Page 1 of 1

Extreme Boost Control

PostPosted: Tue Aug 27, 2013 12:28 am
by Fred
First a primer on normal turbo operation.

The vast majority of turbo systems regulate boost using a wastegate. A wastegate contains a valve for exhaust flow, diaphragm of some type with area > the area of the valve, and a spring to keep the valve closed. Assuming an efficient manifold, boost pressure is set pretty accurately by the area of the diaphragm and the tension in the spring. In this mode intake pressure from *in front* of the throttle is plumbed to the bottom of the diaphragm, and when boost builds it eventually lifts the valve open. Because the area of the diaphragm is much larger than that of the head of the valve, and our presumption of efficiency (implying ~1:1 exhaust and intake pressure ratio), the pressure in the exhaust manifold has little chance of opening the gate early and lowering boost capability.

So what's wrong then?

Well, when the boost is *nearly* at the level required/desired, the exhaust back pressure starts to play a bigger part, and can cause the wastegate to prematurely leak a bit and bypass still-needed exhaust gasses. What this does is slow the spool down and push the boost RPM threshold up higher. Not ideal.

Another issue is that you must choose a single boost pressure and live with it. Enter boost controllers and related hacks. What the forum kiddies call a MBC (Manual Boost Controller) is actually just a bleed to lessen the effect of intake pressure on the wastegate and thus raise boost. Fancy versions use a ball and spring and screw as a sort of pressure regulator. There's plenty of info here, though, as usual, some of it is shit:

However, with a soft spring you get the exhaust pulses pushing the gate open more and earlier than otherwise. I was thinking about how to remedy this and I came up with this chart:


Basically, you could use a good vacuum source and some clever valving and control algorithms to force the wastegate *OPEN* to *lower* boost below the spring's natural boost level. And/or use vacuum to utterly seal the manifold for ultimate spool with no leak. With a good vacuum source you could expect to get a 12+psi spring to open up the valve completely providing *NO* boost. Nice for a valet mode, too. Or, with an 18psi spring (such as mine) you could get it down to 8psi for first gear, for example! :-)

Assuming you use a tank and the engine (trailing throttle and/or idle) to provide vacuum, the question is, how to vary between vac, atmo, and pressure, without leaking too much air into your vacuum tank and losing vacuum.

I might be over-complicating this, but I thought you could use nested PID loops to control top and bottom pressures independently with multiple valves, and source these pressure ratios from a normal boost control open or closed loop setup.

In the simplest form you could have vac/no vac for the top port on a simple on/off switch based on gear.

Locked until I get some more posts in.EDIT: Unlocked, comment away.


Re: Extreme Boost Control

PostPosted: Tue Aug 27, 2013 1:05 am
by Fred
If I drop 2 of the 5 states, and use just atmo/boost or just vac/boost then a 4 port solenoid with a single PWM input will do the trick. Perhaps I'm over-complicating things?

On the other hand, if we use vac/boost on both sides, where is our steady state? Some bleeding of boost into the vac container? Yuck.

And if we don't use vacuum, and use boost/atmo on both sides, then we can't go lower than our spring and are stuck with a weaker spring than we'd like.

Also, the more extreme we make our control of the wastegate, the touchier the system will be duty wise.

Perhaps the solution is to ONLY use an on/off valve for vacuum to be applied to the top during 1st gear, and use atmo/boost for the remaining gears to control at or above the spring.

On a different note, my plan to use two 3d tables per gear with target boost looked up by TPS and RPM and base duty to achieve that target looked up by TPS and RPM, is exactly what (some? all?) Subaru owners use:


A reasonably priced 4 way valve that might be useful: http://pneumaticautomationcontrols.airi ... /410-12vdc

On top of that they have asymmetrical non-linear PID control for closed loop operation. OEMs usually do things right, so it's nice to see that my plans for FreeEMS align exactly with what Subaru actually did:


Pics pulled from this post:

The guy knows what he's talking about with respect to control and what is needed. Pretty well explained.

Can I make it more sophisticated, though? Or is the juice not worth the squeeze?