SSA3032x bandwidth bug at around 1.03, 2.07 and 3.12 GHz?

[rfspezi]

I just noticed, that there is something strange going on at 3.12 Ghz on my SSA3032x when switching from a lower BW to 3 MHz BW. 
Find attached a screenshot comparing the same 3.12 GHz source at 0.1, 1 an 3 MHz BW.
Also attached is the initial sweep where i discovered the "bug"?!

[EDIT]
When you let the source sweep across that frequency range, you see a lot of different variants of distortions - but only when BW is set to 3 MHz.

[EDIT2]
Similar Problem at around 1.03 and 2.07 GHz! 

[rf-loop]

I just noticed, that there is something strange going on at 3.12 Ghz on my SSA3032x when switching from a lower BW to 3 MHz BW. 
Find attached a screenshot comparing the same 3.12 GHz source at 0.1, 1 an 3 MHz BW.
Also attached is the initial sweep where i discovered the "bug"?!

[EDIT]
When you let the source sweep across that frequency range, you see a lot of different variants of distortions - but only when BW is set to 3 MHz.

[EDIT2]
Similar Problem at around 1.03 and 2.07 GHz! 

The title is wrong or least alternative trumpth. Bug is: You have forced it working out of specs and then you claim with question mark it do not work right.

Who have promised there is 3MHz RBW? No one. Just turn off this 3MHz RBW and your selfmade "bug" is repaired.

Data sheet: Resolution Bandwidth 1 Hz~1 MHz, in 1-3-10 sequence.

If you force it for working outside of specifications only you are responsible how it functions or not.

You can hack 2.1GHz model without options to 3.2GHz model with options (but after then cal certificate is not valid)  and without opening unspecified 3MHz RBW.

[rfspezi]

First:
You could have told me that in a normal and kind way.

Second:
I did formulate the topic as question. Thats what questionmarks at the end of a sentence are used for.

Third:
Thanks for informing me that 3 MHz BW are not officially supported.

Fourth:
The true reason why Siglent is selling a lot of devices is their (inofficially supported/wanted) hackability.