View Single Post
Old 04-04-2005, 12:50 AM   #32
Scooby Specialist
Member#: 56468
Join Date: Mar 2004
Chapter/Region: NWIC
Location: Seattle, WA
04 Improved STI
Dirty White


Hydra Boost Control in a Nutshell

Three main players:
1) Boost Target (Select-->Control2-->Boost target)
2) Boost Control Settings (Select-->Settings-->Boost Control Tab)
3) PWM Map 4 (Select-->2D PWM-->PWM Map 4)

The discussion below pertains to a boost control plumbing like the stock (restrictor, T-fitting, solenoid normally closed) or in an 'interruptor' mode (restrictor or not, solenoid IS the T-fitting, 3 way solenoid normally open).

Boost Target
These values at each RPM have the least to do with actual control of any of the boost control parameters. I would go so far to say that the main purpose of the value in each column is to set an "boost cut" value which is 3 psi above the set value. From brief experimentation it seems that the target value may have a very minor effect on the actual boost level, but that would in the +/- 1 psi range.

Boost Control Tab Settings
There are four main values here: RPM onset, Proportional, Integral, and Derivative. Phil's Tuning Guide goes through these values but I'd like to add some additional insight based on altering these values.

The RPM onset value is the RPM at which the boost control system starts to operate. The only thing to consider here is that you begin trying to control boost before boost is present. A lower value in the RPM box will give you a bit of a running start if you are trying to raise boost with a small solenoid. In other words, opening the solenoid at 1500 RPM may permit higher boost than opening it at 2500 RPM.

PID controllers are not really applicable to things like a solenoid that only has two states--on or off. PID control is easier to understand with something like an oven where you can vary the "proportion" of the heat by turning up the gas incrementally.

A boost control solenoid is not an oven. You have two states, on or off. In terms of proportion you can only be on. The same is true for adjustments at the D, derivative level, you are either adjusting on or off, no in between. The integral, I value, is the only one that matters much in boost control You can look at it as bias toward being on or off. If the value is over 1/2 of 255 (127.5) the bias is toward "on" or an open solenoid valve.

Phil has the PID values set at 175, 165 , and 45. You can change these values all you want, but the integral value is the only one that will have much of an effect. Raising the integral will tend to raise boost (somewhat). We are talking at most a couple of psi here, no more. These values, like the boost target value, are really fine tuning boost.

Generically, if you need to "activate" something to raise the boost:
P must be >= 127.5
I can be varied somewhat to raise boost, values between 125 and 200 seem to work best (always less than P)
D only effects how "active" the boost control system is, that is how often it "checks" the boost. Values of 25-50 all work fine, anything more does nothing.

So what is really controlling boost? PWM Map 4

PWM Map 4
PWM stands for pulse width modulation. The basic idea here is that the Hydra is sending a current to open (or close) the boost control solenoid. The faster the pulse width the more the percentage of time the valve is open. Its the same as the fuel injectors.

I experimented with all different shapes of curves in Map 4: linear increasing, linear decreasing, asymptotically approaching a value, parabolic, hyperbolic..... Here is the long and short of it. You want to start in the 5-15 value range on the far left (0%) and you want to end at 100 on the far right (100%). In between, you want to keep values low until you get close to the 100% column, at which point the values rapidly climb to 100. This gives you the fastest boost rise, and holds boost.

If you log data with the Hydra you'll see that the majority of the time under WOT the Hydra spends it's time in the 80-85% range of the PWM map. At low RPM it's at 100% and the solenoid valve is "open/active" all the time. As you get close to YOUR boost target (not the one you set in the Hydra) the solenoid backs off on it's duty. As the boost peak is reached the solenoid duty will drop into the 70s or 60s and quickly bounce back up into the 80s.

If you are like me and have a dinky turbo, my boost tapers from 18 psi to 12.5 psi. I can make the turbo hold a bit more boost by raising the values in the 80-90% columns. Thus, instead of a pure exponential curve, I have an exponential curve with a bump in it.

Here are my values across the board:
16 16.8 18 19.2 21.2 23.2 24.8 26.4 28.8 31.6 35.2 39.2 56.4 66 72.8 100

For the VF39 and GM solenoid that gives you a peak boost of 18 psi tapering to around 13.5. Use a different turbo, a different solenoid, and your curve will be different to achieve the same effect.

The two values that will most effect the "boost peak" are in the 73% and 80% columns. Raising these values will have a linear effect on the boost peak if the shape of your curve is like mine. Determining which of the columns will raise or lower the peak is just a matter of logging the data to see where the "boost valve %" is just prior at at the boost peak. Lowering the values at the logged percentage will lower the boost peak.

The one thing I feel pretty sure will always remain no matter what components are used, is the general shape of the curve--like an aircraft take off.

Summary So the summary is: PWM Map 4 is you main boost control point, the other systems/control points are pretty much fine tuning. The "Boost Target" is really there to set a "Boost Cut" which occurs 3 psi above the Target value.

If you are using something like a stepper motor to control boost, the PID values will become more important. Remember the Hydra was not built for the Subaru, but as a generic control system. For our cars, and the majority of boost control systems that Subaru users use, PWM Map 4 is the money for controlling boost and I believe the above advice will hold true for those single solenoid based systems.
* Registered users of the site do not see these ads.
bboy is offline   Reply With Quote