Regarding RF attenuator linearity

[cjrathi]

At first sight, the results suggest incorrect level measurement.

What do you meany by incorrect measurement? The attenuation at 0dB is around 7 to 8 dB. This is bcoz Insertion loss of 3 att parts and also some connector losses

 

[FvM]

What do you meany by incorrect measurement?

I mean the unexpected too high attenuation at -80 and -90 dB. I don't see how HMC425 should generate it. All posts have been assuming insufficient attenuation due to crosstalk, which can be achieved easily. How do you detremine exact attenuation values at -80 and -90 dB? Looking at the uncertainty in your table, it seems like these values are lost in noise?

The base loss in 2.4 GHz band is more or less according to part specification, I think. The 5.7 GHz band attenuation is slightly too high which possibly refers to PCB design problems, also gain ripple at some attenuation steps.

I'm presuming, that the HMC425 get's correct Vdd and digital input levels.

 

[chuckey]

The first thing to attend to is the big jump in attenuation between 2.4 and 5.1 GHZ. Are all your RF lines of the correct impedance?
Frank

 

[cjrathi]

Even i don't understand why is so much insertion loss in 5 GHz band even though IL as per datasheet is same for both the bands..

Also i don't understand the reason of getting attenuation ripples at high attenuation step. Is is because of noise? Then why it is not occurring at low attenuation levels?

I have followed the RF PCB guidelines well enough to avoid any problem. Also sufficient Vdd supply is getting to control pins.

Please suggest something..

 

[hosein_rabiee_g]

I cant see the table. send the measured attenuation value vs expected attenuation in at least two different Frequencies.