View Single Post
#7: 12-02-2010, 07:56 AM
 m sprank
m sprank's Avatar
Title: Vendor III
Location: Escondido, Ca
Car: Stg3 08 5EAT + Stg3.5 08 SpecB
Posts: 29,464
iTrader: (198)
Send a private message to m sprank Find all posts by m sprank Reply With Quote
Learn stuff - Diverter valve leaks, how relevant to engine performance?
The following is an extract from "Diverter valve leaks & the effects on
engine performance and TMS solutions" discussion paper put together by our engineers.

It provides the results summary for tests conducted by our engineers to clarify this contentious issue and dispel the myths. A full copy of the paper is available as an attachment.

Test 1
Pressure_______________ 10 psi / 15 psi / 20 psi / 25 psi / 30 psi
Tight tolerance_________ 136 sec / 119 sec / 97 sec / 71 sec / 58 sec
Loose tolerance_________ 93 sec / 84 sec / 60 sec / 37 sec / 35 sec

A further two tests were performed at 30psi, both with pressure fed only to the top of the valve.

Test 2
Pressure___________ 30psi
Tight tolerance_____ 58 sec
Loose tolerance_____ 35 sec

The purpose of test 2 is to demonstrate that the leak is confined only to the amount of air that can leak through the piston-to-bore clearance, and that no leak occurs on the piston seat side. The importance of this will become apparent in the discussion to follow.


To work out the size of the leak in CFM, a few simple conversion formulae are applied:

Displacement volume = 500mL = 500cc

1 cubic foot = 28,316cc

Therefore displacement volume = 500 / 28,316 = 0.0176 cubic feet (CF)

To convert the test results to CFM:

CFM = (displacement volume (CF) / time (s)) x 60

So the leak results table from test 1 converted to CFM becomes:

Test 1
Pressure___________ 10 psi / 15 psi / 20 psi / 25 psi / 30 psi
Tight tolerance____ 0.0077 / 0.0089 / 0.0108 / 0.0149 / 0.0182
Loose tolerance___ 0.0114 / 0.0126 / 0.0176 / 0.0286 / 0.0302


If we take an example of a 2.5L engine at 7000RPM and 10psi boost, using an engine airflow calculator the maximum airflow is approximately 386 CFM.

With an airflow of 386 CFM into the engine, the GFB valve with the loose tolerance on leaks 0.00295% of the total airflow.

If the same engine were modified and running high boost, at 7000RPM and 25psi it will be flowing approximately 621 CFM, of which the loosest GFB valve would be leaking 0.00461% of the total airflow.

Other boost leak sources

As an interesting comparison, there is a device fitted to almost every turbo car that produces a leak far larger than the results shown above, but is often overlooked.

The boost control solenoid, or in fact any boost control device, must leak air in order to function. This leak is almost never noticed however, as a factory boost control solenoid is a closed system that won’t show a leak during a smoke test, nor will it even operate until the engine is running and on boost.

A common boost control solenoid valve was connected to the test apparatus and driven at 50% duty cycle. The actual duty cycle depends entirely on the turbo system and boost level, but 50% is a common figure for a mild boost increase on a stock car.

Test 3
Pressure______ 10 psi / 15 psi / 20 psi / 25 psi / 30 psi
Boost control__ 0.0960 / 0.151 / 0.176 / 0.211 / 0.264

From these results we can see that at 50% duty cycle the boost control solenoid is leaking up to 17 times more air than the GFB TMS valve.


From the results above, it can be seen that the expected leak from a GFB TMS valve represents such a tiny percentage of an engine’s total airflow that it would be impossible to for any engine performance measuring equipment (such as a dyno, boost or air/fuel ratio gauge) to detect the slightest change in boost level, torque or air/fuel ratio.

When boost leaks are suspected and tested for on a car, this is usually because a loss of boost pressure or performance has been noted. Smoke testing or pressure testing the intake tract will often discover a leak from the BOV, but the size of the leak is not taken into consideration by such testing. More can be read on this topic in the "Turbo lag" discussion paper available from GFB and reference sites

What if…?

