I've noticed that the resolution for tuning fuel and spark tables leaves a little to be desired with the 1227747 ecm.

Here in Colorado Springs (6000 ft) atmospheric pressure is roughly 80kpa and my car with its 219/227 112 LSA cam idles at about 62kpa or 11" vacuum.

Has anyone used an op-amp or perhaps some sort of inline processor (perhaps with a pic programmer) to scale the voltage coming off the map to regain some of the resolution used by the the map tables in the ECM? Other than these tables, what other considerations can you all see that should be taken into account.

 

The only right way to expand the tables is in the code, sad but true, I've tried all kinds of voltage dividers, and things to bias the signals, and there is no real good answerr.
One of those things that well everyone else has to put up with it.

I would check everything else out first to makre sure things area aas good asa possible, some lettle gotcha can be a PITA

 

Nothing really to check out... best idle I can get out of it with the altitude is 11"... and atmospheric being 80kpa, only leaves me 3 cells wide to tune the entire driveability range.

Can you give me specifics on the experiments you tried and why they failed.

 

I would think a maf system would be more viable in your situation.

 

I just did some checking on mine-at 7000ft I have 11in. vac and at about 1500ft I have 16in. vac. With a 615 (maf) I do start at cell 00 and progress through cell 15 as rpm and gps go up. So it appears that maf does have a plus to it if you have a cam and are dealing with high altitude. Although I'd think that the latest gm that uses both would be the best of both worlds-too bad that doesn't look viable for us!

 

Two things to remember here. He has TBI so a MAF system would be cumbersome to create. And by him having TBI he doesn't have the same control over his idle as you would with TPI. Now if he went with a 730 TPI system he would probably have the best system available. The second thing doesn't really matter but I might as well say it because according to lots of replies on the GMECM list, the modern cars that have both MAF and MAP are only running on MAF and using MAP as a sanity check for the MAF readings. They can be converted to disregard the MAF and only use the MAP (SD mode) but that again is something else altogether.

 

I wasn’t aware of any issues with 747 idle control VS 730... could you go a little further there please.

As far as MAF goes, its a system I really like (I have a mustang also and its a great system) but its not really practical for my TBI application.

Does anyone know of another speed density ECM that offers better resolution?

 

The reason I say a 730 ecm is that it is used with TPI and MAP. I believe 8 smaller injectors that aren't too dependent on intake tract for atomizing are easier to control at idle than two larger ones. And of course when it comes to high rpm the TPI is much better. Also, and this is personal preference, I believe the VE tables are easier to tune in the 730. You can use MAF and a 165 ecm but you will have to make the intake air tract and still go to TPI as there is only one injector driver and then be limited to the 255 g/s limit. Either that or adapt the ECM to handle another driver to go with TBI and make changes in the source code to handle the differences but for all that trouble you might as well change to TPI. I know this doesn't help you any with your original question of how to make the car idle better with a TBI and I'm sorry if this went off topic but I was just replying to the other replies. My only trick to try and solve your problem would be to raise the idle in search of better vacuum numbers and hope that helps. Either that or make the cam LDA wider and get a better idle there. 112 is good but as a reference most factory TPI cams are 117. Not much help, huh?

 

Ahhh, I see where you were going but I think maybe I didnt define the problem quite well enough. I'll clarify, please dont take offence.

There is not an idle problem in this case... injector pulse width resolution at idle can be an issue with large injectors (this must be what you are referring to) but its not a problem in this case. In fact, it runs great at idle and great at WOT... the problem is in the area in between. Only one cell between the two extremes. Please tell me when you said: when it comes to high rpm the TPI is much better, you meant multi point and not the long runner intake thats mismatched for high rpm applications.

Truth be known it really doesnt drive that bad. Spark table resolution in the 747 is in 5 kpa increments so its better than the fuel table... and if the 70kpa cells are somewhere centered for its fuel requirements the blm correction can help to clean things up if the load is sustained... the a/f ratio never gets far enough out of wack to cause real driveability problems.

