Has anybody adapted a 730 to run in alpha-n mode?
My friend just got a Manta and he wants to EFI it with hilborns, no talking him out of it. I'm just looking into options that are cost effective. I see 2 different methods of converting a 730 into alpha-n so if anybody has any other ideas please reply with such. First is to disable a bunch of tables and switch the TPS into the MAP pin on the ecm resulting in an RPM vs TPS fuel table. Another idea I have is using the map for baro only, then changing the lookup for the VE table to the TPS input. Then disable the tables that would cause headaches. I would think the first method would work well enough for a mild 350hp 350. This car won't be driven daily but it will see highway and road racing.

 

Sure, why not.........

If you have to go that way, think about the 58 code.
It has AE by RPM increments.
Losing the MAP AE might be a concern.

Thing of it is, you have to be soooo conservative with timing to keep it out of detonation.
Can you run a knock sensor?. Maybe tinkering with that would help with the timing issues.

Might also look at the 808/165 box.

 

Cool, thanks. The AE from RPM table would be nice, it would be even nicer if it were in the $61 mask... maybe that's why the idle likes to be lean, hmmm.
What is in the 808/165 that would be useful?

 

Why not provide a central MAP reference point, from each runner just run a small hose to a junction block and keep the MAP. It seems easier (at least to me) to modify the manifold than to get into some one-off custom code modification. I have seen this done before on intakes with no common plenum.

 

Really simple code.
When you want a basic system SOMETIMES, the simpliest matching answer is best.
8D code is kinda like using Artillery for Ground Hog Hunting, for an A-N, IMO.

The other thing is maybe using the 6E is without a MAF. Most of you 6E cars run really well with the MAF unplugged. Running a 6E on a bench would probably tell you what you need to know about doing an A-N with a MAF system. I'm just not sure if without the MAF if you can still tune much with it. Again, simplier code to work with.

Then on the 808 stuff they seem to be able to play all sorts of games with them with good results.

 

Might try reading up on the megasquirt system. ......I think it might work.

 

Thanks for the ideas. The common point for MAP was an idea we had but I'm worried about the signal with the individual throttle bodies. The system should run pretty good with alpha-n so I'm going to try it first.
Before I make and get any of the parts I'm going to check out the cam profile. Any suggestions on the cam vs what system to use? I don't know if the cam profile would make much of a difference but if it has a large lobe seperation I think it would be even harder to get a good "map" signal.

 

My v6 idles at 70 K/Pa, and cruises at ~80 K/Pa.
Hard to have much less then that, and it runs just fine. The answer in part is table resolution. Considering I have 30% fewer pulsations, the signal is all the worse then on a v8.

 

I think I'm going to do some Virtual 4-stroke simulations and try and get some idea of what the pressure pulses are doing.
Has anybody seen a setup that used TPS with Vehicle speed to calculate theoretical engine load in a table? I think that would be the best way to go about this.
Bruce, what experience have you had with the alpha-n systems? Anything with GM electronics?

 

You got a CamAm car with a speedo?.
You need to know which gear, speed and TPS to do much of a quasi Load calc., IMO.

I did a quasi A/N with the original GN code for when the MAF pegged.

 

I have run a GM ECM in N-Alpha mode by adding some tables and the code to support it. The method was to create a MAP term from a few table lookups. The main table was a 3d table with TPS% and RPM as the axis. The table data was a MAP value. This table was filled with empirical data pulled from log files.

Another table was used to compensate for the IAC airflow. This compensation value was then adjusted for RPM.

The resulting MAP value was used in place of, or blended with the actual MAP sensor input. All-in-all it works quite well.

I found an interesting paper that may shed some light on what you are attempting: http://www.simcar.com/literature/sae.../sae940759.htm

BobR.

 

Yeah, he's adding a speedo to the car for some strange reason. I told him to just memorize gear/rpm to derive speed, but he's terrible at math (art major).
I don't see why gear is needed. Alpha from normal closed, engine speed, and vehicle speed should be enough to derive a load and vol eff. Assume engine speed and alpha as constant, higher VSS = higher load, lower VSS = lower load.
Bob, thanks for that link, good reading material.
Now here's a question, anybody ever figure out a way to control fuel delivery through closed loop with a wideband and an AFR type table with RPM or TPS as an axis? I've been here a while and if it slipped through my fingers then please post a link. From what I gathered, it'll be a source code project. I think this whole thing might end up being a source code project, lol. Then I can run the alpha-n type style for quick transients and under slow TPS transients, use the wideband.
Thanks for the ideas and suggestions.

 

What might get you by on a 305, is one thing, a CanAm motor is an entirely different animal.
The areo loading and pulling high gear at 180 MPH is an entire universe away from having a 2.5:1, 1st gear, at 25 MPH,
at least in BruceWorld.

How about swapping a few wires around in your car, and play with a sorta A-N, and then see what you have to say about things?.

Maybe, machine a recess in one of the stacks, and install a late TB mounted v6 MAF, in it.

 

Hmmm,
how about EGT for a sensor for load?.

The real problem is going to be WOT at lower revs, in the various gears, IMO.

