DIM Electronic Advance Ignition

-

KitCarlson

Well-Known Member
Joined
Jan 17, 2008
Messages
2,624
Reaction score
461
Location
Middle Tennessee
Hi,

DIM means : Do It Myself.

I have been working a couple weeks on developing an electronic advance distributor. This is done by removing the flyweights, and locking the rotor with a setscrew, and using a micro-controller to adjust the timing, based on a timing reference signal. The micro-controller also determines the dwell time to charge the coil such that a ballast resistor is not needed. I am using an optical trigger and will also add a variable reluctor (Mopar or GM style) input.

I am bench testing the unit at the moment. I have used a drill to spin the distributor, and have simulated reference signals using a micro-controller. I am using my hardware from a previous engine management system that I developed in 2003. I have removed all the fuel injection functions and rewrote the ignition code for this application.

I also modified the user interface by pulling out the EFI related components. I have not spent time re-dressing the user interface, so it is ugly. The advance curves for the RPM and vacuum (MAP) are modified using two 2D tables. This will provide similar adjustments without playing with, springs, stops, and vacuum dashpots.

I hope to test the unit in the near future on my Barracuda.
 

Attachments

  • simlog2cpt.jpg
    28.5 KB · Views: 1,107
It sounds like you reproduced the basic functions of the spark system used on GM TBI V-8's ca 1985-1995 (trucks, Camaros, ...), known as the "small cap" or "computer" distributor. Mopar's "no weights, no vacuum" distributors of the same era (K-cars) with "spark control computer" were probably similar, was Ford's TFI system (I think). You will be able to "change weights" and "trim vac advance" from the driver's seat.

You might leverage most components from a GM "8-pin module" and coil. As I understand, it outputs the base timing signal on one wire and your computer returns that signal with spark lead desired (similar to Ford EDIS interface many Megasquirt owners control). I got a junkyard set for all my old Mopars. I will use a Holley Commander 950, which interfaces to the 8-pin module, and acts on a knock signal.

Grab the GM knock sensor (block drain passenger side) and module at the same time. It screwed directly in the block drain of my SB and BB. On my slant, I needed a bushing (3/8 to 1/4 NPT?) for the larger drain. It outputs a single-wire digital signal: high = no knock, low = knocking.
 
Bill,

I am using an optical wheel off the Borg-Warner unit, and a newer opto switch that has automotive temperature ratings. The opto is a zero speed trigger so it helps in development. I plan to add a component for the VR interface.

I am using an ignition driver component, so there no need for a GM module. I think the complete package will be smaller than a GM module. It can easily be hidden away.

There are 6 extra analog inputs, so it is possible to interface with knock sensor, O2 and other data logging signals. The tuning is done real-time, and the data logging interval can be as short as 10mS.

You are correct, it is similar to mid - 80's 2.2L mopar. I have done that before, and used sequential drive for quad coil ignition. There are advantages in that. I did knock control on the 2.2L turbo. There is windowing in the signal processing to reject other engine noise outside the ignition intervals.

For simplicity with OEM appearance, this unit will provide improved tuning and stability. If only there was more play time.
 
You can reduce the effort required by starting with a Lean Burn distributor, which contains no flyweights, has no vacuum advance, and has locked/fixed advance (and more room in the distributor to add whatever you might be adding). Another advantage is that the Lean Burn distributors are in much lower demand than the advancing types, so there are mountains of NOS items available inexpensively. I know Old Car Parts Northwest has lots of them and would probably make you a juicy deal on one or more. They come in single- and dual-pickup versions; if you're putting in your own optical or other trigger, it doesn't matter whether you start with single- or double-pickup distributors.

Please keep us up to date on this...seems to me whatever you come up with would be readily applicable to small and big V8s and slant-6s as well. Please do take a look at the "how/why it's better" links off the HEI upgrade page and comment on how your circuitry will perform vs. HEI.
 
And last, don't forget the issue with "rotor phasing"
 
I really appreciate your comments. It is a lonely job doing this work, and finding those that understand.

Dan,
My first car over 40 years ago was a /6 Dodge Coronet. I built an electronic ignition for it in 1972 that used an optical pickup and a built from scratch CDI system. I will get to the /6, I still have a distributor, but no /6 car it this time. You may be the first to do testing if you wish.

