I am working on replacement for Quadrajet engine control module.

I have a 1980 California Corvette, the first model that had computer controlled Quadrajet. My goal is to create a small circuit board with a microcontroller that will be able to control air/fuel mixture more efficiently than original computer. It will be compatible with 81 and 82 year models, maybe also with other cars that have E series Quadrajet.

After having a lot of trouble with original ECM I started investigating what prevents it from performing efficiently. I found too many problem with that old piece of ancient technology. Some signals are too noisy, some are out of range. Old computer is simply not able to compensate for that. Modern technology provides variety of ways how to interpret dirty and inaccurate signals and how to calculate control signals efficiently.

At this point I have a microcontroller, signal filtering/processing and power supply modules, microcontroller program to read and interpret sensors data and computer software to display it.

I have also found the way to control the carburetor, but I have not finished the algorithm to calculate the control signal.

Here are some videos of my the first test in a real car. They demonstrate sensors data visualization and carburetor control:

https://vimeo.com/92306680

https://vimeo.com/92306768


What you see in these videos is not the final result, it is just a first working prototype built at home. Final result will be printed and assembled on a single small circuit board and will have same connector as original ECM (so it will be possible to put it into old computer box).

You may notice on first video that my carburetor was a little bit out of adjustment: on 50% duty cycle exhaust was too rich (almost no oxygen). Right after taking the video I fixed that by screwing in fuel adjustment screws till I got exact 50% sensor reading at 50% duty cycle, my carburetor works much better now. I found this system very useful for adjusting electronic Quadrajet.

I am also planning to replace exhaust oxygen sensor with modern heated sensor and to add manifold air pressure sensor. Both are needed for calculating optimal air/fuel mixture.

I will also add a bluetooth module to communicate with mobile devices and computers. It is possible to implement an algorithm that will learn driving habits and will adjust the mixture for maximum economy or for best performance, but this requires a lot of processing power, so ECM will be communicating to mobile phone for calculations.

So, my plan is to create auto-calibrating self-adjusting ECM that will provide better fuel economy, emissions and performance than any other Quadrajet setup and will friendly provide interface to read it's status and make adjustments.

Is someone else interested in device like this? I am looking for new ideas and prototype testing.

 

A2----very interesting project you have going on---but---I have to ask-----WHY??

it seems like a ton of work to end up with a system the will still have limitations, don't get me wrong, it will be better than the original but without a MAF the system can only make mixture adjustments "after the fact". dollar for dollar, you would be better off swapping in a TPI setup from a bit newer car, or if you want the original "look" under the hood, go with an aftermarket TBI setup.

don't get me wrong, if you're doing it "for something to do" then go for it

keep us updated on your progress

 

I think what you are doing is terrific. I know that when I get an idea for improving my car, I put more time/money/effort into it than it's worth, MOST of the time. But, it's a labor of love, just to see what I can do with stuff that was designed 50 years ago.

Now, the 'business model' for doing that kind of work is akin to "slim, and none". But, WTH....

If you can develop a circuit that will do the same job as the CA '80/ US '81 ECM and improve performance and/or economy....SUPER!

I'm interested in following this thread to see how it goes.

 

I've thought about doing this as well, have written a few small pieces of code for polling a TPS sensor, etc. Got a little bogged down when it came time to do the work for reading an O2 sensor or firing the MCS, as I don't have the electrical engineering know-how to set up safe circuits for those to interact with my board.

I'm not sure why you think a MAF is superior to a MAP, they're both based on calibration estimates and if either is calibrated incorrectly will require "after the fact" corrections. Once the base tables are correct, either solution should be spot on.

Building a microcontroller to control an E4ME quadrajet will be *SIGNIFICANTLY* cheaper (5-10% of the cost) than going with one of the aftermarket setups. It also involves a lot less work. Any vacuum components will have to be refactored, new fuel lines, fuel pumps, etc. Obviously building the system doesn't account for his time, but if he is a proficient real-time system designer, he should be able to build a prototype in about the same time it takes to convert a carbed C3 to EFI.

 

Jig A Low, I am going to install MAP sensor into the carburetor filter manifold. It not as accurate as MAF sensor, but it gives pretty accurate air flow estimate that is more than sufficient for calculating MCS control signal. It will not be required for ECM to operate correctly, but it will greatly improve accuracy (noth economy and performance). It would not be easy to install MAF sensor into the air intake that I have in my car.

7T1vette, thank you! It is labor of love for me too. My initial goal was simply to get some experience with microcontrollers.

Shark Racer, you are right. In case of mass production the cost of all components, circuit board printing and assembly should be below 20$. This does not include the cost of design, development, testing and fail-proofing. Currently I have code to pull data from all sensors, but it requires some calibration that most likely will be specific for every car.

The biggest problem when developing engine control system is to guarantee it's reliability and safety of car electronics in case of failure (humidity, heat, overvoltage, short circuits are just some of problems that will affect ECM regularly). ECM should never send overvoltage signal to sensors, should always be able to light "check engine" indicate in case of failure. There are a lot more things to think about before it can be used in a daily driver car.