As far as tuning goes, I cant get as aggressive with spark as I'd like in the mid range because I have to tune it knowing its going to be a little lean at times at higher loads (and rich at lighter ones). There is just no way that I can nail down my blm's and do high quality tuning in the mid range with this lack of resolution.

So in any event, the problem is not with idle. heck, it idles at practically the same vacuum my stock mustang does. The problem is the lack of load resolution in the fuel table above idle and below wot.

Now when you mention the 730 as being easier to tune, I think were getting somewhere. Why do you say that? By what factor does the 730's fuel table increment? The 747's fuel table goes in increments of 10:
20,30, 40, 50, 60, 70, 80, 90, 100 kpa.

 

Well it's my opinion that a dry manifold is better. Let the valves do the atomizing. Less fuel is used and it's better mixed. Also, the theory behind TBI is to use the throttle blades to help with atomization. A TBI is only better than a carb because it can be electronically controlled not because it's a better fuel atomizer.

OK, VE lower table is from 20-60 kpa in 5 kpa increments and 70-100 in 10 kpa increments and rpm from 400- 1600 in 100 rpm increments. VE upper table is from 20-100 in 10 kpa increments and rpm from 1600-4800 in 400 rpm increments. The PE fuel table goes up to 6000 I believe. Pick up a version of winbin for kicks and see all the other neat stuff. Gotta go but I wish you and all the others a Happy New Year. I'll get back to you tomorrow if you think I can answer any more questions.

 

hector, tbi is a little more advanced than a carb with electronic controls. Tbi only relies on tB blades at idle to assist in atomization the rest of the time the blades are pretty much out of the equation. A carb relies on airspeed through the venturi to introduce fuel, thus its very sensitive to the vacuum signal the engine delivers to it. TBI doesnt care, it will inject fuel in free standing air if you wanted it to, at low airspeeds through a carb the fuel atomization and cyl distribution goes to all hell. You can get a carb to really smoke any injection system in terms of atomization but for a very narrow power band. Any form of injection is gonna display a much better average across the rpm range, TBI included. You could argue TBI is better because it has less injectors. Minor intake/exhaust tract differences and injector flow rate variations make a port injection system succeptible imbalanced a/f ratios for different cyls since each cyl recieves the same amount of fuel regardless of the air entering it. In addition, a TBI system can take advantage of tried and true wet flow manifold technology that has been under constant refinement since the beginning of the automobile.

 

Just to get a bit back on track as this isnt intended to be a thread of TBI VS TPI. I'm well aware of the advantages and disadvantages of both systems and the difference between a theoretical disadvantage and how things work in the real world for both systems. This thread is about improving the resolution of the main fuel and spark tables for the 747.

I'm still looking for a reply from bruce here about his experiments and specific problems with scaling voltages from the map sensor...

I've gone through the tables and compiled a list of those that use map voltage... I may not have got them all but at least these areas may need to be addressed:
Main Spark, Main Fuel, Coolant compensation Spark, Pump shot vs differential map and knock retard vs map.

I've been through the hack a couple of times and nothing that would prevent this from working has jumped out at me. Again, I'm very curious why Bruces work in this area failed...

 

I agree with Dan, Pablo. This isn't a discussion about TBI vs TPI vs carbs. However, I do know a few things about carbs and such and a carb will inject fuel in free air also. It's called a pump shot which is the only time you have free standing air, in the transition to WOT. There are other things we could discuss about your post if you like off-list and since we both live in the SoFLa area maybe we could even hook up some time. Actually, I do some autocrossing down at Metro Zoo with Equipe Rapide. You should try it if you haven't done so already. I have been around cars for a while and I actually work on them for a living so I can talk cars with the best of them!

