CIS to EFI & ignition

[tbitz]

Some of you may know that I have converted my 78 911SC to EFI and supply parts for other to do the same.

I've recently started play around with some sort of ignition control. Folks on this board have done ignition control with MegaSquirt and EDIS, but I thought I'd try and keep it simple and use the existing distributor & CDI.

For the last couple of weeks I've had a system working. The way it works is the trigger signal from the VR sensor in the distributor is intercepted by the MegaSquirt ECU. Based on a 2 dimensional spark table megasquirt adds a "correction" to the trigger timing accordingly and the sends it to the Bosch CDI. The CDI unit fires when it receives the trigger from the MegaSquirt ECU.

This allows you to advance or retard the timing on top of the mechanical advance in the distributor. All vacuum hoses are pulled off and MegaSquirt has full control of vacuum advance.

The limitation of this system is the spark table in MegaSquirt does not show the actual spark advance given to the engine, but rather the "correction" added to the mechanical advance.

Also if you change the position of the distributor it will mess up the ignition curve.

This system has required some modifications to the MegaSquirt V3 ECU to be able to trigger the CDI box.

[tbitz]

Although the system works, I am not really satisfied. I've decided to lock down the distributor so there is no mechanical advance curve and have MegaSquirt fully control the spark.

Here is a pic of the distributor with the mechanical advance. You can clearly see the weights and springs.



The weights and springs were removed. As shown here:



I then added some steel wire where the springs used to be to lock down the mechanical advance of the distributor.




The distributor was assembled and put back in the engine.

[tbitz]

Once installed in the engine. I played around with the position of the distributor such that with 32 crank degrees of advance (last tick on the pulley) the rotor was still pointing infront of the correct cylinder as shown here. To do this I sacrificed a distributor cap to see what is happening inside. Here is a pic with the crank pulley at 32 degrees showing the rotor still pointing at CYL #1.







It is important to get this alignment correct, because although MegaSquirt has control over ignition timing, it has no control at which CYL will get the spark. The distributor still has this job. If the rotor is not infront of the proper CYL for all anticipated advance timing you may get firing in the wrong CYL.

Once I got this the way I think it should be I screwed down the distributor. Using a timing gun I determined that the fixed mechanical advance set by the distributor is 29 degrees. I set this in MegaSquirt. I also set megasquirt to ignor the spark advance table and supply a constant advance of 20 degree regardless of RPM or MAP pressure. Took the car for a ride around the block and it looks good.

Now I need to come up with a ignition curve to set this up properly. This is a 2D table which contains ignition advance for every RPM/manifold vacuum entry.

This is where I am today. The problem will be getting a good ignition table and making sure the system works for all timing.

Anyone have an ignition curve I can start with??

[Zendalar]

hehe, i was asking the same thing the other day..

Need Mechanical Advance Curve table for 2.4T

But 3d map...well i hope you find one too and share it
I tried to make one 3d map according to other peoples experiences, but once i was finished and saved the file, megatune crashed and bye bye, there went my 3d map...

I will check if i have any version of the map around...

[Zendalar]

Reply nope...just had the 3d map which holds only 2d info, advance curve that does not count the MAP value..

7 degrees at 1000rpm, and 38 degrees at 6000- RPM
..

[A Quiet Boom]

I did a similar thing with my distributer when I set up my MSD timing computer. The MSD requires that total timing be set then it removes timing based on how you set it. There are 3 screws one for total timing removed or idle timing, another for the rpm at which total timing is restored and a third for what rpm the timing starts. Mine curve is pretty simple, total timing is 34 degrees, idle at 14 degrees, timing curve starts at 1200 and is all in by 3000.

Rather than wire my distributer I welded the weights in place. Next time I have the engine out I'm gonna modify the rear mount to accept a crank trigger wheel that I made on my CNC. It will work with the stock MSD crank trigger kit. Basically it's a wheel with 3 magnets on it. I need to make clearance in the mount as well as a slot so I can adjust the sensor.

Anyway I've found that at least with carbs my system works really well with the advance curve I've programmed so you might want to try something similar. I also have a start retard installed that pulls an additional 4 degrees out when I turn the key making it easier to start. My old V8 drag car was a pain to start before I intalled a start retard on it since it was locked at 36 degrees timing with no advance or retard (drag cars don't need advance since most of them run at more than 4000 rpms when making a pass and there is little load on the motor idling around the pits so detonation isn't an issue). Does the megasquirt have a provison for a start retard? I assume the system would work with an MSD box, is this correct?

[tbitz]

Thanks for the input.

Christian,

MegaSquirt has all the ignition features one could dream of.
-ignition timing for a RPM/MAP table.
-Crank timing (fixed value or table)
-Timing for cold engine
-trim angle (quickly shift all table entries up/down)
-fixed timing (ie: lock timing to fixed value regardless of RPM, MPA, etc)
-RPM & fuel cut for over rev

Support for EDIS, distributor, HEI, etc....

I noticed the following when I had fixed the timing to 15 degrees. I had the timing light on the pulley and it was locked to 15 degrees. When I reved the engine, I saw the timing advance/retard as the RPMs went up/down. I believe this is due to the fact that MS only gets 6 trigger for every 720 degrees of rotation. There just aren't enough triggers happening fast enough to get proper crank position as the engine speeds up. This is a limitation of using the VR sensor as the trigger input.

