23>47GHZ doubler  using HSCH-9401

ON4IY  -- March 2010


0. Intro

Parts:
-    PCB DB6NT
-    Housing Hubert Krause
-    Epotek H20 group buy frhyper
-    HSCH-9401 Ebay

I wanted to make a small beacon or cw tx on 47GHz to support our experiments.  I used wire bonding as well as the more classical glueing. I thought the bonding would be superior but that is not the case....I only obtained about -10dBm at 47 GHz using bonding whilst glueing delivered a few mW without a lot of tuning. So for 47GHz, you need a lot of bond wires - or flat bond wire - and the flip mounting does not help at all either.

See Dubus 1/92 and Technik IV for more info. I received PCB DB6NT version 2, and the holes don't line up with the screws in the box. So, first align the waveguide on the PCB with the hole in the milled box and work from there - this is easy if you use backlighting. The PCB was way too big - just cut excess. I did not glue the PCB to the box since with 2 M2 screws and the waveguide end block held the PCB well. Tuning was not critical, some snowflaking needed to get the 24GHz power into the diode.

1. Project status

2. Setup

The test set-up is simple, 122MHz from the generator goes to a DB6NT 11GHz multiplier chain, this is attenuated to 5mW and goes into an Ebay doubler delivering 100mW at 24GHz used to pump the 47GHz doubler. Please excuse my French :-).


3. Testing and initial units: results

Working resistance of the diode was 210 ohms for about 2V at 100mW on 24GHz.



 
-10 dBm on the analyser using 10dB coupler. I read -8 dBm on the calibrated power head, so Pout estimated at 2dBm.




PCB mounted in the milled box
          

      
Mounting of the diode on its back and gold bond wires
(the mess I made with the glue was to see if I could improve results - I could not)




Classic glueing using Epotek H20


Author:  on4iy@qsl.net. Version: Sat Jan 30 17:06:14 CET 2010