Boost Control on a Subaru WRX and STI

What Boost Control Limits

The job is to hold the turbo back. The FAQ’s picture of a turbo with nothing limiting it is one that keeps climbing, on the order of 50 psi, until it fails. Every device in this discussion — the factory solenoid, a manual controller, an electronic controller, a stiffer actuator, an external gate — is a way of opening the wastegate before that happens.

On a stock turbo car the two pieces that do this are the boost control solenoid on the passenger-side strut tower and the internal wastegate on the turbo. The tune is what tells the solenoid when to act. The person who has to support the car is the person who picks the hardware. That list, in this thread, includes a manual or electronic controller, an upgraded wastegate, a restrictor pill, a helper spring, an external wastegate’s type and size, and the plugs and injectors that have to match the new airflow.

A change of about 1 psi at an adjustable rod is the only hardware step the FAQ treats as something you might get away with before a tune, and even that is a maybe. Anything above it belongs on the “yes” rows. Most cars in this thread needed none of the hardware. The tune already controls boost through the wastegate and the solenoid that are on the car.

Stock Wastegate and Rod

The internal wastegate is a spring and a diaphragm. The valve starts closed. Pressure on the nipple of the actuator canister pushes the shaft, the flapper swings open, and exhaust bypasses the turbine. Less flow through the turbine wheel means a slower turbo and less boost.

Whether you can set that spring from the rod depends on who built the turbo. IHI actuators in this FAQ have a fixed rod. MHI actuators have a rod you can shorten or lengthen.

• Shorter rod. More preload on the spring, so the gate opens later and boost can rise, as long as the turbo is still in its efficiency range. The valve also cannot open as far. Too much preload becomes mechanical boost creep that a tune cannot remove.
• Longer rod. Less preload, so the gate opens sooner and peak boost falls.
• Not enough preload. Boost can wander about 2 psi either way even while wastegate duty stays constant.

Duty Cycle

The factory solenoid is a bleed. The ECU pulses it. While the turbo is spooling, that solenoid is open and vents pressure so the wastegate sees less of the signal and stays shut. A higher wastegate duty cycle vents more, which is higher boost or faster spool. As the target gets close, duty should come down a little so the gate can hold the number instead of sailing past it. The opening post said the solenoid was idle during spool and then ramped up to limit boost. Later replies corrected that for this bleed-style solenoid: duty down means the solenoid vents less.

An aftermarket solenoid in this FAQ — a Perrin or Prodrive unit built for these cars, or a GM solenoid retrofitted to them — does the same job with different port sizes and a different strategy. The factory unit is bleed-style and narrow-range. The aftermarket units are intercept-style, and intercept control can respond faster than the same solenoid run as a bleed. That difference is why a stock map, or an off-the-shelf map written for the factory solenoid, is the wrong map. Tiny changes in the length or diameter of the hoses on the turbo and the wastegate move the result enough that the FAQ says not to run a GM, Perrin, or Prodrive solenoid until the car is being tuned. About 95 percent of whether the swap works is the tuner, not the brand on the solenoid.

A 3-port solenoid and the restrictor pill were left unresolved. Owners were told a 3-port can run without the pill. One Prodrive setup in the thread still used a 0.025 inch pill, and the reply was that it was probably plumbed as a bleed rather than an interrupt. Do not pull the pill because the solenoid has three ports. Leave that call to the tune.

Restrictor Pill

The pill is a small brass restriction in the pressure line off the compressor housing, before the tee. It keeps the solenoid and the actuator from being overdriven, which would crack the wastegate open early. On the factory narrow-range solenoid the pill is not optional. Without it you cannot bleed enough signal to add more than a few psi, and the gate opens earlier and farther than it should. That is a performance loss, not the failure mode the FAQ treats as the dangerous one, but the car will not make its boost.

Each of those holes is plus or minus about 0.001 inch. A smaller hole spikes more easily and needs less wastegate duty to reach a higher target. A larger hole is less likely to spike and needs more duty to make the same boost. The weak-spring row is the one to be careful with: it is the spike-prone size, and it takes less duty to overshoot. After a turbo swap, confirm the line still has a pill and that the hole matches the new turbo and the spring. The FAQ’s ways to get another size are a drilled OEM pill, a nitrous jet filed to fit the hose, or a set the tuner already keeps. Swapping the pill is a tune, not a driveway boost turn-up.

