DIY LTCC or similar system for LT1s

[spfautsch]

I noticed some pattern where some coils will stay on all the time. You crank it for some time, than let it off without turning ign off, than crank it again and it will definitely lit some of the coils full time. You can see it in the video.

I'm seeing that in the 2nd video where 2 & 7 stick on (2nd crank in 2nd video). Can you isolate exactly where that is in the log?

It's hard to tell from the video how your wiring is routed but you might try separating the coil driver wiring from the input and uart wiring. You could be getting induced noise on the ground returns or driver outputs that's messing with the EST and opti inputs.

Because of the low resolution I can't tell but it looks like you might have something that could be (conductive) wire used as a loom wrapped around the bundle of wire coming off the controller. Is that the case?

Apologies if I'm missing it, but is there any video of the controller running (leds) while you're driving the two plugs side by side?

[kur4o]

You can count the ign on events to get where you want in the log. I marked the 2 most important areas where led freeze.

Are you suspecting that the est line is not driven by the pcm each events to generate more charge for the first starting event when it doesn`t crank the first time. Or is some bug in the pcm code, the est is run by the tpu so nobody knows what happens there. I think I have logged somewhere the dwell times at startup.


Noise can be a reason but in the test everything that can generate noise was disconnected including injectors, stock coil and icm, All of the COP coils.

I will try to make a video with controller driving #3 and stock coil the rest of the cylinders.

[In-Tech]

Sorry to interrupt, Kur4o, can you clear some PM space so I can send you a message? Or you can email me, efisytems at excite dot com. Thanks :)

[kur4o]

Sorry to interrupt, Kur4o, can you clear some PM space so I can send you a message? Or you can email me, efisytems at excite dot com. Thanks :)

You can try now.


I plan to verify that est line is not activated by pcm on second attempt theory by hooking a led to it and see if it matches the freeze pattern.

[spfautsch]

You can count the ign on events to get where you want in the log. I marked the 2 most important areas where led freeze.

I'm wanting to focus on the 2nd crank attempt in MOV2024.mp4 where 2 & 7 stick on and stay on during cranking. I'll try to isolate this today but might want you to confirm.

Are you suspecting that the est line is not driven by the pcm each events ...

The instances where a "stuck" coil is happening at beginnings and ends of crank events don't concern me whatsoever. But the crank attempt mentioned above is highly concerning because something was seriously malfunctioning.

I will try to make a video with controller driving #3 and stock coil the rest of the cylinders.

I'd like to see it start if possible. Don't try to jog the engine / make it malfunction. I want to see it crank and start. If it malfunctions that's great (well not great, but you get the idea).

Whats the overall state of health and age of your battery?

[kur4o]

I just tried to make est line drive a led but failed. Can I ask for quick confirmation est line output 5v when high and goes to zero when low. If I wire a led it should be driven right.

I will look for some schematics too. When I was doing the est led attempt I managed to freeze #1[It was on first crank], only got it on video, no log. Will upload video soon. Since now I never managed to freeze a led on the first attempt.

The battery is almost new and I guess a little discharged with the extended cranking. I will load it with some juice but for now it looks strong on cranking, given the fact I have some powermaster starter that is amp hungry.

[spfautsch]

Actually, I think I see a problem from the video.

Can you confirm the installation of your electrolytic capacitors? The colored markings are denoting the negative side, and it appears they're pointing away from each other. They should be pointing towards each other - i.e. toward the power ground input between them per the picture in this post [link].

Edit: I don't think the est line is a normal TTL output so it won't drive an led.

[kur4o]

I checked the board and they are reversed. I should have uploaded board picture much earlier. You suspect they can`t filter noise that way. I will try to switch them and see how it goes.

Meanwhile, here is the latest video + one older video where engine starts and run on stock coil and controller is hooked without coils.


Edit: The est couldn`t lit the led so I guess it is too low on current. Some really small led without resistor might work. Gm manual says to check est line voltage with voltmeter on AC scale. It will be between fluctuating between 0.5v and 4.6v


https://drive.google.com/drive/folde...OX?usp=sharing

[spfautsch]

I checked the board and they are reversed. I should have uploaded board picture much earlier. You suspect they can`t filter noise that way. I will try to switch them and see how it goes.

They're going to be fried - the electrolyte will have boiled by now. I'm surprised they didn't explode on you.

You're most certainly getting a very noisy 5v output from the 7805.

Be careful removing them you will likely damage the board. Maybe cut the cans off so you can expose the leads on both sides so they present less mass to draw heat away. The solder pads on the board are very thick and suck a lot of heat away from the component through holes.

The est couldn`t lit the led so I guess it is too low on current. Some really small led without resistor might work. Gm manual says to check est line voltage with voltmeter on AC scale. It will be between fluctuating between 0.5v and 4.6v

I wouldn't worry about it until you find replacement caps. The AVR is comparing the high value to a 5v reference that was fluctuating. Same with all the analog inputs.

[kur4o]

After some desolder torture the caps are off. The red one came fairly easy, but the purple was putting some strain on the board side ways, and give me a good 1 hour torture. Do you suggest not putting them back. They seem in good shape.

