There have been many discussions about how much the stock M275 air filter housing might be restricting performance.
Therefore I decided to build a less restrictive system and test it against the stock system in a repeatable setting with basic measuring equipment. While at it, I also wanted to test with direct plumbing and no air filters at all.
Here is my setup:
2004 SL600 with a 2006 AMG 6.0 M275 with a performance tune and no CATS.
Using the Torque Pro app which is built into my dash head unit and set to display Intake Manifold pressure, Intake temperature and Torque. It registers the max of each for easily comparisons.
Outside temp for all tests was around 40F, so pretty cool. Intake temp when I go WOT was typically around 90F, so barely warm.
Driving on an nearby highway entrance ramp. Traction control is off, shifter in 3. I go WOT at about 30 MPH and accelerate to highway+ speed.
The Torque Pro app shows much lower torque values than the factory claim of 725 lb/ft for a stock engine. First this is measuring wheel torque not engine torque. Still seems low. Not sure why, maybe because it uses GPS, partly because the entrance ramp is a bit up hill. It does not matter because we are interested in differences, not absolute values.
First is a stock air filter test, then a simpler air filter test, then a test with directly plumbing and no air filter. The latter was followed within a few minutes with another test with the stock air filters. Air filter MAP sensors were always installed.

Stock air filter setup:



Performance run a few days ago showed a max torque of 553.4 lb/ft and a max intake pressure of 27.7 PSI, subtracting ambient 14.0 pressure gives 13.7 PSI of boost.



My custom air filter setup consists of these parts. Tubing and filters are 2.5" with a reducer to the 2.25" at the turbo inlets.



On the engine, it looks like this:



Performance run a few days ago showed a max torque of 583.3 lb/ft and a max intake pressure of 30.2 PSI, subtracting ambient 14.0 pressure gives 16.2 PSI of boost.



Directly feeding air to the turbos looks like this. I don't see how this could be more direct. Again the tubing is 2.5" with a reducer to the 2.25" at the turbo inlets.



Performance today showed a max torque of 621.6 lb/ft and a max intake pressure of 32.5 PSI, subtracting ambient 14.0 pressure gives 18.5 PSI of boost.



I immediately swapped back to the stock filters, drove a few extra miles to reduce the intake temp back below 100F and tested again.
Performance with the stock air filters back in resulted in nearly the same as before: A max torque of 551 lb/ft and a max intake pressure of 28.9 PSI, subtracting ambient 14.0 pressure gives 14.9 PSI of boost.



Summary:
Obviously this is not an absolutely scientific test, but it supports the general belief that the stock air filters reduce performance. Since the before and after stock air filter runs were within 0.5% of each other, I think my methodology is reasonable.
Let me say that I could immediately feel the difference without the air filters. It felt like a different car, much bigger improvement than adding a tune.
The filter-less setup improved torque by 12.7%. My simple air filters improved torque by 5.5%.


The AMG M275 is rated for max boost of 22 PSI, but I have never seen more than 18.5 on either Torque Pro or the mechanical gauge I have inside the car. Not sure if true, but have heard that deleting CATs lets engine breath better and reduces necessary boost.
I have experimented with adjustments on the turbo wastegates. I think mine are a bit tighter than stock which should allow higher boost. I once tightened then a lot more but then got misfire codes and limp mode.
I once also tested with the stock air filter removed and the turbo inlets just open to the underhood air. Torque was 605 lb/ft with 18 psi boost; probably less than the no-filters setup today because it was sucking in hotter air.
I have ideas and parts for some other possible air intake systems, but they will have to wait.

 

Solid and straightforward test, reinforces what Fluid Motorunion found when tuning their customers' m275.

I don't think i've ever heard of anyone actually hitting the theoretical 22 PSI. I've heard some people mention there may be some boost leakage with the stock pressure-side (turbo outlet to intercooler) intake hose clamps, so they used t-bolt clamps instead. I bought four for funsies to replace, $13 on ebay.