A missing pill shows up as low boost, and it shows up on more than one car. One 2002 was around 8.5 psi until a dealer pill went back in the line that someone had replaced, and then it made full boost again. One 2003 wagon was at 0.05 MPa, about 7 psi, without a pill, and reached 0.1 MPa after a dealer pill, then hunted between about 0.06 and 0.1 MPa through the pull. That hunt was never diagnosed here. Low boost after a hose or turbo job is a reason to find the pill, not a reason to add a controller.

Springs and Creep

A helper spring is one small spring from the wastegate flapper arm to the wastegate. It adds tension so the actuator has to work harder to open the door. It is cheap and removable. It is also the mod this thread watched people stack until motors came apart.

The damage cases were not one spring on a tired actuator. They were two, three, four, or five helper springs on a VF39 in an STI, or on a VF30 on a 2.0, holding the door shut while exhaust pressure was already high. Helping a weak spring and forcing a flapper closed are different jobs. The FAQ’s own line is that more than one spring is not a repair. If the car is only making about 7 psi, and the number the FAQ uses for “back to normal” in that example is 14.7 psi, a spring is the wrong tool. Find why the gate is opening early.

External Wastegates

An external wastegate shows up for two reasons. It can bypass more flow when boost is high, and it is required when the turbo has no internal gate. On a turbo that already has an internal gate, moving the bypass outside the housing reduces the choking that a small internal passage creates, and the gain is largest on a bigger turbo. It is fabrication: an uppipe with a flange, a gate sized for that turbo, and a plan for where the bypass gas goes.

Where the bypass gas goes

Vented to atmosphere, the gate dumps outside the exhaust. The pipe has to be aimed so it does not heat the bay, and it will be louder. The FAQ’s noise option, if there is room, is a small motorcycle or two-stroke muffler on the end of that vent. The quieter path welds the dump back into the downpipe. That joint wants to be at least 18 inches past the downpipe-to-turbo flange, past the catalysts if the car still has them, and at a shallow angle so the flow joins the exhaust instead of a 90-degree junction that pushes pressure back upstream. Even a careful plumb-back is less efficient than a vent. Street-driven cars keep their catalysts. This is a routing note, not a reason to remove one.

Size and spring

The two sizes this thread actually uses are 38 mm and 44 mm. The sizing theory runs opposite of the usual shopping instinct. A wastegate limits boost, so a lower target needs more bypass and points at the 44 mm, and a higher target needs less bypass and points at the 38 mm. The turbo vendor and the tuner can override that theory, and the FAQ says to ask them. The 44 mm is also easier to service in a tight space because of the V-band. That is a real convenience. It is a bad reason to buy one when a 38 mm is leaking at the gasket. In the FAQ’s 2008 example, replacing that gasket was a few dollars, and stepping up to the 44 mm was on the order of $200 more. Those are that post’s prices, not a current quote.

Spring rate follows the same “ask, then check the theory” rule. The theory is a spring near half the boost you want to hold. A 15 psi target wants a spring as close to half of that as you can get. External gates are meant to be opened and resprung by the owner, using that gate’s own instructions. Most internal actuators are not. The FAQ treats them as a whole-unit swap and names Deadbolt and AVO among the sellers of upgraded internal assemblies. TiAL was the consensus name when the thread argued brands. The decision that matters is size, spring, and where the dump is plumbed.

When You Need a Tune

Tuning the boost system is the slow part of a calibration. Showing up with a solenoid or an actuator the tuner did not ask for spends that time twice. The FAQ’s 2008 picture of the waste was an actuator around $175 plus an extra dyno hour around $300, against a restrictor pill the same tuner might swap in a few minutes for a couple of dollars. Treat those figures as the example in that post. The point is to spend the hardware on the fix the log actually needs.

A manual controller is open-loop. It does not hold one boost number across the gearbox. Set it in third and fourth and fifth run higher. Set it for fifth and the lower gears fall short. A wide-ratio transmission, a 4EAT or a 6-speed in this thread, makes that spread worse. Closed-loop electronic control is what the thread calls the consistent answer, and it still has to be tuned. “Manual” means the driver is part of the system.

