Repair-It

Some info. It doesn't seem to acknowledge the typical failure mode after miles: the linkage can break due to carbon buildup internally.

There are kits to fix broken linkage though cleaning inside seems like a good idea.
https://www.mercedesmedic.com/mercedes- ... ever-flap/

Where DAS goes is to check the dashpot with an independent vacuum source to rule out anything on the control side and it basically wants you to verify that the dashpot moves the linkage AND that both side swirl valve hall effect sensors indicate the mode has switched (can measure at the sensor as well as data PID but I don't know if the data PID splits right/left or is just a single value). So this would be easy to replicate with bidirectional control. The 0521 is probably set because one or both swirl valve sensors are not reporting the position that is expected.

The hall effect sensors are actually at the rear of the intake manifold, if you picture the swirl valves on a shaft (with one valve per cylinder) on each side.

In case this escalates.

They have two in stock but DAMN. https://www.mercedesbenzpartsshop.com/o ... 2140240164
$1500 is rough. So that's probably out.

Rockauto has various options including OE Pierburg. https://www.rockauto.com/en/moreinfo.ph ... 36&jsn=590

And there is this kit from FCP: https://www.fcpeuro.com/products/merced ... urg-516475

On the cheap end there is a Uro part. https://www.rockauto.com/en/moreinfo.ph ... 36&jsn=586

So there are options here.

From the Mercedes medic article was a video. I believe this is the kit from the video: https://www.rockauto.com/en/moreinfo.ph ... 24&jsn=568

But really if the flaps internally are beginning to get sticky due to carbon, making the external linkage stronger doesn't seem like a real fix.

Perhaps you could seafoam the intake and jiggle/actuate the swirl valves during the procedure.

Alright so I posted DAS direction as well as various part solutions. I think the diagnostic approach is fairly simple at least (and of course it's killing me that I don't have the car here).

1) Visual inspection. Checking for all linkage to be connected, nothing broken, no cut or disconnected vacuum lines, etc.
2) Actuate swirl valve via scan tool, probably at idle. Look for dashpot movement and follow motion through the linkage to both side swirl valve bell cranks. There will likely be a problem here and the possibilities are:
a) No motion at all. Completely seized linkage. Or dashpot not actuating.
b) Partial or slow motion.
c) Dashpot moves fully but there is a break somewhere and the motion isn't making it to both swirl valves on either side.

With situation a) or b), I would lubricate (possibly pop off some of the rods if access allows) and actuate. Also I would consider trying to feel swirl valve motion by hand to see if they are gummy or partially stuck. If the dashpot may not be trying to move, see if there is a disconnected or broken vacuum line. Consider connecting a vacuum source and see if that does anything (this is where DAS takes you). Dashpot diaphragm can also be torn internally. Any of the dashpot type issues would probably be a best case scenario (other than linkage lube) but I don't know how much of that stuff can be replaced with the intake installed on the engine.

With situation c), that pretty much puts it into needing a mechanical repair. But at that point, you'd want to make sure the swirl valves themselves (on both sides) are fairly free to move. Since the manifold has to come off to replace the broken linkage, you could try to clean and free up the swirl valves without taking the manifold apart. What I haven't posted yet is the take-manifold-apart option. I think there are kits for this.

You would want to plan to have the car down for a while when removing the manifold because inspection of the manifold may suggest replacement is more prudent. And you wouldn't want to have one of those linkage repair kits on hand if the decision changes after removal. After removal, you'd make the call to either select one of the several replacement manifold options, or order from the one or more repair kit options.

Based on relative location and access on the V8, inspection and lube are the only possible actions with the manifold on the engine.

Here are the relevant work instructions for reference.

I am still looking for the official theory of operation for all of this stuff to understand the impacts if not fixed.

Finally found the theory of operation. This is the best overall description of how everything works. The intake manifold features are designed to optimize the torque curve (broaden - low end and high end) across RPM operation. It does this with two primary systems:

System A) variable runner length. It has a short and long runner switchover capability. Long below 3500 rpm and short above 3500 rpm. The M272 has two dashpots for this and the M273 has one. I don't understand why the difference but it doesn't matter. Default/resting operation is short runners so the engine at least makes rated power output. It may not use the long runner at low rpm unless engine load is above 50% according to the literature.

System B) Tumble or swirl valves. These are basically valves that partially close, about halfway, the intake runner right before the air enters the cylinder head. Rather than RPM based, these are controlled based on engine load. WIS notes that "at partial load" the tumble valve closes and restricts the port volume to increase air velocity and swirl as the air enters the cylinder head. This promotes better mixing which probably benefits fuel economy, emissions and part load performance/response. Default/resting operation is no swirl valve obstruction of the intake ports, so the engine at least makes rated power output.

All controls are via vacuum with control solenoids, dashpots (or perhaps vacuum servo if that sounds fancier) and linkages.

It is system B) that is the issue re: 0521/P2004.

Now I will post the rest of the details about the systems.

FWIW, the code does not seem to specify which mode it thinks the swirl valves may be stuck in. They could be stuck closed, open or somewhere in between, in one or both banks. And frankly those are a lot of combinations of issues so I guess I understand why the code basically directs you to take a look at the system, since it is not responding correctly.

What I do know is that "activation" or "operation" is to the closed position. Also it cycles this stuff one time at engine startup which in theory should ensure everything gets exercised at least once per drive. I guess that isn't enough. Or linkage gets sticky and the problem exacerbates from there....

Here are the conditions for activation of the swirl valves. They rotate about 56 degrees from fully open to fully closed, which gives an idea how far the shafts would actually move on the bell cranks.

It stands to reason that if the valves were stuck fully closed/activated, this would significantly reduce power output and high rpm performance. I note that this does not seem to fit the symptoms. If the valves were stuck fully open/deactivated, this would probably lead to decreased low rpm and light load performance/response and maybe some rougher idling. I would submit that the car may have those things. In a partial or one bank situation, it is hard to really say but those are the extremes and other symptoms would probably fall somewhere in the middle, I guess.

I might say tip in was a little soft? It probably needs the throttle damping stuff turned off, but it probably also needs the swirl valve controls working, too.

And here I have taken the liberty of reviewing WIS description of operation and made a map of how the functions are controlled across the range of engine conditions. This always comes with a caveat: actual engine calibration may differ but this should give a general idea. The tumble valve is only used at lower rpm and lower load conditions to close off the intake ports. Everywhere else it is open. So it operates and closes the intake ports only for the lowest engine airflow situation.

You can also see the relationship between the tumble valve control and intake runner length control. Similar but opposite in load, the long runner configuration is only used at high loads and lower rpms. The tumble flaps are never closed during long runner operation, as a result. The air probably builds enough velocity/resonance and closing the tumble valves would just decrease performance/torque at that point.

Here is an improved version of the table.