DIYEFI.org Forum

I have an idea for your timing light woes:

pull the opposing lead out and fully ground, or not.

I've had trouble with timing lights on wasted spark setups while cranking too, fwiw.

however the engine was running... why are you repeating that job? :-/

Okay, I'll try that. I had tried just pulling one plug wire in the pair hoping the bulk of the spark would go to just the one, but I didn't try grounding it.

Every time I pull the CPS I have to reset the timing. I could turn the engine to be TDC and get it close if I removed the cap off the CPS, but its a pain in the ass with various things installed (power steering, some coolant hoses) so I just do it like this.

Can't you key it before removal such as with two matching centre punches?

Pulling and leaving open one wire is a fail that can damage your coil from excess voltage arcing internally. Wasted spark coils rely on the other plug to BE the ground. Which is ground depends on what's in the cylinders. Having it open is sure to stop your timing light working :-)

I'm afraid to start making holes in things unfortunately. The CPS is geared to run perpendicularly to the cam itself, there's not really an easy way to see how it is lined up without taking the cover off of it. I could line things up if I drained the coolant again, removed power steering reservoir, but I'm lazy. Either way, I got it timed again, and I have the exact same problem. Tried it with 75, 150, 330, 1k ohms and .47 uF, .33 uF, .1 uF, and 0.01 uF. 75 ohms and before and it struggled to pick up the signal, anything above was pretty much all the same. I can usually get it to start, run for maybe 3 seconds, then it dies, more or less no difference from before the rewiring.

The zeal board appears to use a 300 ohm and 0.01 uF cap on the input -- thats the only difference I can see, and I've reproduced as per above, and it didn't seem to make much of a difference.

Thinking maybe the stock igniter wiring is subpar (I can't even see where it grounds?), I'll hook up the new igniter tomorrow with a good ground line. I think its more of just hopeful thinking at this point to think it might help, but seeing as I'm basically out of options until the new vr board comes, I might as well hook that up. Unfortunately I feel it'll probably be weeks before it arrives...


EDIT: nope, it loses sync about as frequently even with both ignition and injection disconnected. So new ignition won't help at all. I guess that leaves me with no options then.

Consider just giving it some throttle and making progress on some tuning, then? Its virtually impossible for the ECU to get out of sync and fire wrong events due to the compulsory single confirm. Therefore there isn't really any reason not to proceed IMO. Giving it more throttle and therefore revs will keep it alive even with losses occurring. Worst case is some shaking and vibration on the mounts. Not a huge deal, usually. Thoughts?

Hi Via,
I am trying to follow this thread but am confused / ignorant regarding your CPS.
In your first post you mentioned NE, G1, and G2.
When I Googled this, this seemed to imply a completely distributor based setup which contained three, single ended VR sensors.
But in your last post you mentioned -

via wrote:The CPS is geared to run perpendicularly to the cam itself, there's not really an easy way to see how it is lined up without taking the cover off of it.

So are all three sensors located somewhere else?
Maybe a picture and a schematic of how you wired it would help?


Thanks,
TonyS

Fred: I can't get the RPMs up past about 2000, it just doesn't last long enough to get up. When it dies, it almost always kills the motor immediately. I suppose with some tuning I could get it to stay up longer, but keeping in mind that when it loses sync it loses a whole cam revolution, its pretty easy to die in that amount of time.

Tony: http://www.supramania.com/forums/conten ... -your-PS2)

That link has lots of good pictures. Yeah, there are three pickups:


The nonturbo version of this car uses a distributor, so I imagine it was set up that way to be similar.
You'll note each pickup has two wires. Originally all the grounds for each pickup feed into one line (G-) back to the stock ECU. In trying to eliminate sources of noise, I have ran my shielded line into the CPS, one ground to each of NE and G1. In that picture, NE is at the bottom (24 tooth), wired to the primary VR and RPM0, G1 is the right (I think...but it doesn't really matter), connected to secondary VR and RPM1. The G1 input works wonderful, the NE misses teeth.

Edit:
Looking at some of the code for the decoders, I see that various ones set TCTL4 to specify how to trigger, where its set to both edges by default in init.c. Is this the correct setting for my 24and1 that triggers on positive edge?

Re the edit, damn it, so engrossed in what you were doing, I forgot about that completely. There is a branch with that edge reversed... and we've done some experiments but I couldn't figure out what the story was. With it wrong, it seems to work better. No idea why. With it right it seems to work worse. No idea why. This has been true of the honda VR stuff too, oddly. Here's the branch, back up your branch, then rebase onto this branch:

https://github.com/fredcooke/freeems-te ... h.edge.fix
https://github.com/fredcooke/freeems-te ... 63efad59d1

Perhaps it'll give you better luck, however you clearly had extra events showing up in otherwise clean cranking. The above "fix" would only manifest as an issue in the first place as the phasing of the edges changed, which doesn't occur at a fixed RPM.

Sorry for not mentioning this earlier! :-(

Fred.

I ran home for my lunch to give it a shot, compiled and loaded. Sync seems perfectly stable on cranking, but I can't get it to turn over. Maybe the timing is a good bit off now with the change? I only had a few minutes, so I'll look at it with a timing light later.

Yeah, I figured the falling edges should work just as well, unless they're really noisy compared to the rising edges or something...

Anyway, thanks for giving me this, I'll mess with it more this evening.

Edit:
Also, does the above discussion about distributor/cam have any bearing on which decoder I am using? It *is* a cam mounted sensor isn't it?

You're using the right decoder. Stable sync sounds good especially if you didn't have it before! And yes, the timing is definitely different with a change in edge.