Trigger wheels like EDIS that have 36 teeth on crank provide many samples to the ECU to determine exact crank position during acceleration/deceleration.

This may not be a problem since during actual driving the engine never accelerates as quickly as it does with no load.

Gonna try fiddling with some ignition curves and see how things go.

Cheers,

[mb911]

Will the MS V2.2 support this as well? I would be interested in doing this as well

[tbitz]

Ben,

MegaSquirt hardware version 2.2 does not have a VR input so it cannot trigger directly from the VR sensor in the distributor. You need MS V3 for this.

Also note that the stock MS unit V3 needs custom modifications to be able to then trigger the CDI unit. This is all very much experimental at this stage, so I will not be including this in any kits until I have a reliable solution that has been tested for a few thousand kilometers of driving.

I will post my progress in this thread.

Cheers,

[bigchillcar]

superb work, tony..keep at it! here's a bump for ya, buddy..
ryan

[911Mac]

Tony,

What made you choose this triggering method over some of the others like EDIS which completely replaces the dist. Was it mainly cost and simplicity?

I've heard EDIS won't support running larger spark plug gaps and that you can have issues with stray electrical emmissions falsely trigger the ignition. But it sounds like a fairly easy way to go the twin plug route. Just curious about your opinion.

Having megasquirt control the ignition could be a real benefit for those running a turbo.

Thanks,

Mac

[JohnJL]

Awesome work Christian and Tony, please keep us "posted."

[tbitz]

Mac,

I choose the distributor because it is a simpler & cheaper solution than EDIS.

Cheers,

Tony

[JohnJL]

I have twin EDIS ignition units and coil sets bolted to the firewall, the distributor installed, a 6400T MSD unit, a tach adapter, a Megasquirt relay box and megasquirt on the inside firewall. I have enough electronics installed in the boot to run two cars!

Why? Because I am planning on getting everything running and rough fuel tuning on the standard distributor and MSD setup, then switch over to twin-plugs driven by Megasquirt and the EDIS coils.
THerefore, I am keenly interested in this kind of development work. I was also just about to search on how to lock my distributor...and there it is!

I also prefer the simpler approach normally, but my JE Pistons CC'd out to 11:1 CR so I will need both plugs esp under load.

thanks guys, keep posting.

[CindyDuck]

Tony,
First let me say thanks for the assistance with my EFI system.
Is the V2.2 compatible with EDIS from a Ford V6? Your distributer system would be a much easier conversion, as I have not been able to locate a 36 tooth trigger for a Porsche crank. Keep us informed.
Mike

[tbitz]

Got this idea for a knock sensor from a number of folks.

Use a mechanics stethoscope to listen in on the engine for knocking. The idea is that the human brain can filter out engine noise from knocking better than any computer can. We'll see....

First I made a small adaptor that will be screwed into a stud on the intake runner. I figure the stud goes right into the cylinder so any knocking noise should go right to this stud and into the stethoscope.

Here is the adaptor made from 1/8" aluminium plate.




The adaptor is screwed into the stud of cylinder #1 and the stethoscope is screwed into the adaptor plate.









I started the engine and put on the stethoscope. With the engine idling I think I can hear the air flowing through the runner. Seems like I can pickup on the finest of sounds coming from that part of the engine.

l'll extended the stethoscope hose into the cabin so I can drive and listen to the engine for any knocking. Hopefully I'll be able to dial in WOT ignition fairly quickly with this method.

Cheers,

[mb911]

where did you end up with this thing?? I am interested in getting rid of my dizzy

[304065]

The important question is, how well did it work with the advance locked out?

Airplane magnetos are fixed timing and use an "impulse coupling" or "shower-of-sparks" device to retard the spark for starting. Using one of the MSD boxes, or the MS computer, accomplishes the same thing.

But what about operation at idle?

There are 360 degrees of distributor rotation. Six cylinders to fire gives you a maximum of 60 degrees of arc in which to fire the plug. In order to avoid firing the adjacent plug, the rotor width must be less than 60 degrees, otherwise it would fire two plugs at once. Let's suppose the rotor is 20 degrees wide. With the rotor aligned with its leading edge under the plug terminal at 32 degrees of advance, doesn't that mean that the rotor's leading edge is well away from the plug terminal when it fires with the engine running at idle at TDC? What happens to the cap and rotor when it has to jump a bigger gap inside the cap?

Not to be a naysayer, but isn't that the reason that the 3,2 and 3,6 Motronic cars had an advance mechanism, not to control timing, but for proper "phasing" of the rotors through the engine's operating range, even though spark was controlled by the Motronic control unit?

[304065]

Here's an excellent page from a Porsche restoration shop in Minnesota. Don't know whether it worked or not.

http://www.the-nurburgring.com/pageresday42.htm

[hobieboy]

Originally posted by mb911
where did you end up with this thing?? I am interested in getting rid of my dizzy

If you are using MS-1 v3 board, why not consider direct coil control? Though still early days & sorting out the last details, but I have done the conversion and the car is running pretty good. See here - Another 930 Megasquirt EFI completed