Sweet, so is it safe to congratulate you now? ;)Originally Posted by spfautsch
And I did run the compile with this version and just as before, no issues. Interesting that you had rain too; we had the first rain in almost a month and I forgot just how slick it can get under those conditions. Had a bit of a brown-pants moment. All good though, safe and sound and looking forward to finally being able to throw this together and put it in the car.
1990 Corvette (Manual)
1994 Corvette (Automatic)
1995 Corvette (Manual)
Kits are $25 each including postage for domestic US. Check your PM for my paypal info.
That cap is on the 5v side of the reg so it should never see anything close to 16v. But I've never officially been a practicing electronics engineer so I'm open to discussion on the benefits if there are any.
A heatsink wouldn't hurt but I don't feel it's needed. I'll have to check my notes to see if I measured what the controller was drawing on the bench, but I used the far less scientific method of feeling it with my index finger after running the car for about an hour, and it wasn't noticeably warm. ;-)
It has done nothing but rain here. Seems like the last time we went more than two weeks without rain was back in November. I normally wouldn't drive it in the rain but I'm tired of postponing. It finally got a much needed oil change this morning and I'm debating whether to give it a bath now. The meteorologists say no more until Friday night, but they completely missed the call yesterday.
Hiya,
Paypal sent, thank you.
My spark tester/distributor machine/8 plug pressure chamber also has a belt driven alternator on it so I can simulate that "noise". I hadn't looked at the schematic that close and you're right, if it's on the 5v side of the regulator that cap should be more than adequate. I'm going to build as the kit was designed, test and post any suggestions I see. When all that's done I gotta find a donor car
-Carl
Cool, if you have the capability I'd be interested to see how it handles the voltage spike when the battery gets disconnected. I'm not terribly interested in trying it on my 25 year old buggy for fear some pricey or difficult to replace box like the CCM gets zapped. Just a forewarning, I've seen some strange behavior with USB to RS232 adapters while firing coils on the test bench. Routing the coil power, spark leads and uart wiring away from each other seems to help. Luckily I've not seen anything like this in the vehicle.
The only thing I think I'd do differently with the PCB design at this point is omitting all the cylinder leds so the ADC input traces aren't cris-crossing some of the led driver traces. Although they were instrumental in initial troubleshooting efforts, there's almost no point in having the leds except on a test bench setup.
Well, actually make that two (or three) things. I forgot to put any silk screening around the ICP header and led ground jumpers to let the user know what's what. But the manufacturer had a limited time 20% discount and I rushed making the changes to get the order placed.
Planning to drive it to work the next couple days without logging anything, and hopefully be able to just enjoy the machine. I managed to scratch a couple year-old items off my honey-do list today, then vacuumed the footwells and washed her down (which makes it nearly certain there will be rain tomorrow). Then I gave my 16 year old nephew a demo he's been dropping hints about for at least a year. He got to smell tire and clutch smoke, there were large smiles on our faces, and we had a good talk in the process.
Speaking of, are the cylinder LEDs integral to the circuitry? If I omit them and their current-limiting resistors will the circuitry still run? Maybe jump the terminals?
Glad you and your newphew had fun. Always great to inspire the next generation to keep these things around!
1990 Corvette (Manual)
1994 Corvette (Automatic)
1995 Corvette (Manual)
None of the LEDs are required but I would recommend putting the power, status and spare output leds and resistors in. This way you can enable the INPUT_TEST conditional and be able to locate tdc and even sequence if you want. I have the spare output wired to a LED next to my cluster in the dash - I'm using it to drive a relay that controls power to my wideband controller. But with input test enabled I was able to put the trans in 4th and roll the engine over to test starts where the low-res beam was in one of the sequenced trigger slots (search the source for: detectSeqInverse()).
However, if you have a rev 1.1 board there are two jumpers - one for the three status leds and the other the coil indicators that disconnects their respective ground planes. This is equivalent to omitting the parts completely.
Oh yeah, having the status LEDs was going to happen, but I just wanted to see if the per-cylinder LEDs were integral to the actual firing of the coils on the rev1 board. Actually just looked at the PCB and saw that yeah, they're coming off of totally separate traces to the same pin, so they'd have zero effect if they were omitted. I do like the jumper arrangement on the 1.1 board though, that's neat!
1990 Corvette (Manual)
1994 Corvette (Automatic)
1995 Corvette (Manual)
I need some quick guide on the controller wiring.
Can I use the opti wiring as a B+ and ground for the controller.
So from engine harness I need to connect Hr, Lr, B+,G, Est and coolant.
Are the grounds on the board are common or on a separate paths.
G3,tx,rx,dtr are needed for logging only.
What is STS.
rev1.0 board.
Any hints are welcomed.
You should be able to hook into the opti wiring for all of those, yes. The grounds on the PCB are shared via a common ground plane.
STS is the optional Status LED.
1990 Corvette (Manual)
1994 Corvette (Automatic)
1995 Corvette (Manual)
I would recommend connecting VIN and GND to the same circuits your going to be powering and grounding your coils with, definitely not the opti power because GM used such small gauge wire for this. I suggest this because the system voltage is measured between GND and VIN and we want a direct picture of the voltage the coils are seeing so dwell can be compensated accordingly.
G2 should be connected to the opti ground - think of this as a PCM signal ground reference.
More importantly for flashing firmware updates.
This allows for an externally wired LED that mirrors the yellow status indicator. O1 can be used similarly for the red OUT1 indicator, or whatever. As mentioned a few posts back the INPUT_TEST mode will show the low res opti input state with the OUT1 LED, and the high res on the STATUS LED.
