M54. Let the foolishness begin.

[evil_twin]

so you gonna let me drive the car like you did before? lol.

[e30polak]

You are the fucking man! That is all. :D

[hoveringuy]

Mechanical miscalculation

I needed to cut the ABS waterpipe because it wasn't routed right and as a consequence grafted the brass fitting onto it. I used fiber reinforced epoxy, the best stuff I could find. I couldn't even imagine pulling it off when it cured and thought it would do fine under cooling system pressure.

It was great, until recently.

Under the onslaught of over 200 hot/cold cycles the epoxy cracked internally and it drips now. I find my engine mount filled with coolant every morning:(

The answer, I think is that the hose needs to go OVER the 1" ABS pipe and then mate with the rest of the 3/4" heater hose. http://www.pegasusautoracing.com/pro...6110&iorb=4764

I think that will work.

The good news is that the VANOS controller has progressed nicely. The car is a blast to drive!

[328ijunkie]

Wow thats badass

[hoveringuy]

Sounds like you're all over it Steve!...now, any chance you could write "PID tuning for dummies?" :)

That's the book I was looking for to do this.

The PID part wasn't that tough once I got the noise out, that was the hard part. The solenoid is an inductive load and doesn't like being switched on and off very fast; it spits out huge EMF spikes that feed back in as valid (but wrong!) input signals. That was solved with larger resistors to the MOSFET gate to slow the ramp-up and some strategic capacitors. I tried just filtering them out in software which worked for my controller but still left them going to the ECU.

The Proportional part looks at the magnitude of the error, the Integral part adds all previous errors and keeps the duty cycle close to optimum while P jumps around. The solenoid duty cycle is computed from scratch at each cam interupt.

The Differential part isn't neccesary because it works best on systems with inertia or stored energy. The VANOS doesn't have much inertia, it can move almost instantly. Plus, Differential is very sensitive to noise and finally, since the set point is almost changing constantly the Differential part never has a chance to see the true change in the error because the set point is always changing which drives error.

The result is a "PI" controller... very common.

My current task is I've added an AND chip that will only activate my circuit output when the ECU VANOS is on. This gives the ECU the "vote" I've wanted to give it.

Next I need to program a map into the EEPROM.

I'm still looking on suggestions on intake/exhaust profiles.

** update **

The AND circuit works, my output switches on when the ECU VANOS switches on. The problem is that the circuit continues to compute duty cycles even when it's output is disconnected so when it doesn't see the cam responding, it continues to ramp-up the signal and the compounding error adds to the Integral term. When the output is connected by the AND chip, the Integral is wound-up and the cam overshoots. Easily fixed, but that's about how this goes. One small step at a time.

[Drew325is]

Sounds like you're all over it Steve!...now, any chance you could write "PID tuning for dummies?" :)

[hoveringuy]

PI controller

More progress on the cam controller...

I re-wrote the PID part to a different algorithm and it's working great, in large part because I tracked down what was causing noise to feed back into the sensor signals. It is also only a "PI" controller. The D part hasn't been necessary.

It is rock-steady now.

If I command 14 degrees of cam advance it stays locked right there at maybe +/- 1/2 degree, even as revs go up and down.

At 1000 rpm it takes about 2 seconds to advance the full 20 degrees. At 2000 rpm it can be anywhere within 1 second.

What I have found is that it really doesn't care about oil pressure or oil viscosity like I originally thought. The reason is that the same oil is pushing on both sides of the diaphram so changes in pressure are felt evently on both sides. Increases in pressure will allow it to respond a little quicker but that's about it.

Now that I will begin mapping advance to throttle position and rpm, I really have no idea of what that map should look like for intake and exhaust. I think for intake I'll have lots of advance at 1500 rpm and taper it down to 0 at 4500, what should the exhaust profile look like?

Anyone have some good experience with cams that can summarize what's going on when intake is advanced and/or exhaust is retarded?

[TwoJ's]

Wow, I don't know how it took me so long to find this thread. Very nice work man! So do you work on Bangor? Down at the shipyard in Bremerton?

I'm also an engineering student (mechanical design) and my electronics knowledge is limited at basic understand of circuits and power systems. Keep it up, and maybe get a video?

[StereoInstaller1]

most electronic stuff is greek to me.

that's for sure...

[CorvallisBMW]

I have a Smorgasboard of experience, I'm drawing on lots of different things I've learned over the years to get this done.

-I learned how to program Z-80 processors in assembly language as a kid. Remember the TRS-80 computers? I'm doing this controller in C++ but it's the same idea.

- My BS is in Systems Engineering which is all about the theoretical side of control systems. I also taught that for 2 years. Feedback, Error, Stability, etc. Sound familiar?

- MS in ITM (Information Technology Management). More computer stuff and LOTS of signal analysis, FFT and such.

