DIY LTCC or similar system for LT1s

[spfautsch]

It's hard to put a finger on what all the issues might be, so I can't say positively that noise is one of them but it's highly likely. I unwrapped the right bank upper harness and tapped the 14ga pink wire for power Sunday. In the process I created a small rat's nest of wire that could be contributing also.

It doesn't help that I'm only getting about 90 minutes a day to work on it, and those are usually distracted minutes. Words cannot describe how anxious I am to get this nailed down and move on to yanking the engine and fixing the oil leak, etc. I'm sure that impatience isn't helping.

I managed to get an order placed yesterday with digikey for some rc filter parts, a spare boarduino kit and a couple proto boards. Hopefully I'll have that this weekend and can try cleaning up my wiring mess. Between chasing noise / count errors and soforth I've been contemplating how to package the thing for underhood conditions - would love to hear ideas here. Mine is going to be dead simple - potting the whole thing in epoxy. Possibly an external m+f weatherpack connector set for the power, uart and inputs (10 pins total if memory serves) and then either 10 or 16 loose-ended pigtails (8 ttl pins and either two or eight ttl grounds) for the coil igniters coming right out of the epoxy.

Work's been busy lately so I haven't had much time to research and audit code. Today I was hoping to look for a wiring diagram because the info at http://chevythunder.com/lt1_electrical_page.htm isn't close to what my Y body has - there is no tach wire on my black coil connector.

[vilefly]

What I am going to suggest might seem obvious, but it could be a lot of the problem. I suggest twisting the positive and negative voltage supply wires around each other. I remember the early fuel injection computers having issues with their injector drivers until they did that. Twisting the input signal wires around each other is a good idea, too. (5v signal + gnd). Twist all inputs and outputs if nessesary, but start with the heavy current users 1st.
I was putting together my Rostock MAX 3D printer 3yrs ago, and noticed in the forum that they have feedback issues, so I twisted all my wires in +/- pairs, and never had a problem. So from then on, I twist everything. ABS wheel speed sensors have twisted wires going to them as well, as they are touchy amplified inputs to begin with.

Worth a shot.

How about some PCM pinouts?
95 corvette pinvoltage chart1.jpg95 corvette pinvoltage chart2.jpg95 corvette pinvoltage chart3.jpg95 corvette pinvoltage chart4.jpg
95 corvette ignition schematic.png

[lionelhutz]

I would have to say potting can have it's downsides, especially with a hard material like epoxy. I have been involved with some items at work that were potted and failures were happening due to the potting and components expanding and contracting with temperature at different rates. Components were breaking and solder joints were getting damaged. Then, it can't be fixed either. Just saying, it's great to seal but can still create problems.

Now, if you can find a silicone based material that remains soft then you might be better off. Still can't fix it, but at least it avoids mechanical thermal stresses.

[vilefly]

I think a light layer with RTV silicone followed with epoxy for the rest of the cavity would balance the need for speed (curing time) and retain some degree of toughness. If it were fast, I'd use 100% RTV silicone, but it would take too long. Silicone dielecric grease would be nice(thermal conductivity), but may leak out before it "dries out" in high temps. A nice "chewy" potting material would be perfect, like the potting material used in Ford TFI ignition modules, and such. Chewy=tough.

[dzidaV8]

Have a look at ELASTOSIL products, I have used their two component curing silicone before with great results. I don't remember the exact number, but you mixed it with hardener, poured it into the case like water and it cured chemically, becoming reristant to over 150*C.
It is a german company, but I'm sure you have something similar in US :)

[spfautsch]

All good info, thanks guys. I hadn't given much thought to the possibility of thermo-mechanical stress problems from hard potting. The catalyzed silicone sounds intriguing, but I may go with some combination of silicone around the soldered components and epoxy over that for impact resistance for mine.

I didn't see any mention of a tach driver in those diagrams, but at least the colors match what I have. For now I'm going to labor under the assumption that the 94-96 Y bodies have a digital gauge cluster. Edit 2: found an addendum to the ltcc install document stating that only a few LT1 cars had the tach driven off the coil - 92-94 y bodies and 93 f bodies. Sweet - sounds like I can eliminate that with conditional compilation for those who don't need it.

I'm sorry to report I haven't made much progress over the last few days but I did get about an hour on it yesterday during which the outside temp dropped from 62f to 32f finally settling in at a balmy 16f. It's back to "stupid" cold. :-\

After cleaning up some changes I hadn't debugged fully I was able to get a better picture of what the microcontroller is seeing. This looks likely to be a case of the EST line causing induced or capacitively coupled noise on the low res input. Edit: the test coil was disconnected at this point so the microcontroller was only driving leds.

R1000:S29:D27:C1:E0:L0
R1000:S29:D27:C4:E1:L0
R1000:S29:D27:C3:E1:L0
R1000:S29:D27:C6:E1:L0
R1000:S29:D27:C5:E1:L0
R1000:S29:D27:C7:E1:L0
R1400:S29:D37:C1:E29:L0 << looks like low res rising edge is caught about 29 degrees early here - eerily similar to spark advance
R600:S0:D16:C8:E59:L0 << low res falling edge is then delayed another 30 degrees
R1000:S29:D27:C4:E59:L16
R1000:S29:D27:C3:E0:L0

If I decide to brave the cold tonight I'll try some external pullup resistors on these inputs since the AVR's built-in pullups are advertised to be in the neighborhood of 20-40K ohms.

Another possible source of noise may be a ground loop - I added a power ground for the microcontroller to one of the fuel rail bolts but didn't disconnect the other ground at the opti pigtail connector. I'm open to suggestions on what the best practice is here - do I need some type of filter for this secondary signal ground or should I do away with it?

[dzidaV8]

LT1 PCMs have separate Tach output, so don't worry about that function.

[Fast355]

LT1 PCMs have separate Tach output, so don't worry about that function.

The flash based LT1 PCMs do, the older chip based LT1 ECMs do not.

[vilefly]

Just in case, I would check the alternator output for ac voltage and voltage drops from the positive lead to the alternator connection. Mine had some issues, and I added an extra positive connection to the underhood fusebox when I replaced and upgraded the alternator. Since the optispark runs off of 12V, I figure this could help.

I tested the optispark with a regulated 5v power supply, and it worked. It was a few months ago, so I hope I am remembering it right. With a clean power supply to it (grounded to the power supply of the controller), it could help eliminate any 12v entanglements.

I wonder if feedback was designed into the coil drivers like nissans do. One can hook an oscilloscope to a nissan's coil trigger signal and see a 5v representation of the secondary output. Kinda neat, but can cause problems if you are not ready for it. I suggest running your coil trigger outputs through a buffer IC with 8 input/outputs. I think the proper term is "unity gain amplifiers", but they will probably be sold as buffers.

Probably could use a big, fat capacitor right at the controller for anything else I missed.