Here's an answer from the Go fast Bits engineers
The spring pre-load of all atmo or recirc venting bypass valves is determined by a number of factors, mostly based upon the design of the valve. For the design of the GFB valve, the spring pre-load is designed to hold the piston shut at idle – only just. The adjustment range of the spring pre-load should be such that the amount of force holding the piston shut at idle ranges from almost none to about 1.5lb. To accurately test this, don’t bother with vacuum pumps and the like, connect a vacuum hose from the manifold of your idling engine to the top of the GFB valve and then feel the force holding the piston shut – it will be almost zero. Therefore, when the throttle is closed after having been on boost, it only takes a very small amount of boost pressure to push the piston open and vent. This will be obvious during use, because even revving the car in neutral should cause the piston to partially open and vent audibly.
It follows from the above that the spring pre-load is not related to boost, but rather closed-throttle vacuum. The adjustment range simply changes how easily the boost pressure in the pipes can push the piston open when the throttle is snapped shut. Note that at the other end of the scale, i.e. when you’re at wide open throttle and full boost, the design of the GFB valve is such that the piston will stay shut regardless of the spring pre-load adjustment setting. This is because the area of the piston top and bottom is equal, and both the top and bottom are exposed to the same boost pressure at WOT, therefore the forces on the piston cancel each other out, leaving even the smallest amount of spring pre-load the simple task of keeping the piston shut – in the GFB valve the lowest amount of spring pre-load is about 10lbs, so it doesn’t matter if you are running 5psi or 50psi, there will always be at least 10lbs of force holding a GFB valve shut at WOT regardless of the boost or spring pre-load adjustment.
A note to anyone who sees boost dropping off as RPM increases, this does not necessarily mean it’s the valve leaking it off – most commonly this occurs because the turbo simply runs out of puff – a stock EVO turbo is really being pushed above 20psi, and as such the exhaust backpressure rises exponentially, the compressor efficiency drops right off and the air becomes excessively hot, to the point of choke – i.e. the region beyond which increasing the turbo’s RPM yields very small increases in airflow.
Going back to the first point about spring pre-load, why is the GFB valve designed to be shut at idle? This is because any BOV that vents any amount of air to atmosphere on a car with a MAF sensor needs to be shut at idle, or the car will stall, backfire, stumble and generally drive poorly – try venting a stock valve to atmosphere and see what happens…
So from this we can establish that if you have an atmo-venting valve, there is a minimum amount of spring pre-load that you can run – if the spring pre-load is too soft the valve will open at idle and cause the aforementioned problems. Now, in regards to flutter caused by the BOV, there is a maximum amount of spring pre-load above which flutter will occur, most noticeably beginning with low RPM and light-throttle lift offs, becoming more and more noticeable at higher spring pre-loads, until it begins occurring during higher boost and RPM lift-offs. Note that it is not uncommon for the minimum spring pre-load and the maximum spring pre-load as defined above to overlap, meaning and ideal solution is not possible – make it too soft to cure flutter and bucking at part throttle, and you may find the car idles poorly. Make it firm enough for smooth idling and you may find it flutters a little during light throttle lift-offs. This now leads into a whole new discussion about fluttering and its effect on the turbo.
In my opinion, if you are running a stock-cored turbo higher than factory boost, then you are placing quite a bit of load on the turbo. How much? Running some rough numbers off a TD05H compressor map, going from 18psi @ 5000RPM to 22psi @ 5000RPM increases the turbo RPM by about 20,000. By the time you get to redline, this number gets a lot worse. Then consider how much time you spend on boost compared to the next guy – there’s one of the biggest variables right there – two cars running the same boost and with the same miles on them may have seen the turbos do vastly different amounts of work simply be differing driving styles. Now throw flutter into the mix – the loads on the turbo created by a flutter when lifting off from 3000RPM at light load pale into insignificance. Compressor surge as defined as occurring when the turbo compressor is incorrectly matched to the engine and is being driven by the exhaust whilst it is surging, that’s an entirely different matter – that type of surge CAN do a lot of damage to a turbo – a small flutter at low boost and RPM in my opinion won’t.