- My mechanical background comes from being German. Is that stereotyping?

- I flew helicopters for 15 years. I expect things to work but plan for them to break.

- I have a P.E. in EE which is where I learned a some other useful tricks.

- I have the Google machine!

Put it all together and I'm comfortable doing this. My weakness will be in the theoretical and practical side of exactly when to advance/retard the cams for best performance and tuning.

Currently I have the closed-loop control working marginally well. I've increased the gains to get faster response but that's given me some overshoot and hunting. I'm hoping that more tuning of the parameters will smooth that out. I have a little potentiometer that I can turn back and forth to adjust the cam. It's my "dial-an-advance". :grin:

That is awesome! I'm a Manufacturing & Industrial Engineer, so most electronic stuff is greek to me. Keep up the awesome work, you are an inspiration sir!

[hoveringuy]

OT: Can you perform an autorotation in the types you're rated for? I've often wondered how effective they are at saving lives. I'd guess that different single-engine choppers have much different survival chances (think: JetRanger vs. Hughes 500) and altitude would be a factor.

Yep. In the JetRanger we routinely autorated to a full landing.

The SeaHawk was heavier and the rotor dynamics were different, but we would autorotate it down to around 20 feet. The few people that HAD to do it for an actual emergency rolled the aircraft at the bottom of the touchdown but were Ok.

Sea Story: I had to fly a ship's supply officer into the San Juan, PR. He was being a complete ass at lunch and kept calling us the "ship's taxi". He fell asleep 10 minutes into the flight as all of our passengers did and after an hour at 5,000 ft it was time for our descent into San Juan. Instead of a nice, easy airline descent I briefed the crew and entered an auto. If you've never experienced one it's like having the cable cut on an elevator... an immediate, violent drop. The rotor speeds up to about 120% of normal and the transmission screams. Things floating up off the cabin floor, your stomach in your throat... it's unpleasant if you're expecting it.

It's a rude way to be woken up.

His eyes got as big as saucers. It was beautiful!

We dropped him off and flew back to the ship. Captain thought it was hillarious. The Supply Officer never gave us anymore crap and I never again heard "ship's taxi"

[m3_eater]

- I flew helicopters for 15 years. I expect things to work but plan for them to break.

OT: Can you perform an autorotation in the types you're rated for? I've often wondered how effective they are at saving lives. I'd guess that different single-engine choppers have much different survival chances (think: JetRanger vs. Hughes 500) and altitude would be a factor.

[hoveringuy]

What kind of education/work experience do you have in electrical stuff? You're going nuts on this thing and seem to know exactly what you're doing. Very impressive, I wouldn't touch it with a 20ft pole.

I have a Smorgasboard of experience, I'm drawing on lots of different things I've learned over the years to get this done.

-I learned how to program Z-80 processors in assembly language as a kid. Remember the TRS-80 computers? I'm doing this controller in C++ but it's the same idea.

- My BS is in Systems Engineering which is all about the theoretical side of control systems. I also taught that for 2 years. Feedback, Error, Stability, etc. Sound familiar?

- MS in ITM (Information Technology Management). More computer stuff and LOTS of signal analysis, FFT and such.

- My mechanical background comes from being German. Is that stereotyping?

- I flew helicopters for 15 years. I expect things to work but plan for them to break.

- I have a P.E. in EE which is where I learned a some other useful tricks.

- I have the Google machine!

Put it all together and I'm comfortable doing this. My weakness will be in the theoretical and practical side of exactly when to advance/retard the cams for best performance and tuning.

Currently I have the closed-loop control working marginally well. I've increased the gains to get faster response but that's given me some overshoot and hunting. I'm hoping that more tuning of the parameters will smooth that out. I have a little potentiometer that I can turn back and forth to adjust the cam. It's my "dial-an-advance". :grin:

[CorvallisBMW]

What kind of education/work experience do you have in electrical stuff? You're going nuts on this thing and seem to know exactly what you're doing. Very impressive, I wouldn't touch it with a 20ft pole.

[matt]

Some of the issues I will need to overcome include, How to I give the 413 a "vote" in positioning the cam? I can't completely usurp its will because it switches maps internally. Plus, I've found that standard VANOS isn't even on that often. Not on when the motor is cold, at very low RPM's, high RPM's, low throttle settings or anytime you are decelerating (low throttle setting).

That's because it's smart...

If the oil is too cold, the oil pressure will be too high and you'll blow out your VANOS seals. (that's just conjecture, but still it makes sense that you don't need max power on a cold motor)

All of those factors are adjustable in the 413 software. :)

I'm excited to see the rest of the results of your work on the vanos controller... writing microcontroller code has never been my thing, but from my experience reverse engineering it, it's not easy.