi site that you may be interested in: http://www.diy-efi.org/gmecm/papers/747fuel.txt

The Async Fuel Mult vs. RPM table isn't covered there. This table is used as a lookup to take the math load off the ECM. It is specific to the number of cylinders to convert a sync PW to the equivalent async PW.

The other tables that you changed have noticable effects as you found out. The further from stock an engine is the more these tables need to be tweaked. Isn't tuning fun?

RBob.

 

Heres the apap bin, I had to rename the extension .txt so I could post it, so you'll have to change it back to .bin. Also for the prop duraton O2 error @ full error I used .3sec rather than .2sec.


Rbob, I read your papers about a year ago, thats what got me chaning these values in the first place. The only thing I disagree with is in your fueling paper you say that for a fire breathing motor, the prop vs airflow gains can be reduced. SHouldn't they be increased, and the low flow maybe unchanged.

ie: A 400hp motor at 3000rpm will require a bigger delta PW to correct an O2 error, than a 200hp motor would. Hence the larger prop flow gain. But you did also make your point about gains to correct steady state surging. Anyway, Great job on the papers, It was pretty much the only place to find advanced tuning info for the 747.

The thing thats really gets me is the prop duration, I went from 4 sec's to .3sec's. WHich the prop term acting so quick, I would have thought that the process would go into sustained oscillation (constant sputering).

**** Do not use the V6 bin as a starter bin, get your VE, spark, PE, and accel enrich down with your normal bin. Then just try messing with above tables. I would say the less of a hot rod motor you have, the more prop duration O2 you need.

 

This area gets tricky. For a 400 hp vs. 200hp engine the injector flow rate should be double. This is required to supply enough fuel. Then with the same PW change the amount of fuel flow change is doubled. So it's even up.

With an additional reduction in proportional gains preventing the engine from surging. The surging is caused by the engine responding to the fueling change. A lo-po engine is not going to respond in the same manner.

Currently I've set the proportional tables to give either 1 or 2 counts of proportional gain. The gain vs O2 error is all set to 16. Then the gain vs airfow is set to 16, 16, 32, 32, 32. This causes the proportional gain term to be nothing more then something to keep the O2 sensor/integrator logic happy.

Quote:

The thing thats really gets me is the prop duration, I went from 4 sec's to .3sec's. WHich the prop term acting so quick, I would have thought that the process would go into sustained oscillation (constant sputering).

It is difficult to see the effects when reducing this value. It only comes into play when the O2 value is either changing states (R/L) or the O2 error is too low for an INT update.

RBob.

 

RBob,

What kind of motor are you running these settings on?,

 

327 SBC 30 over, L98 alum heads, Isky 280 hyd MegaCam, forged flat top pistons, 10:1 CR, Torker (original) intake, 2bbl TBI 2" bores w/BBC injectors @ 16psi, cold air, headers, 2-1/2" exhaust, manual trans, posi, and lots of tuning.

It would be better if you had questions about a particular area of tune. The stuff I run is far from stock. Wouldn't be able to see what I was using as it is all moved around and changed.

RBob.

 

RBob,

when you say, "Currently I've set the proportional tables to give either 1 or 2 counts of proportional gain."

could you define a "count".

Just out of curiosity, how many "patches" have you added to your bin, and do you think that patches can be used to increase performance, or just to eliminate annouying things.

Ryan

 

A count is defined as the resolution of the injector PW: 15.26 usec

If the PW is 3.08 msec the ECM will use the number 201. (201 * 0.01526 gives the 3.08 msec). The 15.26 usec comes from the inverse of the frequency that the ECM hardware counters are clocked at, 64KHz. Add a count of 2 for 203 and the PW is now 3.10 msec.

I don't really patch the bin any more. I assemble source that produces a new bin. Hard to say how many changes I've made. It is quite a few though.

Hmmm, "increase performance, or just to eliminate annoying things." I would say both. Although I think the annoying part is truer. That and driveability. The driveability is why I went to EFI. Functionality also. Stuff like rev limiters are nice to have.

RBob.

 

RBob,

Your using the 747 right? Just wondering what scanner your using?

And about prop duration time,
I don't know how else to say it.

Will a 6sec duration do 6x the change to the prop term as a 1 sec duration. Or will the 6 and 1 sec term both do the same change, but the 6 sec would just be 1/6 the speed.

I believe the second part of that statement to be true.

Ryan

 

Yes, a '747. No scanner, Lockers. It is an edge card setup.

The prop duration does not change the prop gain except to clear it. (I hope I get this correct) the prop gain is calc'd 40 times a second against the current O2 error level. The prop gain can change each time it is re-calc'd.

Same with the prop duration. It is re-calc'd 40 times a second. If the prop duration timer expires the prop gains are cleared (zero'd) for that pass. Of course 25 msec later (40 Hz) the prop gains are re-calc'd and reapplied.

The prop duration timer is cleared whenever an O2 x-cnt occurs.