 

gotcha, im temporarily living out of state for the next 4 years being that I joined the military. You shoulda stopped by bout 7 months ago when i was around all the time. Definately woulda been cool. I go back all the time though. Ill let you know next time im down

 

I'd strongly suggest catching up on TBI vs carburation operation.
Reversion, pressure waves in the N/A manifolds are and always will be a problem, At least with TBI you can try to min the reversion effects.
If you look at the best of manifolds (N/A), you'll see the the Independent runner (weber) when tuned right will always make the best Torq/HP. If you go to an extended Injector you might maybe pick up a couple %, but this is at the limits of best tuning.

Dry or wet manifolds all suffer the same for cylinder filling air wise. Look at the fast cars and you see how they are combining the above ideas.

Wet flow and dry flow manifolds have both been around almost for ever, if Detroit was still trying to get it right they would be coping the MB 300SL system from the 50s.

And no the atomisation that takes place at the venturii is not the beginning and end of when atomisation is taking place. It's the atomisation of the fuel laying on the back of the intake valve getting the blast from the exhuast charge are overlap that atomises the vas majority of fuel and gets it ready for in cylinder burning

Feel free to look up N.A.C.A. paper and start reading there.

 

There is no big miracle to be found with a higher resolution map, the numbers are interpalitated (?). As a matter of fact the large MAP tabled ecms can be more of pain then the smaller ones.

If you look at like some of the really expensive aftermarket ones they use the same size tables as the oems.

 

The engine doesn't care how many cells it's working in, set the tables for the best you can do and be happy with it.

I rigged several dozen 4066's and ron them off the TPS so as to run a *smart* Alpha-N system. After months of work I was able to get it almost as good as oem.

The switching and blending has to be very fast. Some drivers can only sense a .1sec change in pedal response, some .05, I happily had to be a luckly one that easily senses .03 so it takes a high level of sophistication for me to be happy. Also the TPS needs broken you into 2-3% intervals,, I think I had one point 20- 4066 for TPS and 40 or so for MAP switching, plus related logic. Then a whole mess of resistor dividers to get the make values, and the changes had to be made by changing resistor rather then using a simple hex editor.

We won't get into the vac switching ideas.

 

Could you please expand on this.

[QUOTE
The engine doesn't care how many cells it's working in, set the tables for the best you can do and be happy with it.
[/quote]
I understand it like this, correct me if I'm wrong: the engine goes from one one area that has a lower VE then up to place in the original table would have been 4 cells away. The result is a change that caused the vehicle to run extreamly lean for a short period while the BLM catches up with the change since the BLM has even less resolution than the fuel table. This lean period can cause hesitation backfires, etc not to mention issues related to detonation during that period when the spark is all over the place. Without better resolution the car is essentially running off the oxygen sensor.

I have little knowledge about Alpha -N... Basicly I know it runs off the TPS instead of the map. Have others done Alpha N work on the 747?

If when you say vac switching you are talking about scaling modifying the map voltages the ECM sees, thats exactly what I would like to talk about.

If your worried about timing, from the math I've done and the statistics on parts I was looking at, the whole process should take much less than .001 second, possibly more like .0001 depending on how long it takes the processor to make the conversion. For example, one of the slowest parts is the a/d converter an its signal converson rate is 13usec.

 

It's all a waste of time.
What you need to get into is the code end of things. While the calc might take a nanosec, the processing and filtering makes it milsec long.
If you want to pursue this, have fun, but, it's a waste of time in my book.

 

It took me a while to decide to respond to this... but what to heck. I didn’t ask if it was a waste of time. I was asking for specifics on what you and others had already tried. Please give me the courtesy of letting me decide for myself what’s a waste of my time. If you have info on exactly what you tried and what the results were, I can use that as info to make that decision. "It wont work" statements are not exactly constructive w/o the specifics behind how your opinions were formulated.

While I have a great deal of respect for you Bruce for all the work you have done in this area I don't believe any of us are infallible or that we see solutions to every problem that comes up. In other words, I understand that you were not able to make this work. I am looking at it with a fresh set of eyes and would like to know the specifics of others work so as not to waste time repeating their efforts.

