hraday93

The biggest speed bump I see here, is how to attack the tuning process itself. I personally don't understand much about how the VVT is implemented into our platform, but take something like V-TEC which is hydraulically triggered vs something even more advance like what runs in the newest Koeniggsegg engines which is all electronically controlled, and has a very flat curve of open and closed valve timing.

The second option is more precise, but it's incredibly expensive to implement. V-TEC is on the other end of the spectrum, but wouldn't it create mountains of issues for the tuner that wants to optimize that open-close cycle for the most efficient power delivery? I wouldn't even know where to look in TunerStudio to have that kind of control on our platform, as I imagine it's not electronic anyways, but closer to the Honda's implementation.

With a cam rotating 25 times a second, I imagine that would become a tuning nightmare trying to get the process down correctly. But I guess they manage to do it anyways, with things like Hondata etc. But are those tuners relying on the benefit of pressure to create that VVT action, rather than the tune optimizing its potential?

AlwaysBroken

Tuning would not actually be that complicated, I think. You select a cam profile, do some dyno runs. Select a different profile, do some more dyno runs. And you repeat this until you've mapped out where each possible cam profile is strong or weak in terms of RPM, manifold pressure, etc. Then you construct a map to apply different cam profiles at different manifold pressures, rpms, etc. You might end up adding in corrections for ACCEL or other transient engine conditions if they benefit from different overlaps or whatever. Dunno.

If you can modify the firmware yourself, it shouldn't be that hard (in theory) to write code that constantly recalculates the target cam profile for the current engine conditions. I'm assuming the ECU you're using has spare CPU cycles, memory, etc to add features. I have no idea how much support there is for this off the shelf right now.

AlwaysBroken

Two points:
1) 25 times a second isn't a huge burden for an ECU. Mathematical computations take nanoseconds and 25 times a second is like 40ms, which is an eternity unless you're doing something like talking to stuff over a network connection, doing string manipulation or searching large data structures.
2) You aren't going to be drastically changing settings 25 times each second. Your RPM isn't going to go from 4000 rpm to 8500 rpm in 40ms, it will take a few seconds at least, plenty of time to adjust settings. Certainly if something like a 20-30 year old Link or Haltech ECU can keep up with spark timing and injector pulsewidth in this sort of timeframe, controlling valve timing is also possible with a modern ECU.

z31maniac

Not sure if trolling, or willfully ignorant.

VTEC IS NOT VARIABLE VALVE TIMING!!! All VTEC does is switch to a higher lift/longer duration cam profile in the upper RPM range.

Modern VVT can typically advance the cams +/- 40° from base depending on load/rpm.

hraday93

I guess I didn't fully grasp the scope of what some of these processors can do. My main question after reading that is purely from wondering if those separate instances (ones where you mention the different cam profiles) could all become one tune? Or would these ECUs (Megasquirt as a singular example) be actively remapping the car?

I've dabbled with Autotune a bit, and it seems as if the Megasquirt can modify settings based on data it receives but I've still have had zero luck getting the car to start on MS, so I keep putting miles on the engine with the stock ECU as it seems to idle, drive, and go through the entire RPM range without issue even though my setup has changed drastically compared to a stock Mazdaspeed.

anyways, I digress...are things being remapped? Or is it all just widespread data being concisely grouped into a singular map? VVT seemed much more simple before my brain started to fire on the intricacies of tuning. Perhaps my understanding of tuning in of itself could use a better foundation

AlwaysBroken

I'm not really trolling, just attempting to be funny. I know how vtec works (I was a honda guy in the 90s) and I know how modern variable valve timing works as well. Which is why I run 20 pounds of boost on a non VVT head.

My point was that if you're talking about the benefits of approaches taken by non BP engines, you're talking about fabricating motor mounts and transmission adapters to take advantage of those approaches. Once you've admitted that you're willing to do that, the next question is automatically "why not swap in a small block?"

AlwaysBroken

No no I'm not saying that megasquirt or whatever can handle whatever it is that I'm talking about. I don't run a VVT head.

What I'm saying is that if you can actually edit the source code of the firmware for your ecu (wasn't MS originally an open source project? or are the MS PNP things all proprietary?) you could rig together something to control a variable valve timing system of your choice. I dunno how well supported these things are by any given ECU.

The way I understand VVT is that there is a sensor that tells you how much advance/retard you're running but the PWM signal doesn't directly control advance/retard- adjusting the duty cycle advances or retards the cam. So basically one bit of code would set the advance/retard target based on engine conditions and then the computer should do a closed loop adjustment of the PWM signal to make it chase after the target. It should actually be simpler than implementing a boost controller, I would think.

z31maniac

Yep, modern ECUs are incredibly powerful these days. There is no reason to go an aftermarket ECU unless you're building some monster race car.



I don't think he's at the "OMG what crazy motor can I swap in here" more about trying to understand what the advantages are in general. That seemed pretty obvious from his OP.

aidandj

iTrader: (5)

Megasquirt supports up to 4 VVT cams. There are lots of options out there. Not sure how this discussion turned into "how to adjust VVT" because thats already been done. So you are arguing about nothing.

Duty cycle controls the advance/retard of the cam, but its not just set it and forget it. you need a control loop, because there are different variables that can affect the cam angle, other than just duty cycle. Oil viscosity, rpm, solenoid, voltage, etc etc.

18psi

iTrader: (76)

Emilio briefly started talking about it when we were discussing ND tuning in I believe the ND thread here. Since it seems to be so difficult to discuss this without mentioning a specific engine, let's apply this to the ND dual VVT 1.5L.

I guess I"ll start searching for ND tuning discussions (if those even exist, since craputec and others seem to want to completely cut the self-tuners out and support only pro-tuners who charge an arm and a leg)

Art

.

hraday93

FWIW I wasn't trying to discuss "adjustment", I'm still very much on the tuning subject because I truly don't understand the scope of what is, and isn't, tunable. I guess from a fundamental standpoint, I'm trying to figure out what the course of action is. For me running a non-VVT head, in my time on MT I've at least grasped that I should be learning about spark, injector duty cycles, AFRs, etc... then VVT is adding another variable. What are you accounting for? Are you accounting for more at all? Is the process or line of thinking inherently different from the jump? etc.