Now it gets tricky. As long as the prop duration timer has expired AND the O2 error level is less then that required to move the INT the prop gains are held at zero. And note in the cal how the prop gains are very small with a small O2 error level. So how much of a difference does it make to zero out the gains that are already minor?

Now even if the prop duration timer has expired as long as the O2 error level is enough to move the INT the prop gains will still be applied.

The end result is that the prop duration has very little effect on the whole scheme of things.

RBob.

 

Thanks for the Info

 

RBob,

Could comment on "slow O2 filter time constant, and the prop and int delays"

Ryan

 

wow...

once again I'm amazed...

 

As you look at the fueling cal & code it is apparent that the airflow through the engine is used quite a bit. In the '747 this term as grams of air per second is not accurate. The calcuation used is quick and simple. It does provide a usable relative value.

It is interesting to note that the gms/sec term is used to denote two different items: the airflow (as a transport delay term) and the BPW. The higher the airflow the lower the transport delay and the greater the BPW.

When the gms/sec term is used to lookup the INT delay it is as the transport delay.

When the gms/sec term is used to lookup the proportional gain multiplier it is as the BPW. The larger the BPW the greater the prop gain required to change the AFR.

The integrator delay is required to prevent overshoot and oscillation of the fueling. The delay is to give the O2 sensor time to see and respond to the change in fueling (an INT change). It takes a finite amount of time for the change in injector PW (change in amount of fuel) to travel through the engine and out the exhaust. There is also a delay within the O2 sensor.

The table at $48F, INT delay vs airflow, has less delay as the airflow increases. If the O2 sensor position is moved further downstream this delay needs to be increased. This is the transport delay term.

The table at $4CC, INT delay vs O2 error, has a two fold function. The first is to deal with the response rate of the O2 sensor. With a large O2 error level the sensor will be at the bottom or top of the ZS shaped O2 sensor response curve. Large delays are required. As the O2 sensor starts to get into the upright portion of the response curve the INT action is sped up.

In the second case this table can also provide a derivative term to the PID loop. In the case of the ASDZ cal this is true. As the O2 error decreases the INT delay is increased. This effectively stops the integrator from updating. Thereby putting the brakes on INT fueling changes.

Slow O2 error later. . .

RBob.

 

Thank you and you are welcome.

RBob.

 

Very good info!

 

I've got an Idea!!


It would seem to me that with large delays to the int at low rpms, that if your VE cells (747) were out a little, that the prop term would simply take care of this error before the Int could act. This would meen that the BLM wouldn't change, therefore you wouldn't change a VE cell. You would think your tunned when your not (with a scanner like winaldl, which I doesn't show prop counts or PW).

So... couldn't you just lower the prop gains to the lowest value above zero, and maybe decrease the int delays vs. airflow. Then the int would take more control of the error, thus changing your BLM.

It would seem to me that this would be a usefull way to tune very accurattly. Then once the VE cells were nailed you could go back to you old prop and int settings.

I might be way out to lunch here, but I got this Idea when I increased the int delays. This caused the BLM to pretty much lock @ 128.

Ryan

 

Another Idea might be..

You could max your int delay right out, and max out the O2 counts for int action. Then go out and run and see how the prop term reacts. If it seems to climb slow and steady from say .300 v. Then you would know you need to pump up the prop gains and maybe increase the duration.

Ryan

 

Proportional gains can make tuning via the INT difficult. It is very easy to have an INT of 132 (adding fuel, +4) while the prop gain is -10 (removing fuel). It would seem that area of the VE table is lean when in fact it is rich.

As the proportional gains change so quickly it can be difficult to use them for tuning. Also need to have the R/L status bit at the exact same time. Otherwise the prop gain is meaningless.

RBob.

 

I'm starting to understand, I think the INT delay might be key to my crappy closed loop idle. I relocated my O2 sensor about 2 feet down from the OEM location when I installed my headers and duals.

I now feel kinda lost since I don't have a wide band O2. I kinda don't want to mess with any of this since I now see that the BLM and INT may say one thing, when in fact the ECM is removing fuel.

I have however learned what injector pulse width numbers are normal for my engine. I'm "cheating" with WinALDL so I don't have PW in msec, it's unconverted decimal. This was very useful in determining what was causing my idle surge.

 

As far as i can see there is no 'gain vs. O2 error'

4A8 PROPORTIONIAL VALUE vs SLO o2 ERROR
4B5 PROPORTIONIAL TERM DURATION o2 ERROR
4C2 PROPORTIONIAL DURATION OFFSET TBL vs AIR FLOW
4C7 PROPORTIONIAL GAIN FLOW FACTOR vs AIR FLOW

care to clarify ?

 

The table at 4A8 is Proportional Gain vs. O2 Error.

RBob.

 

gain = change in output / change in input

so for 4A8 your input is O2 mv, and your output is a term that changes your pulse width at your injectors.

So bigger gain (values that you enter in this table), means bigger change in your pw, and smaller = smaller change.