One 2003 wagon on a Hallman Pro RX and the stock turbo stayed tighter than that warning. Peak boost was about 1 psi higher in fifth than in second, under 0.5 psi apart between third and fourth, and about 0.5 psi higher from 5,000 feet down to sea level. The reply was that most manual controllers are less steady than this one car, and that a 2002–2003 ECU does not react to the same hardware the way a later car does. The FAQ allows a manual or electronic controller as a standalone only on a 2002–2003 WRX or an STI, and only with that knowledge. On other cars it belongs next to an engine-management system that was tuned with it installed. Ask before it goes on. If a two-port manual controller replaces the factory solenoid, cap the intake port the solenoid used to feed. Leaving it open is an unmetered leak in the signal path.

The thread also never states a factory boost target for the 2.0 or the 2.5, and it never answers how much a stock turbo “can take” as a spec. One owner drew a personal line at 17.5 psi on a stock WRX turbo, and said 19 psi at 5,000 feet made more power and wore that turbo faster. That line is one car at one elevation.

A 2015 WRX question shows up at the end of the thread and never gets a method. This packet’s solenoids, pill holes, and rod rules are the EJ cars with an internal gate and a strut-tower solenoid. A direct-injected FA still needs its own tune. Do not copy this hardware onto it.

Order of work

1. Log the boost you have. A number near 7 or 8 psi after a hose or turbo job is a missing pill or an early gate until the log says otherwise.
2. Pressure-check the charge system. That is the one boost step the FAQ does not tie to a tune.
3. Confirm the pill. It belongs in the compressor line before the tee, and the hole has to match the turbo and the spring. A swap is a tune.
4. Ask the tuner before new hardware. Solenoid, controller, actuator, external gate, and external spring are all tune items. Hose length counts as hardware.
5. Touch an adjustable rod only in small steps. IHI rods do not adjust. On an MHI rod, about 1 psi is the FAQ’s outer edge for skipping a retune. Past that, schedule the tune. Too short and the creep cannot be tuned out.
6. Use one helper spring, or none. It helps a weak actuator. It does not replace a diagnosis, and a stack of them is how VF39 and VF30 cars in this thread broke.

FAQ

Do I need an aftermarket boost controller?

Usually no. The tune controls boost with the wastegate and the factory solenoid. Add a solenoid or a controller when the tuner says those pieces cannot hold the target, and do not drive the new part on the old map.

Does a boost controller raise boost?

It limits boost. The target moves when the tune, the spring, or the duty cycle opens the gate later. With nothing limiting it, the FAQ’s picture is a turbo that keeps climbing until it fails.

Can I adjust the wastegate rod?

On an MHI turbo, if the rod is made to adjust. IHI rods in this FAQ are fixed. Shortening raises the opening pressure. About 1 psi is the only step treated as possibly safe before a retune, and too much shortening creeps in a way duty cycle cannot fix.

What size restrictor pill should I run?

Stock center hole is about 0.040 to 0.048 inch. A larger turbo or a stronger spring is about 0.050 to 0.055 inch. A weaker spring on a stock-size turbo is about 0.030 to 0.040 inch, and that smaller hole spikes more easily. All of those are plus or minus about 0.001 inch, and changing the pill needs a tune.

Should I buy a 38 mm or a 44 mm external wastegate?

The theory says 44 mm when the target is lower and you need to bypass more, and 38 mm when the target is higher and you need to bypass less. Ask the turbo vendor and the tuner. A leaking 38 mm is often the gasket, not a reason to jump sizes.

Is a manual boost controller steady from gear to gear?

An open-loop manual controller moves with load and gear. One 2003 wagon stayed within about 1 psi. The thread’s warning is that most do not, especially on a 4EAT or a 6-speed. Closed-loop electronic control is the consistent option, and it still needs a tune.

The car only makes about 7 psi. Can I add a spring?

No. The FAQ uses 14.7 psi as the “normal” number in that example and says a spring is the wrong fix. Look for a missing pill, a leak, or a wastegate that is opening early.