DIY LTCC or similar system for LT1s

[vilefly]

HAPPY HAPPY JOY JOY! Congratulations on your success (in spite of lots of obstacles thrown in your way)!

So what makes you think the #6 piston is cracked? Unplug an injector or coil, and it went silent?

Looks like you need to run a RAMBo board in the 3d printer. It's what runs mine. But they are not cheap, yet robustly protected.
OCTOPRINT firmware for the raspberry pi can make it wirelessly managed on the home network.

[spfautsch]

So what makes you think the #6 piston is cracked? Unplug an injector or coil, and it went silent?

I'm just going by what my stethoscope is telling me, but thanks for the idea. I'll try disabling #6 in eehack tomorrow.

It's a strange noise that was hard to pinpoint before I yanked the head back off and looked it over carefully. Turns out one of my el-cheapo hardened guideplates wasn't so hard and the pushrod had for lack of a better term sawed a notch in the #2 exhaust guide slot. So it was making some high pitched valvetrain noise when it reached full open that made me think I had multiple valves making noise. I just wish I'd have noticed it before pulling the head off in-car.

Now that it's just the one noise it's easier to hear clearly. Sort of a high pitched type of piston slap that starts up after a few minutes of running.

I hope to yank it out before turkey day and pull the slugs to find out if I'm right or wrong. But more importantly this gives me the excuse I need to go to a stroker kit that I've wanted to build for a while but lacked the funds.

Looks like you need to run a RAMBo board in the 3d printer. It's what runs mine. But they are not cheap, yet robustly protected. OCTOPRINT firmware for the raspberry pi can make it wirelessly managed on the home network.

I'm not too worried about using my printer at the moment. I stuck it in a corner and haven't looked at it since but it could be the jb weld I used to affix the thermocouple ate into the bed heater traces and shorted. No telling. That project can wait. LOL, @ your raspi comment. I have enough embedded linux boards lying around to make half of the household appliances network enabled. Actually almost half are. But I prefer wired ethernet to keep the 2.4g noise floor to a minimum.

Anyway back to topic - I've spent most of the last several days working on a major rewrite that cleans up a lot of cruft that'd been hanging around since the proof of concept phase. The routines looking up dwell and managing when to energize the coils have been completely rebuilt and I believe are quite a bit more efficiently handled. So much so it may afford enough overhead to allow for cell smoothing / interpolation. I have a few obstacles I'd like to conquer before updating it on github. One such goal is removing the need for monitoring the falling edge on the low res signal after sequence is detected. So look for an update in a few days once I get to drive and test. The last test drive back in September was somewhat disappointing due to some intermittent stumbling. I'm hoping that's resolved now but it might not be, or it might turn out that it's completely unrelated to the ignition system. Unfortunately I chopped my harness up enough it would be a solid weekend project to go back to the distributor.

I'm also hoping to get time over the weekend to write up a quick assembly how-to for those receiving the sample kits.

[spfautsch]

Version 0.9.13 uploaded to github (link in sig). This version has NOT been tested extensively. Due to rain today I can't drive it because I've yet to seal the enclosure and don't want to risk water infiltration. Yesterday's test drive wasn't great but some small bugs have since been fixed. It felt like some cylinders might have been missing intermittently.

Major changes:

1) revised dwellLookup so it returns total # of degrees for dwell requirement in two components - # of cylinders ahead of current, and remaining degrees. Logging reflects this new convention so for example if spark advance is 28 and dwell requirement is 38, dwell will be reported as D0.66. Anything over 89 degrees will be D1.xx and so forth.

2) #1 was done to simplify the dwellCoils() routine so less math is required

3) main loop has major cleanup and some prioritization trees so more time critical functions are acted on with higher priority

4) some ISR cleanups and optimization

5) removed the low res falling edge toggle code in the ISR after sequence is detected

6) map sensing for accel comp adder has been improved and tested - optional output #1 on pin A0 is used to mirror the state of accelComp variable for in-car testing and tuning of map threshold

That's what I can remember offhand anyway. :-) As always smarter coding suggestions are always welcome.

I'm not going to have a chance to write up a proper assembly guide for several days, but here's a quick stab at what is fresh in my mind.

