JasonC SBB

Breaking off from another thread...

LOL I've been talking about PTFB and the solution since about 2001. This is why I use a 5 psi wastegate can and then use EBC to raise it to 10.

One of the nice things about the AEM is you can use a 3D MAP of target boost, with TPS and RPM as the axes.

JasonC SBB

On a slightly different topic

Larger can diameter = it holds the wastegate swing valve shut with more force and thus less likely to blow open when it doesn't want to open. This is because force = area * pressure, and area is proportional to diameter squared. This is more important with low psi cans. I like using a 5 psi can because I can use EBC to make MAP proportional to TPS. With a 5 psi can I can start controlling MAP starting at 5 psi.

Inside the can, a longer, lower spring k (softer but scrunched in more), means higher loop gain in control system parlance. This is apparent in the ratio additional psi per mm of actuator motion if you test it with a bicycle pump. More motion per change in psi is more gain. i.e. a 5 psi can may start cracking open at 5 psi, and move 5 mm at 6 psi. This is 5 mm per psi. An actuator that moves 2 mm per psi has less gain and boost will be less flat.

Thus it will hold the boost curve flatter.

y8s

iTrader: (8)

Jason, the adaptronic has a tickbox for "multiply base DC by %TPS" on the wastegate tab.

haven't tried it yet but am curious.

JasonC SBB

I looked at it and it's not ideal but a good idea nontheless.

Ideally duty cycle only starts to open up at some part throttle value such as 50-75% throttle. So between 0 and 50% or 75% throttle, you get mechanical boost in the intercooler piping (pre throttle).

y8s

iTrader: (8)

would it make more sense if "boost target" was multiplied by %TPS over a given range?

so if your normal boost target is 15 psi above 3000 rpm

then from 0%TPS to 100%TPS it would range from 7psi (WG can) to 15 psi?

or something like that?

JasonC SBB

If you are thinking about closed loop boost control, the system should only go closed loop once you are targeting boost which is greater than mech. boost.

BTW see attached actual data of typical curves of TPS vs. MAP for an NA motor, at different RPMs. A turbo motor will have different shaped curves, but this is for illustration.

If you want EBC to feel "natural", you should only raise boost at TPS values above the point where the curve starts to go flat.

This is why the ideal boost control system will have a 3D curve of target vs. TPS and RPM, then a 3D table of duty cycle vs. boost target and RPM. Then closed loop from there.

MeOughtta

Imagine a throttle that directly controlled torque.

How close can we get to that?

Research: Measure torque vs. RPM vs.. MAP

1) For a given throttle position, desired torque is % of max torque, i.e. 50% throttle seeks 50% of max torque.

2) Starting with desired torque and current RPM, use data collected in (1) to look up desired MAP

3) implement PID wastegate control which utilizes difference between actual MAP and calculated MAP as its error signal.

Possibly necessary to get stable control at partial throttle:

Research: Measure pressure drop across the throttle as a function of position and flow rate.

Requires: pressure sensor before throttle plate.

4) Using MAP and RPM, calculate flow at desired torque

5) Using flowrate and throttle position, look up pressure drop (deltaP) across throttle plate. Add that to MAP value determined in (2)

6) Use delta between pressure calculated in (5) and pre-throttle pressure sensor as error signal to PID controller.

Not that hard to do using a single-board device control computer.

Coming next: How we could use the new Garret turbo RPM sensor and a servo-driven blow-off valve to maximize responsiveness and minimize turbo lag.....

M.

JasonC SBB

Some OEMs with throttle by wire do what you say (TPS is torque command), such as with turbo Audis.

With a mechanical throttle, there will always be a region where you get way more % of full torque than % throttle opening. It's worse at lower RPM.