COP Ignition Development

-
The following picture, shows the decode of the opto sensor wheel and 7 of the 8 ignition outputs. My logic analyzer only shows 8 channels, so the 8th channel is viewed along with the reference signal on my 2 channel scope. This is a good start for the ignition system. From this point, addition of other features should be straight forward.
 

Attachments

  • OptoWhDecode.jpg
    21.4 KB · Views: 483
I found enough time to work on the data acquisition system that reads the MAP and battery voltage. A terminal is used to display the USB data every 200ms.
View attachment COPvals.PNG
I have also demonstrated timing and dwell control and captured signals with logic analyzer. The waveforms show 7 coil drive channels with an ignition timing of 25 degrees advance and a dwell of 1.6ms.
View attachment COP25d.jpg
The simulator ramps RPM, show below is a ramp 600 TO 6400 RPM.
View attachment 600to6400.jpg
 
Interesting project! I hope it's ok that I post a few pics showing my homework...

I'm using the D585 coils and converted the original distributor to a cam sensor using simple home tools. For the crank sensor I welded a tooth wheel to the pulley, using a VR-type sensor. The cam sensor uses a Hall-type sensor. It all works well.
 

Attachments

  • PC043582.jpg
    29.7 KB · Views: 470
  • P9273267.jpg
    29 KB · Views: 449
  • P9073199.jpg
    26.2 KB · Views: 434
Nice work! I am having difficulty understanding the interaction between the Hall sensor and wheel. A post of the waveform would be helpful.

Are you using Megasquirt?
 
KitCarlson said:
Nice work! I am having difficulty understanding the interaction between the Hall sensor and wheel. A post of the waveform would be helpful.

Are you using Megasquirt?
Click to expand...
Yes, MS3X for full sequential injection and ignition.

The cam sensor consists of a round piece of thick aluminum, on which the half-circle piece of soft sheet metal is mounted. The Hall-sensor has a pullup-resistor, and the piece of sheet metal is mounted in a position giving a high signal (5 Volts) to the ECU when cyl 1 is on the compression stroke.
It's now about 2 years since I did that, but if I remember correctly the output wave is a pretty square wave, 5V. The signal goes high approx. 180 crank degrees before cyl 1 TDC on compression stroke, and goes low approx 180 crank degrees after TDC.

The only function of the cam sensor is to tell the ECU if cyl 1 is on the compression stroke or not. The crank sensor gives the actual position of the crank.

I could provide a wave form some days into the new year if you need it.
 
Here is the sensor waveforms applications planned for my system.

The top is a 4 tooth wheel on the crank, it would be used with distributor 1/2 circle. The half circle is OEM distributor sensor on some magnum engine. This is simple and requires little decoding by micro controller. The cam sync can happens at 360 degree crank intervals. The cam/crank method could also be incorporated in distributor housing by adding second target wheel and sensor.

The bottom is a single sensor tab wheel with window in one tab for sync. The sync happens at 720 degree intervals. The single sensor is low cost, and mechanically plug and play.
 

Attachments

  • TimingSensors.jpg
    16 KB · Views: 461
So the intended application of this is COP/CNP on carb'd engines? Is it for use with stock Mopar ignitions? I understand that it would offer some control of the coils, but what do you hope to gain from that?
 
It is for mainly carbed engines, most modern EFI systems include programmable ignitions. See point #6 below, it provides crank cam signal for sequential EFI. It is for Mopars, requires distributor modification, does not use Mopar ignition box or ballast resistor.

Some of the gains are as follows:

1. Electronic advance, elimination of flyweights and vacuum dash port for accurate programmable advance curves. Electronic advance provides more adjustment than can be done with flyweights, slots and springs .... An example is timing retard for rev limit, timing retard for turbo.

2. Elimination of distributor rotor and long plug wires.

3. High energy multi-spark. Longer duration than MSD CDI.

4. Programmable rev limit.

4. Simple real-time tuning.

5. Data logging and presentation of RPM, Ignition advance, manifold pressure, battery voltage and engine temperature.

6. EFI ready timing reference signals, including tachometer signal.

7. Fuel pump keep-alive signal, for running engine, shuts off pump 0.2s after stall condition.

There may be more ...
 
Nice work!

I've been interested in a programmable advance curve form some time now and last year even bought myself a BlackBox from CB Performance.
Unfortunatly the box developed some backfiring/timing issues after awhile, and this being on my daily driver I had to disconnect the ignition system.
 
Thanks for the info, it seems like a neat project.

Any idea on price point for the controller? You said something like $10 per coil earlier? Much more than $25/coil all in and you can get a 6al2 and a lean burn distributor to cover #1-4 and most of #5. I also don't believe 4 crank / 1 cam is enough info for full sequential, is it?

KitCarlson said:
It is for mainly carbed engines, most modern EFI systems include programmable ignitions. See point #6 below, it provides crank cam signal for sequential EFI. It is for Mopars, requires distributor modification, does not use Mopar ignition box or ballast resistor.

