JasonC SBB

Coolio.

What's your control strategy? (e.g. PID, hysteretic, etc)

Reverant

iTrader: (10)

I'll start with P and work my way up to I and D if needed.

Jim

AbeFM

iTrader: (3)

Just what I was thinking.I have an unemployed electrical engineer as a friend, and I keep trying to get him to do something like this. Guess I'll have to come up with something else for him, now. :-P

Thanks for the pointers on the schematic.

mistersolo

I finally got around to getting a scan of the VVT cam and crank signals with the Oil Control Valve (OCV) plugged in and unplugged. Please excuse the crappy pictures. The top picture, the cam overlapping the crank signal, is when 12v is applied to the OCV.

AbeFM

iTrader: (3)

Awesome!

That's a lot of shift, wow... Anyway, the first thing I notice is that the two pulses can fall on opposite sides of the crank tooth - it'll take a little work with the MS code to figure out if it matters. I think it uses the number of teeth between crank teeth as the sync signal. If it uses the number in one full rev we're ok, if it's between adjacent teeth we're in trouble.

Also, unfortunately, the gap in the trace kinda blocks out part of the trace - really, putting the scope in "single" mode would work a lot better! But either way, that's a big help.

Will reflect on it. :-)

Joe Perez

iTrader: (8)

Are you MS2 guys looking at rising edges or falling edges of the CMP? Because it looks to me like all of the rising edges always remain on the "correct" side of that TDC CKP pulse- only the first falling edge of CMP appears to cross the line.

Take off every 'ZIG'!!
You know what you doing.
Move 'ZIG'.
For great justice.

JasonC SBB

The problem with P only is the steady state error. The problem with a (varying) steady state error in your VVT and a speed-density (MAP) system (as opposed to a mass air (MAF) system), is that your AFR will vary with the VVT phase error.

What is the mechanism of the VVT? What's the transfer characteristic of VVT solenoid to phase? This characteristic will greatly affect the control strategy. It's possible that the mechanism has an I characteristic, so what you actually need is a P and a D, and perhaps the "null" point should be 50% duty.

AbeFM

iTrader: (3)

I believe it is the falling edge, I know it's the same edge of both. And I also think it's the trailing edge (actually rising) since the leading edge is (curiously) unreliable.

Heh - I sure as heck don't know the transfer characterisitic of VVT. I think that's one thing that makes engineering princebles harder to apply outside of the design phase, you don't always have the answers you want.

I would imagine, however, that it depends on oil pressure.

My thought was a fixed offset, especially is small, would be ok, since all that really matters is how the motor runs there.

Joe Perez

iTrader: (8)

So which is it? The trailing edge is the rising edge according to those scope shots. If it's the riding (trailing) edge, then you guys should be in good shape.

AbeFM

iTrader: (3)

Sorry, I can never get used to this negetive logic, it's the same slope on all traces, user selectable. However, using other-than-trailing (rising) edge is a bad idea, the sensors are unrelaible there.

The VVT pulses are different shapes, interestingly. However, as the MS-II code is now, I don't think it matters. Only number of edges matters. But I don't like how there are two edges on opposite sides of a crank pulse. Guess one could either: try it and see, ask the MS-II guys, or move the cam sensor on an earlier car to emulate it. :-)