To the best of my knowledge the heart GM module is analog. I am talking about the dwell control circuit. The circuit is fairly well documented made by Moto. Recently internal circuit components have become obsolete to the general public. I am an old analog guy, after learning about micro-controllers 20 years ago, I switched. The smarts of a micro-controller, win in control. The GM module by its self does not do timing control without mechanical means, or ECU intervention. My circuit is an ECU. That is the major difference.

67Dart273,
I have the phasing issues figured out. Base timing of 10 degrees, yields a total timing control over 50 degrees.

Testing is going well. My simulator generates signals for cranking at 200 RPM, and revving from idle to 6000 RPM in 4 seconds, then repeats.
 
I fixed a problem in a calculation, it was a rounding issue in the integer math used to calculate RPM and timing advances. The bumpy green line in the user interface plot is now smooth. Now the system is working well enough to test.

I only have the time to work a couple hours a day, so it will be a few days before it is put on the car. I have a couple other projects in the way.

The pdf capture of screen shot does not work well after compression to fit the post requirements. I do not have time to mess with that now. I might find a solution over time, possibly a youtube video. The user interface is not required, however the ability to log and present data is useful.

By making acceleration runs, changes in timing can be checked to see how the engine responds. A sort of on car dyno.
 
Here is a real time plot of the engine RPM (red), Ignition timing (green), and manifold pressure in (blue). The RPM is simulated by a second micro-controller, it starts with cranking speed, followed by idle, followed by increasing RPM from idle to 6000 RPM in under 4 seconds. A mighty-vac was used to provide the change in manifold pressure. The stair steps in the pressure is the result of small pumps to increase pressure. In a normal engine at WOT, the vacuum drops to zero. For reference the maximum timing is 10 base, plus 26 RPM, plus 12 vacuum. At WOT the timing is 36 degrees. Also at idle, the distributor port will be used, it will not provide significant vacuum at closed throttle. The blue plot is absolute manifold pressure, it reads inverted compared to vacuum. On the left is atmosphere, after the stair-step is maximum vacuum or about 0.02 ATM

I took a snap-shot of only a part of of the screen in an attempt to get a better view. Still not happy with screen capture, then MS paint to crop. Is there a better way?
 

Attachments

  • ims8wMAP.jpg
    30 KB · Views: 1,055
Dan,
My first car over 40 years ago was a /6 Dodge Coronet. I built an electronic ignition for it in 1972 that used an optical pickup and a built from scratch CDI system. I will get to the /6, I still have a distributor, but no /6 car it this time. You may be the first to do testing if you wish.

Yes, plee-uz. Let me know what parts you may want me to send you for experimentation. I still haven't got around to swapping the piece-of-junk Mopar orange box system off my '62 Lancer for something better.

The GM module by its self does not do timing control without mechanical means, or ECU intervention. My circuit is an ECU. That is the major difference.

Sounds like a real nice update! What are you looking at in terms of secondary potential, current, and duration vs. HEI?
 
Dan,

I doubt, the secondary current or voltage will change significantly from the HEI. The dwell control might be improved, however I think the HEI does a respectable job with that. The dwell time can be selected, however at high RPM, it is a matter of available dwell time, and time for the ignition event. That is a reason why coil on plug is better. The /6 has the advantage of less cylinders and typically less RPM requirement than a V8.

The major improvement will be the accuracy, repeatability and the ability to tune without mechanical adjustments. The base timing will still be manual, with a timing light on balancer at idle. Gone will be centrifugal advance, vacuum dashpot and other related components.

Timing is done with two simple tables. The RPM table has 10 data points in equal positions from idle to maximum RPM. The idle speed and maximum, are chosen depending on engine requirements. At each RPM point in table, a corresponding timing value is set. The ECU measures RPM, and in a linear fashion, interpolates values in between the data points. The timing based on vacuum is similar.

I typically work with 3D tables for engine management, the 3D table could be 10 x 10 or 100 cells to fill. For most this is too intimidating. By using the two 2D tables, the mindset tuning conventional distributors stays the same. Since vacuum is near zero at WOT, and closed throttle much of the possible information in a 3D table is non-germane.

The mechanical timing advance mechanisms are fairly crude. There can be friction, looseness, hysteresis, induced vibration and bouncing at stops. The vacuum advance mechanism also has faults of friction, hysteresis, wear and non-linear travel vs vacuum. I find it amazing that they in general work as well as they do. What is hard to do is make adjustments, test and make more adjustments. With the electronic system, change and test is easy!

