Dual (redundant) EDIS & MJLJ setup
Posted: Fri Jan 21, 2011 9:44 pm
I'm converting a Corvair (air cooled flat 6) engine for use on an aircraft. I wanted to use electronic ignition but was concerned about having a single thread system that would kill the engine if there was a problem with it. In aircraft, reliability is at the top of the list. I like the basic Ford EDIS controller and coil pack. They are reliable and if the EDIS loses the timing controller, it reverts to 10 deg advance. That'll keep the engine running even if poorly. To increase reliability even further, I wanted the option of having two sensors and two EDIS units, and figured that I could get by with a single coil pack since even a coil failure would still leave 4 cylinders firing. I had never seen two EDIS units connected to a single coil pack. I decided to think this through and experiment with it. I used one MJLJ v3 and EDIS-6 I had on hand, and obtained another MJLJ v4 and EDIS from AutoSport Labs. The bottom line is that it all works perfectly. Here is what I did.
I made a trigger simulator using a 4" dia 36-1 tooth wheel from DIY Auto. The trigger wheel is driven by a small electric DC motor. Variable voltage from a power supply allows variable RPM from 500 to 6500. I have two sensors mounted 120 degrees apart. Each sensor feeds it's own EDIS-6 unit. Each EDIS has it's own MJLJ unit for timing control. The MJLJs have the same tuning file loaded although they wouldn't have to have the same file. Each MJLJ trigger offset is used to insure timing is coincident for each one and that value is loaded separate from the tuning file so the same file can be used. The EDIS outputs on pins 10, 11, 12 is connected to the coil normally for EDIS #1 which has the sensor mounted normally at 60 degrees BTDC. The outputs on pins 10, 11, 12 for EDIS #2 has to be connected to the coil differently. That is, since my #2 sensor is 120 rotation degrees after #1 sensor, I have EDIS outputs 10, 11, 12, connected 1 firing sequence after the EDIS #1 connection. The EDIS outputs 10, 11, 12, fire the coil A, B, C inputs (coil plug pins 1, 2, 3) in the sequence of A, C, B (1, 3, 2). Therefore the EDIS #2 outputs 10, 11, 12 are connected to the coil plug pins 2, 1, 3. If the #2 sensor is mounted at some other 120 degree relationship to the primary sensor, then the #2 EDIS outputs will have to be connected to the coil differently. That will have to be figured out.
Timing relationship was checked using a dual trace oscilloscope with sensor #1 waveform on channel 1 synced on the trigger wheel gap. The spark timing was checked by monitoring the EDIS outputs on channel 2. The spark timing is in the correct sequence for all three coils whether ingition system 1 or 2 is in operation. The coil primary firing spike is 350 volts and 8 usec in duration. The spike amplitude and duration doesn't change whether one, or two, EDIS outputs are connected to the coil.
I cannot run both EDIS/MJLJ units simultaneously because I get timing jitter which I suppose is to be expected. Running one unit at one time is fine and I can switch from #1 to #2 with no perceived interruption. Running it on an actual engine will show what happens. It might miss firing one or two cylinders upon switchover. That's acceptable. Ignition #1 or #2 switching is simply by switching power to the EDIS/MJLJ #1 or #2. +12V stays on the coil pin 4 continuously. With both EDIS/MJLJ units off, the coil draws no current so the coil can remain connected to the battery without a switch. Not having a coil switch removes one more failure point.
I have run the system for 10 hours on one system, then switched to the other system for 10 more hours. No problems seen. I was concerned that the coil primary spike may damage the output of the "off" EDIS, but the built in EDIS protection seems to be adequate for this configuration. The coil draws 1.75 amps changing slightly with RPM. The EDIS and MJLJ draws about 180 ma. The coil temp rise is about 40 F. The ambient temp where I had the test was about 45 F. I think it will be a good plan to provide some fresh air cooling to the coil since it will be mounted in the engine compartment and will get quite hot.
This will work perfectly for my airplane engine to give me the greatest redundancy in ignition with least $ cost and weight. It would work good for a competition car too for the max in ignition reliably.
Here are a few pictures. Please excuse my rats nest test setup. The trigger wheel is spinning, the flash froze the image. Plug gaps are widened to over 1/8" to raise the secondary voltage to make it more difficult to fire the plugs.
Louie
I made a trigger simulator using a 4" dia 36-1 tooth wheel from DIY Auto. The trigger wheel is driven by a small electric DC motor. Variable voltage from a power supply allows variable RPM from 500 to 6500. I have two sensors mounted 120 degrees apart. Each sensor feeds it's own EDIS-6 unit. Each EDIS has it's own MJLJ unit for timing control. The MJLJs have the same tuning file loaded although they wouldn't have to have the same file. Each MJLJ trigger offset is used to insure timing is coincident for each one and that value is loaded separate from the tuning file so the same file can be used. The EDIS outputs on pins 10, 11, 12 is connected to the coil normally for EDIS #1 which has the sensor mounted normally at 60 degrees BTDC. The outputs on pins 10, 11, 12 for EDIS #2 has to be connected to the coil differently. That is, since my #2 sensor is 120 rotation degrees after #1 sensor, I have EDIS outputs 10, 11, 12, connected 1 firing sequence after the EDIS #1 connection. The EDIS outputs 10, 11, 12, fire the coil A, B, C inputs (coil plug pins 1, 2, 3) in the sequence of A, C, B (1, 3, 2). Therefore the EDIS #2 outputs 10, 11, 12 are connected to the coil plug pins 2, 1, 3. If the #2 sensor is mounted at some other 120 degree relationship to the primary sensor, then the #2 EDIS outputs will have to be connected to the coil differently. That will have to be figured out.
Timing relationship was checked using a dual trace oscilloscope with sensor #1 waveform on channel 1 synced on the trigger wheel gap. The spark timing was checked by monitoring the EDIS outputs on channel 2. The spark timing is in the correct sequence for all three coils whether ingition system 1 or 2 is in operation. The coil primary firing spike is 350 volts and 8 usec in duration. The spike amplitude and duration doesn't change whether one, or two, EDIS outputs are connected to the coil.
I cannot run both EDIS/MJLJ units simultaneously because I get timing jitter which I suppose is to be expected. Running one unit at one time is fine and I can switch from #1 to #2 with no perceived interruption. Running it on an actual engine will show what happens. It might miss firing one or two cylinders upon switchover. That's acceptable. Ignition #1 or #2 switching is simply by switching power to the EDIS/MJLJ #1 or #2. +12V stays on the coil pin 4 continuously. With both EDIS/MJLJ units off, the coil draws no current so the coil can remain connected to the battery without a switch. Not having a coil switch removes one more failure point.
I have run the system for 10 hours on one system, then switched to the other system for 10 more hours. No problems seen. I was concerned that the coil primary spike may damage the output of the "off" EDIS, but the built in EDIS protection seems to be adequate for this configuration. The coil draws 1.75 amps changing slightly with RPM. The EDIS and MJLJ draws about 180 ma. The coil temp rise is about 40 F. The ambient temp where I had the test was about 45 F. I think it will be a good plan to provide some fresh air cooling to the coil since it will be mounted in the engine compartment and will get quite hot.
This will work perfectly for my airplane engine to give me the greatest redundancy in ignition with least $ cost and weight. It would work good for a competition car too for the max in ignition reliably.
Here are a few pictures. Please excuse my rats nest test setup. The trigger wheel is spinning, the flash froze the image. Plug gaps are widened to over 1/8" to raise the secondary voltage to make it more difficult to fire the plugs.
Louie