Edit from prev post: 3D table means two axis plus values? I guess I was thinking of that as a "2D table", since that's a 2D array. Anyway, there's an boost control table for RPM/TPS which I bet could be easily perverted to some other analog in. So I'm not sure it couldn't be done. I hope not, since I was planning on doing it.


Wasn't someone (reverant?) working on a stand alone VVT board?! Anyway, overall, I think as long as you have a stand alone PID routine, the advancing or retarding could be taken care of externally.

 

There's an experimental, "vanos", phase logging tool in the beta I'm running. I haven't tried it yet. Everytime I upgrade, 4 things break. Right now hot starts are a 1 in 3 thing, there's a skip which wasn't there before, and the accel pump settings have some pesky factor of 8 or 10 or so in them - which you'd naively think you could just multiply all your numbers by the same factor, but they are artificially limited to be only the numbers the guy who wrote the code think are reasonable.

In other words, I could check out the valve, and the code, as soon as I test my adapterboard and get that project off my plate. :-)

 

Seems Kevin is working on designing a controller as well:
What have you done to your Miata/MX-5 today? - Page 81 - Mazda Forums

 

abe:
3D = 3 variables (2 in, 1 out) or if you graph it, a surface plot.
2D would just be a curvilinear plot

 

Huh, someone should invite him over here so we all can compare notes. His card doesn't look terribly complex, does it?


Semantics. Anyway, sure, 3-D it is then.

 

He's on here. Same name I'm pretty sure, just doesn't post all that often. He built a turbo Miata for the 05-08 GRM Challenge. Runs MS in his 90 with 99 engine in CSP. I think DIYAutotune sponsors him... at least I think that's his car in their add.

 

ok, pm'd him. Someone should PM him on the other forum.

 

Greetings all. Sorry I missed this thread as I don't log in as often as I used to. I am indeed building a VVT controller and so far everything is working short of a tune. The board uses the OE CKP/CMP sensors and corrects the signals for use with a typical NA MS installation. Furthermore, the controller operates the OCV via a 2D table via RPM vs. load. I'm considering correction for air temperature as my instinct says I should, pneumatically speaking, but I'll cross that bridge when I come to it.

Thus far, with a loose PID tune, the camshaft advance tracks the set point derived from an interpolated 12x12 table. There's a touch of over/under shoot, but that is easily corrected through the PID settings.

I'll check back and update as often as I can. In the mean time, I've posted a few pics of the process: Picasa Web Albums - Kevin - Miata VVT Con...

 

Nice little setup. I wonder how much you could get those signals from a MS, even an MS-II. Someday they will be able to modulate. Tell us more about the system!

 

What do you mean?

Any in particular? There really isn't much to it other than an Atmel ATMega168 with a bit of programming . It's fully standalone and the symmetrical tach output signals (CKP/CMP) can be used with just about any engine management system. The OCV is operated via a 12V PWM signal running at about 488Hz. The programming interface is laid out similar to that of MegaTune's and connects via an RS232 link. I still need to add a few bells and whistles, but nothing that affects the operation of the controller. Essentially, the MCU programming is complete and the Windows interface is right behind it.

I drove the car for the first time since Sunday this evening (been busy with a few other things on the car). I dialed-in the PID settings a bit closer with only a degree or so of overshoot) and settling in less than a quarter second (according to the logs, anyway). Overall response seems very good. I have a dyno session scheduled in the morning (Friday) to tune out the entire works). Wish me luck!

 

Very nice! Too scattered to think of a good question now.

I want to have the MS do this. I'd like to see the PID code, too. :-)

Oh, what I meant: The MS knows about all the events, from signal processing to timing, you SHOULD be able to get it out over CANBUS or something. The MS-II now (experimentally) reports cam-phase.

 

Agreed. It'd be nice to have this all happen within the MS, but this was a much quicker solution. CAN isn't out of the question. I have an additional SPI interface left on the processor that I can use for such a channel. I'll have to see how this develops in order to explore further options.

 

Nice work!
Is there a possibility of you documenting your work so it can be replicated?
Did you find that the P I and D values have to be different at different RPMs?

 

I don't know how the timing would be handled between them but if you DID you can, that's many fewer inputs you'd need and could spare them for something interesting - even if it's just turning on status lights it would be an AWESOME addition to the MS.

 

Great job!

So what Improvements did you achieve?


bye

Simone

 

Kevin,
Great work! I am in need of a standalone VVT controller. I have a TEC2 running an 01 VVT motor in my 94 CSP car and obviously it is not controlling the VVT. I used an MSD switch for a while but I need to have full control. So, if you decide to sell some, I am interested!

George
1994 MX5 CSP
1997 EK G2

 

Not at this time.

The PID settings did not have to be changed throughout different RPM or MAP bins.

I do have a number of spare I/O points available. I'm contemplating adding alternator control to one of them at the moment



Thanks!

The car picked up 32 HP over the '99 engine I had in there before hand!

PM Sent!

 

I am also in need of a VVT controller. If you are going to make these available in kit or built form I would be interested.

 

Did anything come of the VVT controller?

 

"Just" tuning the VVT behaviour?

That's great!!