My 87 toyota supra hardware build

[via]

Yep, I'm sorry I was writing that from memory and forgot the first two were the rpm inputs.

Good to hear re: injection timing.

[Fred]

Just went through your config changes and have a tiny bit of feedback for you pasting from IRC:

[19:56] * WyomingJAFA wonders if via is still awake
[19:56] * WyomingJAFA adds via's remote
[19:58] <WyomingJAFA> via, not too important in this case, but and nevertheless, you're working from my master, right repo this time, but not the most recent code.
[19:59] <WyomingJAFA> dev branch is a superior option usually, but not guaranteed.
[19:59] <WyomingJAFA> it's certainly superior if you're applying changes
[19:59] <WyomingJAFA> because otherwise you'll have to rebase up onto there anyway
[20:03] <WyomingJAFA> first issue, but i think i told you this on the phone, you have 400rpm hysteresis :-o
[20:03] <WyomingJAFA> lower this to 200 and 100
[20:04] <WyomingJAFA> 200 for ign, 100 for fuel, such that fueling is back on for 100rpm delta before ignition hits
[20:05] <WyomingJAFA> otherwise all looks fine
[20:05] <WyomingJAFA> unsure if it's complete or not
[20:06] <WyomingJAFA> would prefer VIASUPRA or something, does it have a nickname?
[20:06] <WyomingJAFA> My "TRUCK" one is shit...
[20:06] <WyomingJAFA> as is "PRESTO" because presto might have 3 cars...
[20:09] <WyomingJAFA> SEANKLT1_ID, SNOTROCKET_ID, SLATER_ID, PETERJSERIES_ID, DEUCECOUPE_ID, PETERTRUCK_ID, DEUCES10_ID
[20:09] <WyomingJAFA> ^ good ones
[20:09] <WyomingJAFA> either dude+car or car's nickname
[20:09] <WyomingJAFA> the rest are bad
[20:09] <WyomingJAFA> sorry

Sched config looks perfect. Ideally your injection wiring should match your ignition wiring, but it's really not that important.

[via]

I was not awake, sorry. I'm always on so I did read it in the morning.
I will make my fork as per your suggestions/requests, and will switch to VIASUPRA. I'll hopefully have that done tonight.

Today was a big day. There were many ridiculous fuckups on my part however. For example, the most damning one: I started making the connections for the injectors, when it seemed the first pair was already wired up to something. Specifically, it was what was controlling my main relay for power to the ecu. I honestly have no idea how the hell I did that. IG SW, which is what it was supposed to be conneted to, isn't even in the same connector, its on the other side. I fixed that, but obviously that suggests that whenever I have had the ecu on this entire time, it has been powering through the fuel injector. Thankfully up until one switchon today, there has been no fuel pressure, but I'm sure I still flooded it to a degree. I spent a good amount of time cranking with no ignition/fuel just to try to make sure it was relatively clear.

Once I had that straightened out, I tried starting, and at first it only sounded like it wanted to catch, then finally actually did. I had a relatively smooth idle at 1800 rpms for about 30 seconds. It would occasionally miss, but according to my WBO2 it was running at about 10:1, so I'm guessing that is the explanation for the miss? The engine eventually stalled with a nice exhaust backfire.

I tried starting again, but I couldn't get it to turn over. One of the times I tried, the act of turning on power to the ecu, and thus powering the fuel pump, caused a terribly loud backfire in the exhaust. This kinda scares me, as it suggests that something sparked back there (in-tank FP)?

Next fuckup was my inability to do basic math, where in my head I needed to increase VE to reduce fuel. Needless to say I couldn't get it to start, and I hit another loud backfire. Spent some time trying to clear it out with no ignition/fuel connected, but still couldn't get it to start or even feel like it was trying to catch. Hooked up a timing light, had no spark. Checked plus, they're beyond fucked.

So I replaced them, took a while as I had to go into town to get replacements. I got a bunch as I see this happening a few times. I still can't get spark however. This leads into my current problem, the first output for ignition appears to be constant high now. It wasn't like this before, but I'm seeing some weird behaviour that can probably be blamed on shitty wiring but I figured I'd run it past you. I have three LEDs, one for each ignition output. If I apply power to the ecu, IG1 (originally from PT2 to the line driver) is lit. If I remove the connector which also removes the ground for ignition, it is still just the tiniest bit glowing. The other two LEDs also appear to very mildly glow. If I remove the TA board from the sockets, all LEDs stay off. When its all hooked up, measuring the TA board pin voltages for PT2-4, 3 and 4 are 0, but 2 was on the order of 0.2 V, not quite zero. As far as I can tell there is no wiring issue, the TA board is outputting a voltage that is not precisely 0. The line driver is TC4427CPA-ND. I've currently pulled the board out and am trying to debug it to verify there is in fact no wiring issue.

Anyway, until ignition is going to fire, I will not try starting again. Either way, I *did* have at least 30 seconds of idle, so its a borderline success.

[via]

I replaced both my line drivers and I haven't see any issues, no barely glowing LEDs, nothing. Not exactly sure what was wrong with them. Either way I hooked it up, and also reduced the VE for the whole idle area to 40, hoping to bring the idle closer to stoich. I think I need to compensate with the cold enrichment, because it took all of time to get it to start. I can't get it to run for a full 30 seconds like last night, but it did run a tiny bit. Not quite sure whats going on there with the WBO2... but I know its more or less useless in situations like this.

Anyway, in the attached log, I'm getting sync loss, which is probably what is causing it to die, aside from i'm sure fuel being way off. Since I haven't gotten to this part yet, I'm wondering how to interpret how a sync failure occured, since to me it isn't obvious that anything went haywire with the CPS inputs.