Some of the gains are as follows:

1. Electronic advance, elimination of flyweights and vacuum dash port for accurate programmable advance curves. Electronic advance provides more adjustment than can be done with flyweights, slots and springs .... An example is timing retard for rev limit, timing retard for turbo.

2. Elimination of distributor rotor and long plug wires.

3. High energy multi-spark. Longer duration than MSD CDI.

4. Programmable rev limit.

4. Simple real-time tuning.

5. Data logging and presentation of RPM, Ignition advance, manifold pressure, battery voltage and engine temperature.

6. EFI ready timing reference signals, including tachometer signal.

7. Fuel pump keep-alive signal, for running engine, shuts off pump 0.2s after stall condition.

There may be more ...
Click to expand...
 
j.mcconnell said:
Thanks for the info, it seems like a neat project.

Any idea on price point for the controller? You said something like $10 per coil earlier? Much more than $25/coil all in and you can get a 6al2 and a lean burn distributor to cover #1-4 and most of #5. I also don't believe 4 crank / 1 cam is enough info for full sequential, is it?
Click to expand...

If you want plug wires, rotor, cap, single coil, a sloppy VR sensing distributor, CDI, then buy it. I fail to see similarity with COP system.

Yes 4 crank teeth, one cam signal is enough. I have prior experience developing injector control for a Ford 7.3L diesel PZT retrofit, with similar timing system, that reached 2500HP on dyno. Are you forgetting the crank turns 2x, so 4x2 = 8? The cam signal is high for 4, low for 4, for two sync checks per 720 degrees of crank.
 
Here is my test setup. It consists of two $3 Arduino nano boards connected with jumpers. One board simulates the distributor sensor, cranking at 200 RPM, then idle, followed by RPM ramp to about 6000 RPM followed by ramp back to idle.... and repeats. The many colored wires connect to an 8 channel logic analyzer, for capture and measurement of signal timings.
View attachment 20151208_152054.jpg

I found a few hours to make progress on the system. The measurement system that reads the MAP sensor and battery voltage is reporting values via the simple VT100 terminal user interface. The user interface also configures basic engine settings for RPM range, number of cylinders, RPM limit and of course RPM and MAP advance curves. It is simple to change the advance curves in real-time by using keyboard key to select, then increase or decrease by 1 with each press.

The VT100 application has file send and receive capability for data logging and loading or saving configurations. Later a B4J application with more features will be used for an improved user interface. The timing advance curves shown are not realistic, they are being used to test advance vs RPM.
View attachment copgui.PNG

Here is a RPM ramp from idle to RPM limit of 5500. The blank area is ignition cut. The RPM is measured every ignition cycle.
View attachment IdleRl.jpg

This is an example of idle timing. Note the signal fall (ignition event) is near right most part of falling reference signal.
View attachment InitTime.jpg

This is an example of timing at 5300 RPM, note that it is closer to the left side of reference signal, showing about 40 degrees advance.
View attachment 5300Tim.PNG
 
It is raining today so I found the time to get the multi-spark going. The dwell on the first pulse is a coil charge of 1.6 ms, the drop is the spark period of 0.25 ms followed by another coil charge of 0.85 ms. The second charge is less because of the stored energy in the coil. With a single spark, the spark duration would last about 1ms, but the energy would drop over that time. The second spark is full energy, and at idle speed is about 4 crank degrees retarted from the primary spark. Multi-spark is typically used at RPM less than 1500.
View attachment multiSpk.jpg

The firmware is nearly complete, so I am considering the mechanical packaging. Shown are a couple enclosures, the nano ECU board, a MAP sensor, and a coil driver transistor. There will need to be as many coil drivers as coils. Since the design works for 4, 6, and 8 cylinder engines, I have considered making output modules. The ECU board and MAP will fit in the small black enclosure that is 1"x 1"x 2". The larger enclosure is about 1.25" x 3" x 6", it is much larger than needed, but would be a one size fits all, and eliminates some external connections. My plan is to use the large enclosure for prototype, and worry about the packaging and connector details till later.
View attachment 20160115_135023.jpg
 
I started to modify a lean burn distributor. Any distributor will work, but the lean burn has fewer parts to start with. The ignition works with a a shutter wheel locked to distributor shaft. The pickup is an optical vane sensor, but that will be covered later.
View attachment 20160116_131712.jpgView attachment 20160116_132059.jpg
The first step was to disassemble reluctor top hat from distributor shaft. There is a wire type c-clip with tabs in the well below the rotor. Needle nose pliers are used to grab a tab, and pull it out. The reluctor is located on shaft with a roll pin, and a slight friction fit. I got my finger tips at the bottom, thumbs at top and pulled it off the shaft. The base is mounted with two screws, remove those and lift plate. Keep the plate, the optical sensor will be mounted on it.
View attachment 20160116_133145.jpgView attachment 20160116_133210.jpg

