DIY LTCC or similar system for LT1s

[dzidaV8]

Nice project, I think it would be much easier just to use uP for coil switching, we are talking about frequencies of about 25kHz at 8k RPM, that's basically a slow signal for let's say AVR or STM32 running at a few MHz... But I appreciate your effort to make it discrete!

Another point that comes to my mind is that almost all of the optis I had replaced had the optical sensor failing due to high temperature, NOT the HV part. It seems that over time temperature causes the LED and phototransistor to lost alignment somehow.

I think a good solution would be to use vortec style cover with CKP sensor for hi-res signal (180 tooth wheel), and only low-res sensor in the opti. The question remains if the CKP sensor is capable of picking so dense teeth.

[vilefly]

Thanks dzidaV8, I figured that if I make it hardware based, anyone with a processor can make use of it.
yup, heat kills the optical pickup by increasing the resistance of the led emitter until it burns out or is too dim to work. Heat kills hall effect sensors, too. Roughly at the same temperatures, too. It's just the fact that the emitter generates heat in operation is why it probably fails slightly sooner.

I don't think the vortec timing cover fits the LT-1 due to the water pump/distributor drives. I think you mean the 1996-up timing cover that has the extra 4x reluctor and crank sensor. The tooth density that you speak of definitely would raise issues with the stock crank sensor. They are magnetorestrictive in nature, and can be quite touchy.
A custom hi-impedence sensor would work better....something that develops a higher voltage per tooth, basically.

[spfautsch]

It seems that over time temperature causes the LED and phototransistor to lost alignment somehow.

I believe what happens is the phototransistor slew rate slows to the point it basically runs into saturation (on all the time) when it gets hot. Theres an "equivalent" sensor that was used by Chrysler in a traditional distributor application that was used heavily in cheap aftermarket optis that did this. The Chrysler application was designed to work in a thermosplastic distributor body, and wasn't able to handle the heat of operating while attached to an aluminum plate bolted to the front of the engine block.

[Tom B.]

You need to use the 96-97 cover. The crank damper mount is also shorter to accommodate the trigger wheel.

[Aroberson77]

I took the delteq apart and heres some pictures
20170917_213504.jpg 20170917_213527.jpg

[vilefly]

Ah-ha! they used an UV eprom to possibly sequence the coils, which is something I considered as a cheat if I couldn't find the right combination of counters and gates. Could also simultaneosly store an advance curve, too. Can you read the number on the SOIC chip up there? I suspect a data distributor chip of some kind.
He said it uses a northstar engine ignition module, but I have a doubt, since the northstar uses 2 crank sensors angled 27 degrees apart and a very complex reluctor wheel on the crankshaft.
northstar.gif northstar2.gif

[vilefly]

Still trying to work despite constant distractions. Here is roughly where I said I was at. The Digikey website has a free online electronics drafting program linked to their huge catalog. It is called "Scheme-it". Allowed me to post a diagram with a screen capture.
ltcc.jpg

I haven't setup the reset for the counter, and other stuff. but it should give everyone an idea.

[Tom B.]

The Delteq does use a Northstar module and coils.

[lionelhutz]

I've been trying to figure out what you're describing. The 1-8 demultiplexer made no sense. A 5-bit ripple counter could work by counting the high res pulses each time the low res turns-on.

I assume once you have the count you'll latch it into a chip so it's held until the next count? You need to use a latch signal for this chip that transitions properly when the low res signal goes away.

I would have expected you to use the high res as the clock to the ripple counter. I suppose you could and the high res and low res so the clock pulses only occur when both exist. Definitely use the low res as the input. So, the counter begins counting up during the low res pulse and then when the low res pulse ends you latch the count into the latch chip.

I'm curious how you'll determine between the other 4 cylinders which have the equal 2 high res pulses. I suppose you could keep the count from the previous cylinder and then count the 2 pulse pattern on-top. Cylinder 4 would be 9 counts for example which is a unique count. This means you'd need a counter reset circuit that looked specifically for a 2-pulse pattern to reset the counter.

I'm not sure why you'd want to use the ignition module and then use it to drive another ignition transistor. If you use LS coils you drive them at logic levels since they have a built-in driver.

Best I can tell, the low res signal starts when the corresponding cylinder is at TDC. So, don't forget that you have to switch to the #8 cylinder when you've found the #1 TDC low res pulse so that the PCM can fire the #8 cylinder in advance of the #8 TDC low res signal appearing.