Maybe several time in PE fuel adders, or PE adder per amount of TPS/RPM.

 

A maf? Do you really think it will register a smooth output being on only 1 cylinder? This motor is a typical sbc, 300hp 5600rpm double hump heads with edelbrock torker manifold carbed (but not for long).
The EGT sounds like an idea worth pursuing. I only fear that the changes seen will be induced by the sa and fuel than actual load. When we get our large dc electric dyno I'll be able to motor the engine down and watch EGTs on a sbc, that'll be a fun project.
So with the v6 maf, how would you use it and with which ecm/code?

 

Ya know, maybe the simpliest route would be best.
If this is making some really serious HP, then it's going to be climbing thru the gears, pretty fast, and as long as you have sustained WOT right you might be fine.

Something like the GNs have with a PE timer, for fuel enrichment, might be the hot ticket. If he uses the clutch for shifting.

With the 60 PPH saturated injectors being available now, that should be good for like, 950 HP.
What do you think he'll be looking to make?.
If much more then that, he'll be limited to the 749, since that's the only one I know of where it's been public domained on how to change the injector drivers in it, for P+Hs.

The MAF I was talking about is just a hot wire element that fits into the air stream, not at all like the 3rd Gen types. A simple peak detector circuit, should suffice. Just need to have the decay right, on the sampling.

One other ecm to think about is a 165. They run pretty good with no MAF, and in soft fail mode. IMO, it'd be worth hooking one up and seeing what exactly the timing does, and then all you'd have to worry about is the default airflow numbers. Or possibly run say a 747 actively, for timing control, and then the 165 for fuel only in soft fail.

Of course with an ecm with Source Code, things would get really easy.........

 

Ya know, this could be the basis for a ground up source code project. I wish I knew enough about the code to work on something like that. Basically, it seems you'd have just a table for fuel that you would tune using RPM vs. TPS, then the same for spark, then a couple tables for spark and fuel AE using just a delta TPS %. You probably wouldn't even be too worried about coolant temp, as long as it runs pretty constant.

I may be way off here, and may be trying to simplify it too much, just throwing out ideas!

 

GM with all their alledeged faults, has some software that has millions and millions of miles of actual useage on it. For the code to run as well as it does, under all the conditions it does is guite a feat.

Developing code to run, and finding out what you need takes an incredibly long period of time, and countless hours of tweaking and experimenting.

It is much, much, much simplier to start with existing code, first understanding that, and then modifing that for a specific application.

It takes some complexity to make it simple enough for it to run well. You can have code that runs, and easy to work with, but if too simplistic, the drivibility just goes to dodo. Testimony to that is just looking at all the new aftermarket stuff. It's all gotten more complex.

Now with that said, on some full on race cars, they transistion thru things so fast, that huge errors aren't that noticeable. Which is what this Mantra is probably going to be like, and then we get to the issue of application. A full out drag car, with an auto, with 800 HP on tap that weights 2,000# is not going to need the sophistication of a 300 HP 3,600 grocery getter.

Yes, I went tangent again......

 

I totally understand what you're saying Grumpy, about the GM code being pretty darn good, and I guess I was more thinking along the lines of a stripped WAY down version of GM stuff. But instead of going in and removing everything you don't need, starting blank and adding the basics that you would need, that way you're sure to not have something in there you don't need that you may have forgotten about, or not noticed, that may give you problems.

With that said, I'm off to look at code again!

 

It's easier to go item by item and comment out what you don't need, or THINK you need, then delete it and find out, you do in fact need it, or that it was the right place to hook something in.

Properly commented out, it's dead.

 

I see your point. Now that I put more thought into it, it would be easier that way.

I'll shut up now, so that this can get back on topic.

 

Conversation geminates thought, and more concise expression.
There's lots of nuances to code.
1/2 of what I rant about, is for me. ie so that I learn to better express what I'm try to say, and better define strategies.

 

So here's the deal, I think we're going to do it. He's going to get the intake and we'll make the injector bungs, weld, and go at it. We need to take a trip to the junkyard to get some parts but that shouldn't be a hold up.
I've now tuned 3 engines all alpha-n and it's going really well for me. The only real troubles I ran into is that our little engine needs more timing than we can give it but didn't have the time to advance the trigger wheel .
The one thing that's hard to do with a dc electric dyno is get a fast acceleration run. The best I could muster was a lengthy pull from 3000 to 12000 at 100% resulting in some funky lean and rich spots in the pull.
I tuned it from 3000 with 1/4 throttle and just moved straight up with throttle to 100 then back down slowely, then jump wide open and close to check AE. The PE-ltd ecu is okay but the major drawbacks are it's resolution. I only have a 15 x 15 fuel map (every 1000rpm) and fuel trim is in the 625us which SUCKS.
Wish me luck.

 

I'll be even more specific and wish ya GOOD LUCK....
I hope ya keep us all posted...

 

I thought alpha-n mode was a bad thing to run in?????

Now we are encouraging people to use it on applications?

 

Read the whole post. It goes into detail as to why this is a good option and why it should run very well if done correctly.

 

I'll keep everybody posted and maybe when we're done we'll take it for a road trip this spring.