I will reverse them and give them a try. Do you have some p/n for new ones. Do they need to be canned. I can find easy uncanned, not sure about the canned.

I

[spfautsch]

I wouldn't recommend putting them back. You're only going to have so much luck removing them without damaging the board when you find out the hard way they're cooked. They work similarly to an alkaline battery, and the electrolyte is going to be broken down.

Part #s are in the bill of materials file that came in the release archive. It might be named "BOM" something or another.__

[kur4o]

I found some uncanned electrolyte caps in the solder drawer and put them for testing. Will try to find something more reliable later. I hope these are good enough for testing.

Hope to test soon and report back.

The board seem very sturdy and took lots of beating without issues. I wouldn`t have been so nice with the caps if knew they are toast anyway. I had planned to reuse them.

I can try to charge the caps and see if they hold any charge.


I really hope that the caps are giving me the troubles. Can`t wait to test the monster spark output and do some more comparison tests.

[NomakeWan]

It was absolutely the caps, and the fact that all of a sudden it ran like crap and wouldn't run right again is pretty indicative of the problem. Your capacitors are shot, do not rely on them for anything mission-critical ever again. Put new ones in their place, in the correct orientation.

[spfautsch]

I found some uncanned electrolyte caps in the solder drawer and put them for testing. Will try to find something more reliable later. I hope these are good enough for testing.

The bigger the numbers, the better. The ones in the kit were lower ESR (meaning they can charge and discharge more quickly than older chemistry electrolytics). I just copied the arduino power supply section, but I'm relatively certain the low ESR rating was factored into the design.

The board seem very sturdy and took lots of beating without issues. I wouldn`t have been so nice with the caps if knew they are toast anyway.

The problem I'm afraid you'll run into is that after applying a certain amount of energy (heat over time) to the pads, eventually the copper cladding will start to separate from the board substrate and you'll end up with lifted traces.

If you were stateside I'd say chuck it in the trash and I'll send you another kit. But I'm 100% certain you have the only diy-ltcc circuit board in the eastern hemisphere.

I had planned to reuse them.

Trash can.

Think of an electrolytic capacitor as if it were a balloon. There's capacity, or the overall size of the balloon - this equates to the microfarad rating. There's breakdown voltage - the resistance of the balloon to break under pressure. Then there's the equivalent series resistance or charge / discharge rating (ESR). Essentially this could be compared to the size of the balloon's neck - a larger opening will allow the balloon to inflate and (more importantly) deflate more quickly.

What you've done is turned the balloon inside out. It will work as expected for a while, but the internals aren't designed to be exposed to oxygen (reversed polarity usage). Think of the damage that occurred as having holes spontaneously appear in the skin of the balloon (think of how corrosion occurs), until you get to a point where it won't hold air anymore.

This is exactly why the first question I asked after you said it ran for a while and now won't run at all is "did you put the electrolytics in backwards?".

I really hope that the caps are giving me the troubles. Can`t wait to test the monster spark output and do some more comparison tests.

I also hope this is the problem, and I'm on the edge of my seat as well. Listening to your description I think it's safe to say your ICM is shot, or simply can't supply the current the MSD coil needs to charge fully.

[spfautsch]

kur4o has some new caps ordered and is awaiting their delivery. I was hoping to test some startup instances with a dead battery this week to see if I get similar "stuck" coils on initial cranking. As luck would have it I came home to find my PKE system (or who knows what else) had drained the battery below 11v after sitting for 8 days. Unfortunately I don't know if I even have coil indicator LEDs in the board I'm currently running, so those tests may be a week off even though half the work was already done for me. :-\

Meanwhile, since everything else has been great and I've been reduced to tweaking on mundane things like the instant fuel economy constants so the digital dash's MPG calculation matches actual, I thought I'd try pushing the envelope on spark advance. Today I had a first drive after adding about 4% globally to both timing tables (there was some smoothing and extremely high load cells excluded). Oddly, I found some really strange missing at approx 2400 rpm and < 50kpa. After looking at the logs it seems all these were adjacent to instances where 46+ degrees of advance were being commanded by the PCM. I suspect there was no "delivered" spark here - i.e. the EST line wasn't switching low after 46 degrees.

So I'm curious if anyone has ever messed around with the constant at 0x12045 in $EE "Maximum Spark Advance" which is 46 in all the base tunes I've looked at. Since I'm no longer limited to the sweep provided by the distributor's rotor, I'm wondering if I can't just bump this up to some ridiculously large number like 60. I can post a log, but after looking at what I have there's nothing I see that strikes me as actual detonation. The biggest knock events look like they were reverse traction loss when I downshifted from 5th to 3rd at 65mph (even though the rev matching was pretty darned good).

Edit: Overall it feels and sounds better than I recall from 8+ days ago. Unfortunately I didn't have the opportunity today to lug it hard in 6th gear, but the few instances I got close felt like it was on the bleeding edge of misfire. So I think "on load" spark advance may be close to the limit. I just want to push it as far as is possible so the limits are clearly defined.