Benz SLR Gullwing, not sensitive to loading?

Does anyone else here has a Benz Gullwing (or LPS = same generator) wich does not react to input impendance?
I's on a Clearaudio Ovation with TT5 arm. My phono stage is a Graham Slee Reflex C in wich I just soldered 47kOhms resistors when I got the cart, before was 1kOhms.

Now I ripped some tracks to my computer to see if 100, 1k, 22k or 47k sounds better using a Y-connector with resistor. And there is not the tiniest difference in sound at all. Plus the three tracks even have the same volume level. I'm not used to that because all the MC cartridges I heard before sounded (more or less) brighter and the level dropped a little bit when I raised the loading.

Is there something wrong here?

@tealow Not at all! In fact this is good. The reason to use a loading resistor isn't for the sake of the cartridge, which does not 'ring' at audio frequencies, its for the phono section.

This is because the inductance of the cartridge is very low and also has a high 'Q' value, and its in parallel with the capacitance of the tonearm cable. Together they form a very peaky resonance at a very high Radio Frequency (RF, possibly above 1MHz) and when this peak is energized, it creates RFI (Radio Frequency Interference) that is injected directly into the phono section.

If the phono section doesn't like that, it won't sound right and so loading resistors are used to detune (make less peaky) the resonance, thus removing the RFI.

If the phono section does not care about RFI then you won't hear a difference until you get the loading so low that the output of the cartridge actually goes down.

In short, you're in a good spot so no worries!

47K is the industry standard loading and if the phono section is properly designed this will work fine with all low output moving coil cartridges. If you have to apply loading to get the 'right sound' the phono section has a design issue.

In practice, I find what Atma-sphere says to be true as regards loading LOMCs at 47K ohms, although sometimes I slightly prefer lower values of load R just "because" I can imagine I hear a tiny difference. On the other hand, if you don't hear a problem with a 100R load on a Benz LPS or the like, then something is amiss (because there is a mismatch between a cartridge with an internal R of ~30 ohms and a phono load of 100 ohms, such that a large fraction of output V is lost to ground AND you have a high frequency roll-off within the audio frequency range), as Intact says.  You got great advice from two industry professionals.

That explains it. I now know why I always felt an affinity towards 47k values for my LOMCs. Even though I did start with the recommended loading values set by the manufacturer, I gradually migrated to the higher values as each time I went higher, it sounded better than the previous values. 

Even though I did start with the recommended loading values set by the manufacturer

They specify those lower loads because they have no idea what phono section will be used and the simple fact is that many designers do not take the electrical resonance I mentioned earlier into account. This causes them to be sensitive to RFI and thus the loading resistor has a very noticeable effect.

But nothing comes for free except nothing and sometimes you have to pay even for that :)  In the case of a loading resistor, 100 Ohms is 2 orders of magnitude less than 47000 Ohms and so you are asking the cartridge to do more work driving that load and so the cantilever becomes stiffer (less compliant) due to the electrical load. This can have an effect on how it behaves in the arm.

This is not about whether loading should make a difference. The key statement from the OP is:

And there is not the tiniest difference in sound at all. Plus the three tracks even have the same volume level.

If there is no change in level on a 38Ω cartridge with a 100Ω load, then there is a problem with the test methodology.  Either the applied load is not connected or the level measurements showing no change are incorrect.  Mulveling's first response pointed this out and there is no sense in going on tangents until this obvious problem is addressed.

dave