navalhawkeye

I can't win with this thing.

Deets:
'96, Reverant MS3 Basic
'01 VVT

So, two weeks ago at autocross. I drive the car there, it's perfect. Change tires, walk the course, etc. I'm running heat two, so it's 11:30 by the time I'm running. Car is stone cold. Start it up, won't idle, giving it any gas shuts it off. Damnit. Change the cam sensor in grid because I have a spare, no change. Fine, I'll drive someone else's car.

End of the day, I swap tires back, it's 4:30. Car has been sitting since 11:30am. Starts up and runs perfectly fine.

Fast forward. I replace the crank sensor during the week. Drive the car quite a bit on Friday. Hmm, new sensor fixed a stumble that was happening around 2000rpm? Sweet! Drive it to autocross the next day, and get through the morning runs without a single problem - handily had FTD.

Car sits from noon to 2:30 when afternoon runs start. Same thing, won't idle. Tach is going nutso, presumably because it's losing sync. This time I have my laptop with me. TS is flashing between full-sync and half-sync.

Alright, maybe I'll get lucky and I'll be able to drive it home like last week. Nope. A friend was kind enough to trailer it home.

Take it off the trailer, still doing it. Whatever.

Tonight: go to start it and it's okay-ish. First start was a bit rough, and idle was bouncing around a bit, but as it warmed up it evened out. Now if I start it it's fine.

But, in TS the VVT1 err light is flashing every once in a while. It wasn't doing that the other day. It's staying full-sync as the VVT light is flashing, though.

Here are two composite logs...one from Saturday at autocross, and one from just a few minutes ago. Hopefully these can give someone that knows what they are doing a little insight that could help me?

navalhawkeye

Talking to a friend that came up with the following:

"I also notice that the cam pulses fall directly on top of the crank pulses in some cases, this is bad as the wheel decoder expects a cam pulse to always be be before the crank pulse, if it's after, it assumes the next crank pulse is the one it wants, but things don't add up so it loses sync."

https://www.miataturbo.net/megasquir...e4/#post785142

There's also a couple places where the cam doesn't follow the 1-2-1-2 pattern.


He suggested starting with changing from rising edge to falling edge and resetting the timing. I'm going to try that tonight and see what happens...

Tune is attached, with a screenshot of what he mentioned from the log.

Braineack

iTrader: (62)

looks like the crank sensor has reverse polarity from the cam.

a rising edge composite on your VVT motor should look like this:



Notice they are both vertical lines going UP.


That being said, I'd look into the cam sensor itself. Maybe the gap if not replacing it. You should always get 1-2-1-2 on the green line...unless the nubs on the cam wheel are worn or something.

I'm not sure what circuit he uses for the cam signal on his MS3 Basic so I can't suggest adjusting the sensitivity on that circuit.

Reverant

iTrader: (10)

It looks like the camshaft timing signal has slipped to the right.

navalhawkeye

Just to make sure...are we talking about the cam or crank sensor? I didn't think there was any adjustment in the cam sensor, as it just sits in the valve cover. That said, I'm already on my third cam sensor. I put a new one in at autocross two weeks ago, after replacing one last year trying to fix the issues then. The crank sensor is also new and gapped to .030".

I'm learning...what are the potential causes of this? Another friend (Leafy) said he's had the toothed ring on the cam spin, which I'm guessing would do that. Hopefully that didn't happen to mine.

Thank you for the responses...please keep the help flowing

navalhawkeye

Changed the ignition from rising edge to falling edge and reset the timing.

Seems to have corrected some of the issues with the first log, but still doesn't look exactly like the one Braineack posted...

Braineack

iTrader: (62)

that looks good. it ultimately depends on how the unit is built.

Reverant

iTrader: (10)

Check your base timing again, as it will shift several degrees with this change.

Copper280z

I'm the (not-Leafy, obviously ) friend mentioned above. To my knowledge the wheel decoder requires the cam pulses to be between the last and first crank pulses. If that's true, then I think when the cam advances with vvt he will lose sync.

I'm thinking that either the cam is off a tooth (or positioned incorrectly for some other reason) or the toothed wheel on the cam has moved. I'm thinking this because it seems there is no adjustments for the cam or crank sensors(is there?), yet they don't seem to line up as expected with rising or falling edge.

I asked him to take a composite log with the cam fully advanced (which might require it to be running above 3k? I'm not all that familiar with the miata vvt) so we can make sure that the MS is happy with the cam signal at all points.


-Bob

navalhawkeye

I did...I think had to move it 4.5-5 degrees to get it back to 10 degrees.


I'll post a log doing what Bob mentioned (with the cam advanced), and double check that the cams are timed right tomorrow.

codrus

The Miata sync logic in 1.3.4 is pretty simple. Really all it cares about is that there's at least one cam tooth between crank teeth 4 and 5, and between 8 and 1. The double tooth on the cam is only used to get initial sync, once that's established doesn't care any more. A comment in the code suggests that it used to check for the double cam tooth, but that was removed in order to make it work with the VVT heads.

--Ian

Copper280z

Ryan, Leafy, and I worked on this for a while today, well Ryan did at least. He took 4 logs, full advance and retard with rising edge and with falling edge. Today we can see that with rising edge and full retard (connector unplugged) the cam is advanced slightly from the position it was in on 6/27. Now the last cam pulse is clearly before the #1 crank pulse. We don't understand what could be allowing the apparent cam position to change. As the car sits now it the car looks like it should, and does, run correctly. But without changing anything, a few days ago it didn't work, and it looks like it shouldn't have worked.

With falling edge, the fully advanced cam position has the first of the double pulse crosses the #8 crank pulse by quite a bit. From the sounds of it, this would work, but I don't like it because I think it wouldn't have the information required to re-sync if it lost sync with the cam advanced.

None of the logs look quite like this:


Earlier Braineack mentioned that one of the signals looked inverted. I don't know enough about the box here to say, but I'm pretty sure on my v3.0 board I had the option to use the standard signal or the inverted signal. Is there any option to invert the signal of one or both sensors here? If one is inverted then it could cause an apparent shift left or right of one tooth width, since the rising or falling edge option changes how it looks at both sensors at the same time.

I'll let Ryan post the logs.

Leafy

iTrader: (1)

Something's real fucky with this car. Sometimes it makes a full retard log that looks fine, sometimes it looks off like 15deg.

codrus

Check the crank for keyway wear?

--Ian

Copper280z

I'm with you on that, pull the balancer and inspect everything about it and the wheel very closely.

navalhawkeye

Whoops, didn't realize I posted this morning...

navalhawkeye

Big thanks to Leafy and Bob for helping me look into this. Bob has spent way more time than I could have expected looking into it with me.

I'd be happy to take it apart again to check the pulley. All that matters is that the key holding the timing belt gear, and the trigger wheel in place have no play, right?

Copper280z

Yep, what we're looking for is slop in anything that connects the crank trigger wheel to the cam trigger wheel.

Copper280z

Just as an update, for people in the future, the solution was going to the flyin miata 36-2 trigger wheel. It totally solved all of the problems, He installed a new head and cam and got exactly the same problems, no slop in anything. At the DC pro I spent a good portion of the weekend playing with it, fixed many other things, but this wasn't one of them. 36-2 went on and it's been great ever since.