Attached is a bill of materials with resistor color codes for those who don't know them offhand. I would cite a few of the phrases I was taught in order to aid memorization, but that was from a much less politically correct era and I'm afraid it would get me banned. :-)

All values for required components are silk screened on the board. You will probably need a magnifying glass. If there is no value for a resistor or capacitor and only a component number that means the component is optional and will need to be selected based on your particular application for that input or output.

Obviously, mind the key marking on the Atmega.

The 5v regulator can be attached to the board with a 1/8" pop rivet or a 3mm x 8mm screw and nut. I don't believe a heat sink will be required but a small sheet metal style one could possibly be used if you'd like to err on the side of caution.

Due to the size of the component side solder pads, the crystal should probably have a mica insulator to prevent possible shorts. Since I couldn't find anything on short notice I used a business card temporarily between the PCB and the crystal's metal case while soldering to provide a small air gap. Solder these joints without going overboard on the solder and you should be fine.

The polarization marks on the electrolytic capacitors go in pointing toward each other.

Resistor values for R1 and R2 should be measured with a known good meter and recorded (ideally on the board). These values can be fed into the source to fine-tune the voltage divider network for the ignition circuit adc formula.

EDIT3: these values go here

// plug-in your values for R1 and R2 here (voltage divider network)
// be sure to include mantissa of zero or the precompiler will convert to integers and data will be lost
#define R1_VAL 9850.0
#define R2_VAL 3844.0

The LEDs go in with the short leg (cathode / negative) closest to the board edge. If you happen to cut the legs to the same length you can identify the cathode by looking at the internal mechanics of the LED. On these particular chemistry LEDs the larger piece (generally referred to as the "anvil") is on the cathode / negative side.

The plating on these boards is pretty thick so try to get everything right the first time. Desoldering attempts haven't worked out very well for me.

Good luck!

EDIT1: BTW, after programming be sure to disconnect the serial adapter or if you intend to log the datastream remove the connection for the DTR pin or the controller will be reset by the serial adapter. I use a six pin IDC male to male connector between two female connectors and offset it by one pin as shown in this picture from post # 356. Just keep in mind if your engine stalls immediately upon starting logging you forgot to disconnect the programming pin.

Also, for programming choose board type "Arduino/Genuino Uno" from the Arduino Tools menu.

EDIT2: It looks like I might be out of commission for a while now. I've been meaning to get a camera in the #6 and #8 bores for several weeks, and this is what I found on the inboard side of #6 cyl wall.

Pic (click)

I think I might have a broken or bent skirt there. Whatever it is, something's not good. Hoping to avoid the boring machine!

[spfautsch]

Sorry for the excessive posting but I parsed the 0.9.13 source with cppcheck and it found an out of bounds error in the est_isr() and a missing address of operator in getDwell(). Fixes are on github as 0.9.14. Also cleaned up some preprocessor warnings about narrowing casts. This build is untested but I will try to test base functionality (starts, runs, etc.) tonight. EDIT: basic functionality tests were good.

By the way, one added feature in this build that I overlooked in my previous post is the theoretical ability to handle ATDC (aka negative advance) spark timing. I'm not sure how to test this in-vehicle - possibly wait for the next ice storm and go out for a short drive to trip the ABS.

[Terminal_Crazy]

By the way, one added feature in this build that I overlooked in my previous post is the theoretical ability to handle ATDC (aka negative advance) spark timing. I'm not sure how to test this in-vehicle - possibly wait for the next ice storm and go out for a short drive to trip the ABS.

Serious question: What does that achieve?
Mitch

[spfautsch]

If you're referring to tripping the ABS, it's because kur4o has speculated that might be the only time ATDC timing is commanded by the PCM.

If your question is more broadly asking what does retarded spark achieve that would be for torque management - i.e. traction control, abs, between gears on a slushbox (auto trans), etc.

[Terminal_Crazy]

If you're referring to tripping the ABS, it's because kur4o has speculated that might be the only time ATDC timing is commanded by the PCM.

If your question is more broadly asking what does retarded spark achieve that would be for torque management - i.e. traction control, abs, between gears on a slushbox (auto trans), etc.

Sorry, must have missed that bit!
Thanks
Mitch