The electronic pressure (vacuum) sensor is quite accurate and temperature compensated. The ability to measure RPM is also accurate, based on a clock with a quartz crystal. So, the ignition will be more accurate, predictable and easily tuned.
 
I doubt, the secondary current or voltage will change significantly from the HEI

That'd be fine by me; HEI sets a good, high benchmark for those parameters.

The dwell control might be improved, however I think the HEI does a respectable job with that.

Agreed.

The major improvement will be the accuracy, repeatability and the ability to tune without mechanical adjustments. The base timing will still be manual, with a timing light on balancer at idle. Gone will be centrifugal advance, vacuum dashpot and other related components.

That all definitely works for me!

Timing is done with two simple tables. The RPM table has 10 data points in equal positions from idle to maximum RPM. The idle speed and maximum, are chosen depending on engine requirements. At each RPM point in table, a corresponding timing value is set. The ECU measures RPM, and in a linear fashion, interpolates values in between the data points. The timing based on vacuum is similar.

Okeh…and how will these tables be accessed? Needs a PC or...?

The mechanical timing advance mechanisms are fairly crude.

You're being unnecessarily diplomatic.
redbeard.gif


I find it amazing that they in general work as well as they do.

That is pretty amazing, isn't it.

With the electronic system, change and test is easy!

Huge advantage.

The electronic pressure (vacuum) sensor is quite accurate and temperature compensated. The ability to measure RPM is also accurate, based on a clock with a quartz crystal. So, the ignition will be more accurate, predictable and easily tuned.

Very interesting. I'm keen to learn more about that sensor. You've got me wondering if this what you describe could be used to generate a reliable RPM signal for an add-on fuel injection system.
 
Subscribed! I have a few of those lean burn era distributor's, but the one I'd be most interested in modding would be from my 78 400. I'd like to use it in conjunction with a dual g.m t.b.i system, and hei ignition.
 
You've got me wondering if this what you describe could be used to generate a reliable RPM signal for an add-on fuel injection system.

Fuel control is actually easier since you don't need exact timing, unless trying to do "sequential injection". The Megasquirt site says that isn't needed and "batch mode" is sufficient. At anything above idle, the injector sprays longer than the intake valve stays open, so timing doesn't matter. For TBI, batch mode is fine. Batch mode requires no cam sensor, only rpm, plus a few other sensors - manifold pressure, inlet temp, throttle position, or just a mass flow sensor. Injector drive is fairly simple today with a chip like the CS453 that takes TTL from the computer and can drive throttle body or multi-port injectors, high or low impedance.

Re the concern with time to charge the coil, I understand that is less a problem with an E-core coil. Look at the one on Mopar Magnum engines, GM "external coil" (85-95), or similar Ford. Otherwise, I think you must have a distributor-less system to use multiple coils.
 
Dan,

Thank you for the comments. I use a less that $300, 10.1" minibook with USB port. I have a special application that I wrote for development. It is all in one, for updating code, presenting real-time data with meters and charts, setup and table changes, and data logging. I have other ideas to use what most people carry with them every day. That is another project.

The RPM is measured by measuring the period of ignition trigger signals. The reciprocal of period is frequency, scale to revolutions per minute you have RPM.

Bill,

I will pick up an e-core coil and do some evaluation.

I have always developed synchronous fuel injection based on speed density. I measure fuel requirements for each combustion event. Cylinder filling requirements are determined from manifold pressure, temperatures, VE table, RPM and TPS, with WBO2 for feedback. I found that sequential control reduces low RPM emissions and improves idle quality.
 
All,

More about the progress on the electronic advance distributor project. The controller contains the vacuum (MAP) sensor and the ignition driver. No other modules or ballast resistor required. Hardware includes hooks for CAM/Crank sensing. Graphical user interface (GUI) via USB, not required for normal operation.

I took Dan's great advise, and used a lean burn distributor. It is an off the shelf unit that is very simple, without vacuum and mechanical advance. It also has a single pickup coil. Distributors are available for both /6 and V8.

Bill also suggested an e-core coil, and it has also shown excellent performance.

