"We've created the next generation of valve technology, and we retrofitted it into an old SAAB"

While I sound sarcastic, its actually REALLY DAMN COOL that it CAN be retrofitted into other/older applications! I bet someone builds one for a Honda K eventually.

 

I wonder, in all seriousness, what the parts cost would be for sixteen of these, plus whatever ancillary supporting hardware is needed. I don't see anything in the NASA TT rules which mandates the use of camshafts.

And we're going to need more outputs on the MS3. Rev, would CANbus be fast enough to run a 16 channel high-current driver board with sufficient angular precision for this application?

 

Angular precision as in 180-1 trigger wheel 24 inches in diameter?

 

No way. CAN Bus has a ~2ms latency.

 

This is really funny that this video comes out now my room mate and I were just talking about this two nights ago. We keep bringing it up and arguing over pneumatic or electronic actuation and who out of the car companies pouring money into are going to do it. I was expecting bmw to come out with the first working prototype to be honest.

 

So how long till you have MS4 done that can run these?

 

Couldn't a lossless system be developed for pneumatic actuation?
F1 cars use compressed nitrogen at 3000 psi instead of mechanical valve springs, for instance, and that is a closed system. So far, I have seen just one incident where they had to pit the car and top up nitrogen from a smallish canister.

Hydraulic systems operate in a closed system to actuate all sorts of mechanical doodads (very technical term there), ranging from excavator rams to hi-speed valves in diesel fuel injectors.

 

From what they say the biggest key is the push system.. that little 1" by 1" cube looking piece. They start producing those and this technology will spread like aids at a club ricer meeting.

Mounts can be made so no new head will be necessary.

 

Is it not viable to just adjust the timing of the events to accommodate for the latency?

That is not meant to sound like baiting, I legitimately dont understand.

What type of setup would be needed to control it

 

Yeah it makes no sense to me how we can control the timing and the injector timing and cant control one of these devices. Sure you'll need 16 outputs (8 since they will be in pairs of 2) however currently we're at 8 so... injectors and (sequential) spark plug signals.

 

my guess is something super 'spensive that has insane resolution and speed.
basically something that would go into a koenigsegg

 

And will pay for itself in maintenance and economy in a year.

Dann

 

The fuel injectors and ignition coils are driven from output pins connected directly to the CPU itself- there is essentially zero latency in the path.

Putting any kind of "external" network bus between the main CPU and a subordinate driver, be it ethernet, CAN, or even I2C, adds latency; some more than others, and frequently non-deterministic (variable in an unpredictable or uncontrollable fashion.)

CAN was never really intended for high-speed, time-critical stuff. I just wasn't sure how bad it was, as I've never actually used it.


Given the present-day MS architecture, you'd need a second MS3-type CPU dedicated to the task. Not a big deal, really. Just parallel the crank sensor to both CPUs, and run a CAN line between them so that the main CPU can pass target data (maps) to the one driving the valves.


It's probable that the delay is not constant. This would make prediction impossible.

 

Wooo no more timing belts!!!

 

Yea, it'll be more like timing suspenders with this stuff.

 

Thats what i was thinking parrallel ecus of some sort or perhaps the 440 adaptronic ecu. I thought that had more than 16 outputs.

As far as the delay. Correct its not constant so there needs to be an angular velocity algorithm of some kind for the other cylinders.

 

How much current can the MS injector drivers handle? I'm pretty sure these are very similar to injectors on the electrical side, well actually they're probably closer to a boost controller. If I had to guess amperage requirements maybe 300 milliamps. But I'm not an electrical guy.

 

That's kind of a non-issue. The VND5N07s on the MS3X board are limited to 5 amps, but there are only eight of them. Therefore, a new board would have to be designed regardless. It's possible that those actuators may require power in both the on and off direction, but that should be achievable by simply putting two FETs on each CPU pin, with an inverter on one.

 

I'm interested in learning the details of those actuators. ECU controlled valves is hardly a new idea; the first conception probably occurred about 10 minutes after the first conception of ECU controlled fuel injection. The problem is that nobody has been able to create an actuator that works, and not for lack of trying.

A quick Google search turned up a little bit of history, though I suspect there's been a lot more effort and investment by most of the big auto makers. I'm sure I've heard rumors of research at GM.

There's this from 2001: AutoSpeed - Camless Engines

This is is kinda cool. EVIC engine

As I said, I'd be interested in finding out what makes the actuators in the video different.

 

The company that is developing the actuator is Cargine, not Koenigsegg. Check out their website.

Cargine | Free Valve Technology