As surely as the results of this test will answer many questions and put things into perspective, it will also undoubtedly raise a few more questions such as “What if my valve is old and worn”, or “what if my valve is leaking more than the tests results above”?

The results from test 2 show that the leak measured is all coming from the top chamber of the GFB TMS valve. The air source to the top chamber is via a nipple connected to the intake manifold, usually using a length of 3/16” or similar vacuum hose.

It makes sense than, that the largest leak possible from the top chamber of a TMS valve is limited by the vacuum hose – i.e. the vacuum hose will only flow up to a certain amount at a given pressure, even if the leak at the valve were larger. Even though a leak this large would render the GFB valve ineffective and would be immediately noticeable, it’s worth performing the test for perspective.

The pressure regulator was connected directly to the test apparatus using a 400mm length of 3/16” vacuum hose.

The vacuum hose displaced the 500cc of water in an average of 0.21 seconds at 10psi, which was difficult to measure accurately but is still indicative nonetheless. Since the time was so rapid and the test equipment not big enough to test a larger volume of water, higher boost pressures were not tested.

So a 400mm length of 3/16” vacuum hose at 10psi leaks at a rate of approximately 5 CFM. Again, using the example of the 2L engine at 10psi, this represents 1.29% of maximum airflow.

So although this amount of air loss is a grossly exaggerated example and would not actually occur, it is still not large enough to be noticeable when driving, and would be difficult if not impossible to measure at the engine.

Further testing

As stated initially, any amount of air escaping from a TMS valve can be correctly referred to as a leak. However, from the results above it can be safely concluded that even the theoretical maximum leak at the top of a GFB TMS valve will not cause a drop in boost pressure or engine power large enough to be noticeable, let alone reliably measurable.

So the top of the valve can be eliminated as a potential source of power loss. The only way it can leak enough air to cause a power loss is if the piston were to actually open under boost.

The test procedure used in test 1 simulates WOT conditions up to 30psi, and test 2 proves that no leak from the bottom of the valve occurs under these conditions.

It is often thought that BOVs will eventually begin to open given enough pressure. This is true in many factory fitted valves as they are actually designed to do this to limit boost pressure. However, a GFB valve’s design means that the opposing pressures on the piston cancel each other out, meaning that no amount of pressure at WOT will open the valve.

In fact, because the diameter of the Mach 1’s piston seat is slightly smaller than the outside diameter of the piston, there is a slightly larger area on top for the pressure to push down than there is on the bottom pushing up. Therefore the spring pre-load is irrelevant – it is not possible for the GFB valve to open under boost.

As final proof of this, test 4 was a repeat of test 1 at 30psi with the spring removed entirely from the Mach 1 valve. The piston did not open and the results were again exactly the same.


So technically, GFB TMS valves can be said to “leak”. Importantly however, when put into context the following has been demonstrated:

• on a 2L engine at 10psi the leak would represent 0.00295% of the maximum engine airflow

• the largest expected leak at 30psi is 0.0302 CFM

• all boost control devices also leak air, and at an average setting of 50% duty cycle a boost control solenoid leaks up to 17 times more air than a GFB TMS valve

• The amount of air lost to these leaks is so small that it would not be possible to detect or measure a performance loss as a result

• It is a simple matter to prevent the leak altogether with the use of an O-ring, but we choose not to (even though it would be cheaper) for valve performance reasons

To summarise, a GFB TMS valve shown to leak as per the tests above indicate that it is not large enough to cause a performance drop or even be measurable at the engine.

The only way a GFB valve can cause a loss of air large enough to be felt as a drop in performance, or measured as a lower boost reading, power drop or AFR change would be for the valve to open under boost.

However, as all of the tests demonstrate (in particular test 4), the GFB valve remains shut regardless of the boost pressure or the spring pre-load, so the possibility of a GFB TMS opening under boost pressure can also be safely eliminated as a source of power loss.