Attenuator vs Termination?

[rfengg]

Hi all,

I need to test an amp at 10Ghz and rated for 20W output power.
I am torn between deciding to go whether for a Attenuator or  Termination (dummy load : 50ohms) .
The load I was looking at is rated for DC-20GHz and 50W, while the attenuator is rated for  DC-20GHz and 50W and has 50dB attenuation......both are around the same price.

Am I right in assuming that the attenuator would be a better choice since I can hook it up to the amp and then monitor the output of the attenuator on a spectrum analyzer?
If I use a load then I have no method of monitoring the power......any limitations in using a attenuator for this use case?
Even if I leave the attenuator output end open, it should act like a reasonable dummy load with -100dB reflecting back to the amp?

[xrunner]

Hi all,

I need to test an amp at 10Ghz and rated for 20W output power.
I am torn between deciding to go whether for a Attenuator or  Termination (dummy load : 50ohms) .
The load I was looking at is rated for DC-20GHz and 50W, while the attenuator is rated for  DC-20GHz and 50W and has 50dB attenuation......both are around the same price.

Am I right in assuming that the attenuator would be a better choice since I can hook it up to the amp and then monitor the output of the attenuator on a spectrum analyzer?
If I use a load then I have no method of monitoring the power......any limitations in using a attenuator for this use case?
Even if I leave the attenuator output end open, it should act like a reasonable dummy load with -100dB reflecting back to the amp?

You want the attenuator. The load will simply absorb all the power and you can't measure anything really. 20W is +43 dBm so the output of the 50 dB att. will be -7 dBm. You can then use a spectrum analyzer to check out the signal power and other parameters. Whatever power you measure with the spec analyzer will have to have 50 dBm added to it for the real power output of the unit. or you can probably tell the spec. analyzer to add it back in on a menu.

[ejeffrey]

Even if I leave the attenuator output end open, it should act like a reasonable dummy load with -100dB reflecting back to the amp?

Correct.  The return loss of a common attenuator or dummy load is rarely better than -30 dB (equivalent to VSWR ~ 1.06) anyway so a 20 dB or larger attenuator will normally have negligible degradation in matching even when the output is left open. 

[switchabl]

Even if I leave the attenuator output end open, it should act like a reasonable dummy load with -100dB reflecting back to the amp?

This is only the case if you assume that the attenuator is perfectly matched. It is true that leaving the termination open will not make a big difference at 50 dB attenuation. However I would not necessarily expect a power attenuator to have much better than 20 dB return loss at 10 GHz in the first place. Some are quite good (and some loads are quite bad), you will have to check the datasheet.

If you go with the dummy load you will need an additional coupler for your measurements.