[Microwave] Building a noise source/head

[Scott Townley]

I have gathered a few articles on building your own noise 
source.  Hopefully someone on the list(s) have done that before!
1. The series diode configuration on the W1GHZ web pages looks simple 
enough...but there is no DC block anywhere.  Wouldn't that allow a little 
bit of the diode DC bias to appear at the noise source output?  I would 
think that's NG for GaAsFET preamp testing.
2. Bill Sabin wrote some up in QST (May '94) that are shunt configurations, 
and included a blocking capacitor.  But...noise sources are intended to be 
very broadband devices (decades) and the use of a *single* blocking cap 
seems to me to be quite bandlimiting, especially when you're aiming for the 
5-10GHz range on the upper end.  When I was a wee lad I was taught to 
parallel several caps to make a broadband bias-T (e.g., 10pF, 500pF, 
.01uF)...is that technique (or some variant) applicable here to what 
amounts to a broadband DC-block?
3. The noise diode manufacturer shows a very simple shunt configuration 
with a blocking capacitor, and the single-pole rolloff equation on the low 
end, but no word/guidance on dealing with blocking caps above 
self-resonance.  So back to the bias-T question in 2.

Any other experiences or pointers would be appreciated.



Scott Townley NX7U
Gilbert, AZ  DM43di
http://members.cox.net/nx7u