It almost seems that you take offence when asked to share what you already know. I know that you are working on a project with others to understand and modify the program. I know that once done, it would be a cleaner way of accomplishing the same thing... but I'm not interested in reverse engineering code right now.

Anyway, I'd very much like for you to share specifics of what you have done and the results of the experiments if you so desire.

 

Unfortunately, sometimes the only "real solution" is within the instructions.

It's like guys I've come across on the other Boards. They advance the distributor and find they go slower because they triggered the knock sensor and it's pulling timing out. So they next disable the knock sensor and discover they are "pinging" so they back off the distributor. Now, they find that with the knock sensor disabled they are even having "ping" at the stock base setting and are having to retard the distributor significantly to eliminate the "ping". Now their car is even slower.

We both know that they are using "mechanical methods" when the real solution is to alter the spark tables in the eprom. Should I continure to advice them on how to do these "mechanical methods" when I know they won't work properly and the real solution is to work with the spark tables in the eprom?

Source code is like that when you compare it to "simply adjusting tables" in the eprom. Once you start getting inside the instructions, then a whole new world opens up. Things that can only be "at best cludged" can be fixed properly within the instructions.

 

Glen,

What I'm looking for is very specific. My question is not at all like your example but your answer (the tail chasing portion of the example) is something like I'm looking for. Specific reasons or outcomes to specific paths... if you do this, this will happen.

About mechanical solutions... some things need them and I think this may be one case... after all, this is really about a mechanical limitation, (the reduced operating window of the map sensor due to a larger than stock cam and altitude) The ideal solution would probably be to have a map sensor that worked in a 50-100 kpa / 0-5 volt operating range.

I'm not 100% sure that the fix for this one is in the code (although alpha N could be an interesting way to go)... its possible but I see problems. Take for example your current voltage from the map goes through an A/D and is converted into an 8-bit value. Thats all the ECM has to work with... ever. If you alter tables in the ECM for more resolution you are working within a decreased resolution due to the a/d. In my case the map range is 2.7-3.7 volts... converted into 8-bit actually gives you 20% of the original resolution from the ECM's a/d with which to work with for your operating range(less than 6--bits). Now I don't know for sure but this could create some problems. Converting the voltage before it goes to the ECM would allow the ECM and the tables to use the full resolution of its 8-bit a/d conversion.

 

Using the big cam example. Yes, if you ONLY looked from idle to WOT you would only have a range of 50-100 kpa. But when you consider light part-throttle driving and deceleration, I can tell you that you will still get some lower MAP readings, possibly in the 20-30s in deceleration. How would you be able to handle those?

 

You can make an op-amp circuit that will rescale the output voltage of the MAP sensor. You may gain some resolution from that. The type of amplifier you need is called an instrumentation amp, and basically has the formula of A(v2-v1), where A is gain is Volt/volt, and V2(MAP input) and V1(offset voltage) are inputs to the amplifier. Briefly looking at the scaling, from 10kpa to 100 kpa, you have a 0-5 volt swing(roughly). Changing the gain and offsetting the voltage, you can get roughly a 0-5 volt swing from 20-90kpa, using a 1 volt offset and gain of 1.5. For each unique situation, you could change either, or both, to get the range you need, and then scale the prom tables accordingly. Be careful! The output could(and will) go above 5 volts and below 0, so you could damage the ecm if you operate above or below those pressures.(there are ways around that though) The safest thing to do is to get another map sensor in parallel, wire it up, and monitor while you drive, to make sure you woudln't damage the ecm. If you'd like more info, do a search on instrumentation op-amps on an IC manufacturers website, like linear-tech or maxim. Hope that helps.

 

I dug this up while searching around, but it's only a month old so i thought i'd add.

It occurs to me that he doesn't mentioning changing the boundaries of his BLMs. He mentions that he has less BLM resolution than VE table. Well since he appears to only 'use' 3 VE MAP values, then why not alter the BLM MAP boundaries to correspond.

Then the BLMs wouldn't have to accomodate quite the same spectrum...