I am also planning to add cruise control module and speed sensor (using mechanical-to-electric converter for my TH350) to make cruise control more accurate and adjustable. I have some ideas like integration with GPS for adjusting cruise control according to speed limit and stereo cameras for detecting distance to car in front and adjusting to it's speed.

Currently I am working on engine simulation model to use for development and testing of mixture control algorithm. This will take a lot of time.

 

sounds cool.

 

If it were me and I had an '81 and I didn't want to go non-ECM (ie. '80 Carb and distributor), I would opt for the new FAST EFI 2.0 or comparable system. it means new fuel system but from what I have read, it is a quality set-up.

Does '81 have in-tank fuel pump?

 

TedH, it is all about price, looks and emissions testing when comparing to EFI. Clean and well tuned computer controlled Quadrajet can be just as efficient as fuel injection.

 

True. If keeping the CCC, I'd check with the rebuilders to see what options there are to rebuild and improve performance. I love my 34-year-old carburetor!

Wishing you best of luck with your endeavor!

 

To me, this sounds like the easy part. I was thinking of stubbing out a simulator in java and then porting the business logic over to C.

 

Greetings,

I am working on the exact same concept, an alternate ECM for an electronic quadrajet for my computer science thesis. I have a 1979 Oldsmobile Omega (not a CCC car) with a traditional quadrajet. My goal is a functioning ECM controlling an electronic quadrajet. The sensor suite will be O2. MAP, TPS, and coolant temperature.

I am an electrical engineer working on a Masters in Comp Sci, and a Motörhead. I am happy to help in any way I can.

To those who ask why, I say why not? Improve upon the finest carburetor ever produced with technology not available to GM in the early eighties. Plus it is fun.


Michael

 

Michael, the original poster hasn't been on the forum in over 5 years.
You may want to see if you can track him down through the vimeo links he posted.

 

I’m still here, more or less. I’ve sold my Corvette and don’t have any means of testing an electronic Q-Jet setup anymore, though.

The GM setup was pretty mindless. I can’t imagine the software logic being greater than 1000 lines of code. I’d have to storyboard it, though.

 

Good luck with your project!

Please post a photo. The skeptic in me thinks it is just as likely that your thesis is in writing an AI that can subscribe to, and post in, car Forums.

Either way, good luck!

 

Michael, I am interested in this idea. I’ve never thought about something like that and I could be a winner. Please keep us posted.

 

I will definitely storyboard it. I am using an 8085 microprocessor and writing the code in assembly language. I hope it is a lot less than 1000 lines.

 

I think something like that could be interesting on a older car as it could open up some different possibilities

 

By the time I got into it, I'd want to do more than what the GM ECM likely did. There is a diagram in one of the service manuals that suggest it just raises the solenoid cycle rate (or lowers it) based on AFR readings, causing the needles to essentially dwell in a given position. You also have to control ignition function in an 81 or later car.

At idle, it will behave the opposite - you're hanging an air bleed more open or less open depending on what the ECM wants.

Basic subroutines would include:
1. Idle
2. Off-idle
3. Cruise
4. WOT

And you'd also want to have modifiers for acceleration enrichment and deceleration derichment. Also altitude and temperature compensation modifiers. I don't think these early systems had IAT, though you could add one pretty easily to the air cleaner. That should help with smoothing out some drivability.

That's without thinking about how timing impacts all of this.

Will you do wideband or narrowband? You'd need to go wideband if you want more efficiency - burn slightly leaner than stoich if possible.

How controllable will your VE tables be, or are you just going to do a static one? Either way, it would be good to be able to load those from some kind of user-configurable file so that the ECM can compensate for different engine configs without wholly self-tuning.

This is where I see this thing getting big. I've not scratched the surface here.

I was thinking of this in a C level language (or higher) -> what can be done in one line in C is likely very many in 8085 ASM - even after a skilled assembly programmer optimizes.

Good luck, it does sound fun.

 

Thank you for the comprehensive reply.

Taking a baby step at this project, let's say we want to install an electronic Quadrajet in place of a conventional unit. We want no more or less functionality than the traditional unit provided, which is namely:

• The primary metering rods responding to manifold vacuum.

If I am thinking about this correctly, all that is needed to make the electronic Quadrajet "work" or minimally drivable is

• A manifold absolute pressure (MAP) sensor controlling the pulse width of the 10Hz signal, period. If this worked well enough, then more complexity could be layered on, in both hardware and software, modifying the pulse width with sensor readings from the TPS, ECT, O2S, etc.

(My car has a AutoMeter Wideband O2 sensor/gauge that outputs a linear 0-4V corresponding the A/F ratio, greatly simplifying reading the O2S).

Comments requested.

Michael

 

My friend has a electronic carb he kept when he sold his 82 pickup or something Chevy with the electronic carb. I called him last night and told him find it and hang on to it. This is a real interesting thread.