[Jama]

My new project car, a 1969 Mustang has a ford FE engine in it.
My goal is to install port EFI with minimal modifications to the engine bay.
So for now a distributor stays.

I'm using a Holley Hyperspark distributor.
https://www.holley.com/products/fuel_sy ... /565-323BK
Internally, it looks to use the TFI shutter wheel.
I did briefly consider making a higher resolution shutter wheel (24-1), but stalled on that idea.

8 evenly spaced teeth on the rising edge, with one of shorter duration.

image.thumb.png.3fbd1ad3cc6e0e6d4cfb7484343f74e1.png (726.46 KiB) Viewed 7266 times

image.png.469f3f259a4fde29669b3bbd3e77948d.png (3.75 KiB) Viewed 7266 times

My goal is to implement this as a decoder to allow for full sequential with just the one signal.

[PSIG]

Excellent! We have talked many times before about the "signature tooth" detection many systems from Ford to GM to Bosch have used. Keep us updated!

[Jama]

Excellent! We have talked many times before about the "signature tooth" detection many systems from Ford to GM to Bosch have used. Keep us updated!

I did recall seeing it recently, so went and dug that other thread up.
I know there are a few other similar systems, so I'll try to make it adaptable for other patterns in future.

Today's progress.
Hooked up the dizzy to A speeduino and confirmed I got RPM in basic dissy mode.
I then captured the signal with a logic analyser.
And lastly, I've made it into an Ardustim pattern so I can easily feed it at different RPMs to my speeduino.

In the below image, you can see the capture and my generated pattern.
The red mark indicates the single shorter tooth, and the blue the 7 other equal sized ones.

TFI Ardustim.jpg (181.34 KiB) Viewed 7244 times

[NickZ]

great, this is a very sort after pattern. the small tooth does not seem much shorter, would there be enough difference for the firmware and processor to pick it up and isolate for noise and power disruptions?

[PSIG]

I would think you would want to look at both the trailing edge timing from #8 to #1, and the same timing from #1 to #2, for comparison as signature tooth. By eyeball it is about 25-30% difference, if you're comparing those, and not just a ±10% difference in the tooth width. In a way, it would be like missing-tooth, but in-reverse, and tighter tolerances (<30% instead of >50%) This will make sensed edge selection very important.

Of course, comparing both rising and falling edges would be nice. What approach are you thinking for this?

TFI_signature.jpg (19.54 KiB) Viewed 7019 times

[Jama]

I would think you would want to look at both the trailing edge timing from #8 to #1, and the same timing from #1 to #2, for comparison as signature tooth. By eyeball it is about 25-30% difference, if you're comparing those, and not just a ±10% difference in the tooth width. In a way, it would be like missing-tooth, but in-reverse, and tighter tolerances (<30% instead of >50%) This will make sensed edge selection very important.

Of course, comparing both rising and falling edges would be nice. What approach are you thinking for this?


TFI_signature.jpg

Hey PSIG.

I've started my working out the expected crank angle for each rising and falling edge.
This calculation is based on the OEM Ford setup where the rising edge of the narrow tooth is 10 Deg BTDC.

What I'm planning to do is have the narrow signature tooth/gap be the second tooth in my series.
So in theory full sync is possible with less than 1 full revolution, this remains to be tested.

Also, as the sensor in this situation can only be hall effect, I can trigger on change of tooth edge.
So all 16 edges will update the internal calculation.
Giving a theoretical maximum of 55 Deg (Large tooth gap) without a reference signal, which is much better than 90, if using just one edge.

For now, I'm working on it being a hall sensor with nothing inline to invert it.
So the user can not change the edge.

Code: Select all

  // 7 even teeth & gaps, 1 narrow tooth & longer gap
  // narrow tooth starts 10 deg BTDC
  // Runs at cam speed, so 720 degrees and sequential
  // Normal tooth length = 48 Deg
  // Normal tooth gap = 42 Deg
  // Narrow tooth length = 35 Deg
  // Narrow tooth gap = 55
  toothAngles[0] = 620;   // Tooth before Narrow Tooth Falling edge
  toothAngles[1] = 668;   // Tooth before Narrow Tooth Falling edge
  toothAngles[2] = 710;   // (Narrow) Tooth Rising edge
  toothAngles[3] = 25;    // (Narrow) Tooth Falling edge
  toothAngles[4] = 80;    // N+1 Tooth Rising edge
  toothAngles[5] = 128;   // N+1 Tooth Falling edge
  toothAngles[6] = 170;   // N+2 Tooth Rising edge
  toothAngles[7] = 218;   // N+2 Tooth Falling edge
  toothAngles[8] = 260;   // N+3 Tooth Rising edge
  toothAngles[9]  = 308;  // N+3 Tooth Falling edge
  toothAngles[10] = 350;  // N+4 Tooth Rising edge
  toothAngles[11] = 398;  // N+4 Tooth Falling edge
  toothAngles[12] = 440;  // N+5 Tooth Rising edge
  toothAngles[13] = 488;  // N+5 Tooth Falling edge
  toothAngles[14] = 530;  // N+6 Tooth Rising edge
  toothAngles[15] = 578;  // N+6 Tooth Falling edge

  triggerActualTeeth = 16; //Both sides of all teeth over 720 degrees

This is all theory at the moment, but hopefully should be possible to implement.

James

[PSIG]

Good luck!

[5.0thunder]

Awesome to see this being worked on. I was pushing hard for this a few months back and aside from Nick Hay, I couldn't get much traction.

I have a Speedy V0.4 running on a TFI ignition fox platform car right now so I can help test this in any way you deem necessary. The car isn't wired for sequential fuel at the moment because of the 4 channel ECU outputs but I also have a Dropbear that I intended to use on of my TFI cars if we can get the trigger settings working.

[PSIG]

Just a note that paired can still be cycle-timed, so a V8 with 4 INJ channels should still a viable test mule. This would allow a paired setup to squirt on its two cylinders at the same point every sync (not randomly 1/2-cycle out), although that's not the point of the testing.

[5.0thunder]

Just a note that paired can still be cycle-timed, so a V8 with 4 INJ channels should still a viable test mule. This would allow a paired setup to squirt on its two cylinders at the same point every sync (not randomly 1/2-cycle out), although that's not the point of the testing.

Right, it is wired in pairs too because i was hoping to test the trigger capturing and outputs with it regardless.