The modification process is simple, and could even be easier if the coupler was a manufactured part. I started with a 1/2" long, 5/8" OD aluminum rod stock, with a 5/16" ID. A 6-32 setscrew was added to fasten to distributor shaft. A single lathe cut was made to fit the ID of shutter wheel, a press and 5/16 bolt was used for crimp. The setscrew is aligned to a wheel notch, for ease of assembly.
View attachment 20160116_134226.jpgView attachment 20160116_142124.jpg
View attachment 20160116_142312.jpg
 
I installed an optical sensor board today and did the first test using a portable drill to spin the distributor. I used the sensor PCB that I designed for VW engines with modifications. Only one sensor installed, and the board needed to be offset, so it was changed for that.
View attachment 20160118_150605.jpg

A window was cut into one tab for the cylinder sync reference. I drilled a hole and filed it square. The detection means is flexible.
View attachment 20160118_132227.jpg

I made a new longer coupler for the shutter wheel, it extends to the top of nylon on the 5/16" distributor shaft. I started with a part from by VW conversion. They both have the same ID size.
View attachment 20160118_140520.jpg

Shown below is the distributor timing at if was set to trigger for #1 cylinder, the trailing edge of tab following sync tab is aligned with center of opto sensor. This would be done with crank set to desired base timing.
View attachment 20160118_160722.jpg

The following is a capture of the distributor signal on cyl2, and the rest of the cylinders.
View attachment OptoDistTst.jpg
 
Thanks.

So far tests have been at the micro controller pin level. Next I need to order a few parts, connectors and wire, to build a prototype for on car tests.
 
I've been enjoying watching this project unfold, I think it's a great alternative for people who want to stick with a carb but would like to control ignition timing
 
Yes, very nice work on both the electronics and the mechanical.

Have you considered controlling fuel injection in batch mode? I guess it would require only one additional output (and driver stage), and would add significant possibilities.
 
Yes, I have considered one or two banks of injection. I have 4 i/o pins left. I will also need to add TPS, and intake and engine temperature. I am presently using two 16 bit timers for ignition, one for dwell, the other for ignition event. I can use one timer for both, to free the timer for fuel delivery. It adds some complexity, but since Nano board is less than $3, I could add a second controller, gain more pins for things like MPI, evaporative controls, waste gate control, speed sensor ... I would need communications between the micros. I have experience with that by using two EMS2 units on my 2.2 turbo.

My thinking is the COP ignition is a good starting point. I hope the simplicity of use, ignition power, and reliability will change minds. Then there will be a movement to EFI too.
 
KitCarlson said:
...
My thinking is the COP ignition is a good starting point. I hope the simplicity of use, ignition power, and reliability will change minds. Then there will be a movement to EFI too.
Click to expand...
I agree. It's best to finish the current project and then start thinking about next steps.
As you say, the fuel injection will add complexity in both the electronics/software and in the car itself, and it will add quite some cost and work for the end user.
 
I am working on the system wiring connections at the moment. The complete harness will make it a simple plug and play. I hope to post something in a few days.
 
Excellent work & Many thanks for taking the time to do this!!!

As to 'changing minds' ... some may argue more components
& electronics interface are a drawback...
Maybe 15 yrs ago :) Thx Again !
 
mech1nxh said:
Excellent work & Many thanks for taking the time to do this!!!

As to 'changing minds' ... some may argue more components
& electronics interface are a drawback...
Maybe 15 yrs ago :) Thx Again !
Click to expand...

Thanks for comments.

The poor execution of Mopar electronics in the Lean Burn and L-Body LM-PM disasters left some deep wounds for many. Eliminating the rotor from the distributor is a huge reliability advantage, for the associated electronics. The distributor Hall sensor in the L-Body was easily failed, because the shutter wheel was mounted in nylon. The shutter wheel would build a charge, that would flash to Hall sensor. Leaving out a rotor and cranking a Mopar VR electronic, can fail the control box from spark strike to pickup.

The COP unit has low parts count, the parts are also less stressed because the drivers have 1/8 th the duty of a single coil system. If one ever fails, the other cylinders will easily get you home. The components used are automotive rated, and have features that limit current and voltages to acceptable levels.

A MSD CDI ignition has more parts, and the discharge capacitor is a major wear item. Typical capacitors in pulse applications have lifetimes rated in a few thousand hours at most.
 
Here is a wiring diagram example for the COP ignition on a SBM. The wiring has more coils, but is very similar to a regular ignition hookup. The ECU is mounted on the drivers side inner fender well. There are many ways to route the wiring. The route shown minimizes wire lengths, and provides good noise immunity, the disadvantage is it is not pretty. Others may have a much better ideas than me, to make the system pretty.

The red wire going to the coils is common for all, the gray is not, eight separate, one for each coil. The coils will likely mount on an aluminum channel, with short plug wires.

I have parts on order to build a couple prototypes. I will be building one next week.
View attachment COPwire.jpg
 
You must log in or register to reply here.
-

Back
Top