I'll try to make time soon to write up a proper install guide. Documentation isn't really my favorite, and I frequently describe things in terms that only make sense to someone on the asperger's spectrum. So I may need help translating some ideas so they make sense to a wider audience.
So far everything's been good. I've put about 400 miles on this firmware version with nothing I can attribute to the controller save one missed start yesterday afternoon. I believe this may be a minor bug in sequence detection - I've had two instances (over the entire life of the project mind you) where it's refused to start immediately and feels like starting a distributor car with plug wires badly misplaced. I suspect this is an edge case happening only when the engine comes to rest at a particular location and the inverse sequence detection isn't just right. I'll try to log every start going forward but I may have to get more aggressive with trying to reproduce because I'm fairly certain it's only happened twice out of at least 600+ starts.
That clears most of the wiring questions.
Opti B+ is filtered and stabilized through pcm and is not suitable for voltage sensing.
Do I need to do anything with 01,02,03,in1,map. I guess one will be reconfigured for coolant.
That status input is great idea, In the log is there a field with degrees and cyl id to check if the #1tdc is spot on with opti wheel. I am sure some error is possible due to tolerance not being too tight. A great tool for dialing the engine ignition components.
I plan to check wheel position == #1tdc with it.
I will make some install guide when I get to it. It will be just for reference for other guys to avoid some stupid questions I have now.
Still thinking where to get power for coils. Does it need to have a relay or per bank fuse, or should I use the injectors per bank wiring.
To be fair, I've had a few missed starts on my stock car. It's rare, but every so often I'll turn the key and release it and it won't actually catch. Always makes me cock an eyebrow since the LT always fires so quickly. So if only one out of every several hundred starts just happens to miss and you have to go for the key a second time, well...that's not awful. If it turns out to be an edge case that's restricted by the opti wheel design and not your controller, I certainly wouldn't lose sleep over it.
As to your documentation comments, I actually rather enjoy writing documentation. I'd be more than happy to help however I can.
1990 Corvette (Manual)
1994 Corvette (Automatic)
1995 Corvette (Manual)
Nothing for O1-O3.
IN1 is the ADC input for the ECT wire - PCM connector C20
MAP is the same for MAP sense wire - PCM connector C23
Not really. I have a cylinder number static in the sequencer array that will identify what cylinder is being dwelled, etc. because the index for the sequencer array doesn't correlate. I.E.
Code:seqIndex cylNo -1 <sequence not detected yet> 0 1 1 8 2 4 3 3 4 6 5 5 6 7 7 2There is an error tracking mechanism built-into the sequencer struct. I found out early on that the high res interrupt doesn't always count down from 90 to 0 on each cylinder. Sometimes it only counts down to 2 before the low res signal resets the counter, and sometimes it will count > 90 and wrap around to 255 or 254 (the variable used is an unsigned byte). I've always assumed this is due to slop in the timing chain, but who knows. I just know it generally changes as the oil temp comes up. The tracker only pays attention to counts at zero and above, but discards counts > 90 where the variable wraps around.
If you take a look in the file 'diy-ltcc-logging-pids.txt' there's a definition of what you can log. For example if you open a serial terminal to the controller while running (115200,8,N,1) and send the string 0L123458 you'll get some board information (0 is zero) and then a delimited stream every second until you turn it off or send q.
All the data denoting "index" means it's the index being used in the table lookups. So volts index and ect index will tell you what cell is being grabbed from the coil dwell table for base dwell time. Then the map and rpm indexes show what multiplier is being grabbed from the map_v_rpm_mult.h table to increase / decrease base dwell. I'm using this multiplier to reduce base dwell below max for "normal" operating ranges, and moving closer to 128 or higher for cranking and other areas such as WOT.Code:<firmware_version>,<R1_value>,<R2_value> <runtime_secs>,<rpm_index>,<volts_index>,<ect_index>,<map_index>,<error_degrees> (second line repeats continuously)
I've got a start on a lot of this info in README.txt
It just needs to be tied to the switched IGN circuit from the lock switch in the column. I wouldn't recommend putting any relays in this circuit to reduce possible points of failure. It's up to you if you want to get more complex and have some fault tolerance with multiple fuses, but mine is simply tapping the pink power wire that was feeding the original coil. It's either 14 or 16 gauge wire so if you're worried about voltage drop a heavier wire to the lock switch would be warranted.
Just keep in perspective this wasn't meant to be equivalent to a 24 / 27 / 54x conversion. There are certain things it's simply never going to be able to do (like compensate for large instantaneous jumps in spark advance).
Last edited by spfautsch; 05-22-2020 at 07:02 PM.
The ign coil have a 10amp fuse and 0.8 wire that can be fed to coils but it also provides power to auto trans solenoids. Still have to think how to wire it. I might run a separate wire in the end with relay that will feed juice straight from battery with thick wire, to avoid any voltage drop[Will use the coil feed to trigger the relay on ign on]. I do like stock look though.
When I finally ripped the opti apart, I notice that the high res beam goes through 3 slots to the opto transistor. low res have 1 slot. That 3 slots explains the counts inconsistency. Even pcm have built in count errors +-4 high res pulses.
I also found why opties fail so much. It is the only japanese part in the car. There is some dot welds that eventually will broke due to high frequency switching and cause signal loss. Not an easy fix since soldering wont last long there.
That status led is great idea. You can even expand it more to add some blink codes on errors. 3 blinks = you are F...d up, low res signal is lost.
Is there a way to get feedback from coil if it is blown or not delivering spark for some reason.
What do you mean doc, all the best stuff is made in Japan.
1990 Corvette (Manual)
1994 Corvette (Automatic)
1995 Corvette (Manual)
Posting Permissions
Reply With Quote
Bookmarks