I have been doing significant testing on the system with good results. Testing is done with a distributor simulator, it simulates starting, idle, increasing RPM, past rev limiter, then repeats. I also have a hardware setup with 8 spark plugs, a variable speed motor for rotating the distributor, and a timing light method to check advance. A vacuum source is used to simulate engine vacuum.

I have an monitoring system that includes a scope and logic analyzer to measure crucial timing of signals, currents and voltages.

I plan to evaluate and evaluate the best, coils, caps, rotors and plug wires. This testing will also be for component endurance, it will be ongoing and take time.

I have attached a screen shot of the recent GUI changes. The data plot is able to record multiple data at 10mS intervals. It is several times faster than OBD2.

Comments by all are always appreciated.
 

Attachments

  • GUI.jpg
    30.2 KB · Views: 917
Thats cool man, I like it.
I'm going to keep up with this also.
So are you going to end up with the ECU and the control device to make the changes if preferred by the user or are you going towards fully automated Advance and timing curve control?
 
And a messy picture of a test bench. The black box with heat sink is the controller. Under it, is a heavy duty 13.8V power supply to replace the battery.

The controller works fine with the open sparks nearby. That is often a problem for many experimenters.

The setup will improve as time permits.
 

Attachments

  • IMG_0613 (800x600).jpg
    77.5 KB · Views: 924
TrailBeast,

Yes, do not let me stop you. I apologize for the hijack on your thread. I was not thinking at all.

Thank you for the kind words. Not sure mine will go to market, and if it does it could be a long time. I just have this in my blood and need to get it out. Finally doing things I wanted to do 40 years ago.
 
Yes, do not let me stop you. I apologize for the hijack on your thread. I was not thinking at all.

Thank you for the kind words. Not sure mine will go to market, and if it does it could be a long time. I just have this in my blood and need to get it out. Finally doing things I wanted to do 40 years ago.

Kit, i did not mean to throw stones if i did, you have a great thing going here and i would like to see where it goes. i was only looking out for Trailbeast because he is simplifying it in one package and i didnt want it to get off track. my apologies as well.
 
No worries mates.
They almost seem like the same thing in a way except you are literally building the whole thing, from code to implementation with your own mods built in.
Pretty cool.
 
Returning to 273Dart67's comment on phasing, I wonder if mod'ing a regular distributor w/ weights would even work since as you command 35 deg advance the rotor tip will move far away from the tower. It doesn't move w/ weights active, except +/-7 deg for vac advance. I expect that the lean burn distributor has either a wide rotor tip or wide tower terminals to correct for that. MSD distributors have wide tips.

Some other people found this website that is doing a similar effort as yours:
www.rabidgator.com
While it looks slick, it doesn't appear they have a product shipping yet. There is/was a pdf manual on there since I viewed it, but what I saved is just the web page and I can't get back to the manual today. I recall that their box uses the GM HEI 8-pin module as I recommended earlier.
 
Returning to 273Dart67's comment on phasing, I wonder if mod'ing a regular distributor w/ weights would even work since as you command 35 deg advance the rotor tip will move far away from the tower. It doesn't move w/ weights active, except +/-7 deg for vac advance. I expect that the lean burn distributor has either a wide rotor tip or wide tower terminals to correct for that. MSD distributors have wide tips.

Some other people found this website that is doing a similar effort as yours:
www.rabidgator.com
While it looks slick, it doesn't appear they have a product shipping yet. There is/was a pdf manual on there since I viewed it, but what I saved is just the web page and I can't get back to the manual today. I recall that their box uses the GM HEI 8-pin module as I recommended earlier.

The Rabid Gator device/computer does use the 8 pin ECU Bill.
I have the pdf you mentioned here [ame]http://www.letsgocomputers.com/classichei/rabidgatortimingcomputer.pdf[/ame]
Dang man,,,give FABO a couple of years and and mix a little applied quantum physics in there with your calcs and you could be doing -12.87 quarter miles.
You crossed the finish 12.87 seconds before the tree turned green. (Now thats fast) :)

Wait! Does that mean your run never happened?
Or does it mean it's happening every 12.87 seconds for eternity?

Now I'm confused.
 
73AbodEE,

No stones, you made a good call, in a very polite way. I am a bit of a mad scientist, working with machines and electronics most of the time, my people skills are limited. You have actually helped me, by providing feedback.
 
-
Back
Top