Edit: If you haven't already looked at that one, look at this second log, it has a lot more examples.

[Fred]

Yes, noise is certainly knocking it back every time. The RPM falls a bit, then recovers, but some of them are happening a lot in quick succession, which basically means no fuel and no sparks and a dying engine. If you look at the temporal spacing of the secondary increments, it's even. The noise is arriving on your primary input. Some of the codes are "too early" and others "too late" and still more "count too high" which just means it got a noise spike near the end and it knew it was noise because there should have been a sync pulse by then. You can subtly see the step in some of the primary traces. Screeny:

[via]

I intend to spend the day fixing it, but I'm not 100% sure where to look. I figured if something was majorly wrong with the way I wired this, both primary and secondary would be experiencing noise.. So I'll start by making sure the shield for the primary is grounded to the ecu. However, if it is, the problem must either be with the last inch of unshielded wire at the CPS itself, or maybe a dirty pickup inside the CPS? For either of the latter I'll need to drain coolant and pull the CPS out, so I'm hoping its just a shielding issue.

Any other things worth trying?

EDIT: I checked the shielding, rerigged it to ensure its grounded, and it is. No noticable difference, it might even be worse, but that could also just be my battery getting lower and lower. Its rainy and very muddy out, so this sucks a bit. I just pulled the CPS out, and I'm going to put a new connector on it and try to get the unshielded trace as absolutely short as I can get them. I'm also going to check the air gap and try to clean it up.

I do just want to make sure I'm not doing something dumb. There are four wirse from the CPS: G1, G2, GND, NE. Right now G2 goes nowhere. Each of G1 G2 and NE have a line coming from it. Each one is two insulated wires wrapped in shield, with insulation around that. I have the two inner wires joined on each end (its like this for all my sensors).


Another edit:
I redid everything. I rewired the CPS on the CPS end, I even eliminated the connector on the ecu side and ran the shielded wire straight to the lm1815 board. Its hackier now because of that, but its shielded all the way. Checked the clearances on the CPS air gap, both are well within tolerance. I'm waiting on a friend to reset the timing while cranking, but I still seem to lose sync occasionally when cranking. I'm basically out of ideas.

[Fred]

I wouldn't have gone to that extreme. What sensor ICs are you using? LM1815? MaX9924/6? In either case a shunt resistor is a good thing, are you using one on each input?

Read this, especially the section on noise, and see if it helps: http://brickems.com/brickrpm/install/

Routing of your cable is good? not near ignition power/ground wires?

If you have LM1815 then there are, I think, some configuration resistors you can mess with. Maybe try those.

Re better/worse, that log showed some bits that seemed OK and other bits that seemed bad. One comment: You could put some foot into it to bring the revs up and keep it alive, if you want. SyncConfirmations has got your back re the dodgy signal.

Fred.

[via]

I've tried various shunts: 150, 330, 450, 1k, 2.4k, infinite, in combination with various filter caps: 0.01 uF, 0.1 uF, 0.33 uF, various different ways to run the wire, putting pieces of grounded metal between the cps and the coils, and even reversing the two lm1815 circuits (using vr2 for primary, vice versa), and nothing is any better, most are worse. I think I'm ready to give up on this shitty little board. I ordered a max based vr conditioner in hopes that it works better. I'm also considering buying different shielded wire. I'd have to buy that online though.

I'll note that when I reversed the vr circuits, it was the secondary (the 1 tooth) that struggled, so I'm pretty sure its the lm1815 that is being shitty. Buying that board was a mistake in the first place, imho, as the two vr inputs have completely different circuitry, and it was "optimized for the rx7"...whatever that means.

If you have any other suggestions, I'll try them, but otherwise it looks like I'll be waiting a few days for the board.

Edit: in the meantime, I'm going to start wiring the CPS for independent grounds.

[Fred]

Which board did you buy? Super curious now. RX7 setups are the same as the Toyota ones, FWIW. Likewise the humble F2T engine shares it, too. I'm sorry I can't sell you one of mine! Though I'm going to get started on that in the coming weeks if there is nothing more productive to do.

Proving that the VR circuit on the opposite input had the same issue and the issue vanished from the place it was, is valuable. Is there any way you can fix the topology of that bad VR circuit to match the good one? Seems like it could work? Could the lm1815 chip itself have a dry joint that's wiggling and causing a bad signal right at the board? It's not in one of those dual wiper sockets, is it? If so, try removing the socket and direct soldering it? I'm out of suggestions :-)

[via]

Its the Zeal engineering dual VR board.

I re-flowed every solder joint on the VR board earlier, and they are directly on the pcb, so no sockets to mess up.
I completely redid the CPS wiring again as well. The grounds are independent, and the shielded wire goes all the way into the CPS to the pickup. So there are zero inches of unshielded wire in the engine bay now.

However, during cranking I still have the occasional sync reset, sometimes none, sometimes as many as 7 for 19 cam cycles. I'll need to try the separate grounded setup with the resistor combinations above, however, I can't seem to get the timing set. My timing light is firing intermittently at best, and for most of the time wouldn't flash at all. I'm somewhat pissed at my timing light for this. I even took a spare plug and hooked it up, and could see sparks on it while the timing light (connected to that cord) would not fire. I tried playing with dwell some, but couldn't get the timing light to fire reliably. Without that, I can't set the timing enough to try starting... Maybe its temperature dependent? Its a lot colder than yesterday when I had a lot less trouble with it, although even then it would at best fire a couple of times for 5-10 seconds of cranking.