Receive Mixer experiments for 24GHz

My last set of experiments for receive mixers involved two different mixer types for 24GHz.  The first was a Marki Microwave harmonic mixer made for the 22-33GHz (?) range.  Again, digging through my pile of stuff I had bought three of these mixers at Microwave Update many years ago.

A microwave part is just a fancy paperweight unless its used in something.

This mixer has sma connectors for the LO and IF - from what I could find on the net, its IF was in the 1GHz range.  The waveguide input appears to be WR34 (?) or some non standard flange - its feed is smaller than WR42, the hole spacing is slightly smaller than WR42 waveguide.  I found I could attach a section of WR42 with one screw only - obviously the fix is to drill out the holes on the WR42 flange which im not crazy about defacing my small stash of waveguide but its needed to get a good mechanical mount with all four flange holes.

Anyway, hooking this up to my Elcom 11.8GHz oscillator and 432 for an IF detected full scale the 24ghz test source in the shop.  Im planning for this to be a spectrum analzyer extension mixer - it could also work for a bilateral mixer but I need to test it for transmit.

While I had my 24G test setup running, I tried a Celeretek receive mixer.  LO was around 14dBm, IF 432 MHz yielded full scale readouts on the FT817 with the test source.  This mixer will definitely work for my backup 24G radio.

40 GHz downconverter at 47GHz

Many years ago I bought a 40GHz downconverter assembly.  The input was WR28, there was an LO input and IF output SMA.  The IF was in the 1GHz range and the LO input was listed at 8 to 10GHz.
I never used the downconverter as I had convinced myself it wouldnt work at 47 - after all, thats what the surplus id tag read.
I finally decided to open the downconverter and see what it looked like on the insides.  There was a PC board assembly for the detector post amp but what was interesting was the LO and RF input:  There was a hybrid ring and what looked like chip diodes.  The LO transmission line had a tiny chip that must have been the multiplier.  What was interesting was that there were NO filtering on either the RF or LO transmission lines.  So, what would stop me from feeding in 11.8GHz?  And, why not see if the mixer post amp was broadbanded and worked at 432MHz?
I fed the multiplier with 11.8GHz from my Elcom oscillator and hooked up the IF to my FT817 tuned to 432MHz.  I turned on my 47GHz weak signal source brick and quickly found the test signal - I now need to try higher LO drives (I was using 13dBm) and try the gain pot on the post amp board.
So, the lesson is: if you dont try it, you never know if it works.....

Picture 1:  RF input on the left, LO input on the right.  IF out the top.

External view of the 40GHz down converter

Back at it: experiments with harmonic mixers from the surplus pile

Its been a while since my last threat to post something.  I attended Microwave Update in St Louis and it was great to see many of the midwest microwave ops which got me back to building and experimenting.
My primary focus was to get back to building a spare 24 and 47GHz transverter.  47G had some challenges as I have yet to find any surplus made for that band that was not super expensive.  Even the surplus was pricey.  I found out about some Alcatel 38ghz detectors on the auction site that appeared to work good as harmonic multipliers.  I ordered a couple and gave them a try.  There are two SMA connectors and a WR28 input.  One of the SMA's goes to a 'detector' and the other goes to a diode assembly that would act as a multiplier - I dont know what the design intent of these were - power detectors?  Anyway, I fed the multiplier SMA with a 11.7GHz oscillator and listened for a signal with my cheap diode mixer/detector made out of an SMA flange and a diode from a Qualcom 'gold board'.  The IF was 432 MHz.  I found the test signal generated by the Alcatel mixer.  I tried the 'Detector' SMA but nothing was heard.  The picture below is of the Alcatel mixer being fed with the 11GHz brick (X4 multiplication)

I found this picture of the Mixer insides - the top is the 'multiplier' signal path, the lower is the detector signal path.  Supposedly there is a low pass filter in the detector path which explains it not working with the 11ghz brick.

Experiments with 47ghz and 78ghz with 12ghz diodes.....

I decided to do some further experiments at 47GHz and try out 78GHz.  I had obtained a DB6NT transverter at Central States VHF conference a couple years ago and have built a CW transmitter/downconverter for 47ghz (power out < 10 dBM) but I wanted to try a simple spare 47GHz system using the BAT-15 or similar diode assemblies soldered onto an SMA flange or .141 hardline.  I have enough odds and ends for testing/detecting a signal.  No way to tell power output though outside of building a down converter to hook up to my HP 141 spectrum analyzer but without a calibrated source, its all guesstimates.  My thought is - build it, take it out and see how far I can be heard....

First thing to resolve was the IF frequency for the transverter which will drive the local oscillator requirements.  Working the math assuming operating frequency of 78.192 GHz and 47.088 GHz, and assuming 432 MHz for an IF.

For 78.192GHz, the LO would be 12.960GHz and for a beacon, the LO would be 13.032 GHz.
For 47088, the LO would be 11.664 GHz (432 MHz IF)

I picked up an Elcom phase locked DRO from the auction site for the 11.664 GHz - but for the 78GHz LO, I lucked into some frequency agile bricks from CTi and also MA/COM - and even better - the CTi bricks run on +15 Volts!  The MA/COM one runs on +20v but its nice to be able to have LO source voltages that are not positive ground like the ones built for telecom plant.

I used my trusty HP synthesized oscillator to find the magic frequency to lock the brick at both 12.960GHz and 13.032GHz - so the first step was to build up a LO and include varactor tune to lock the LO.   I built a test source LO to get things started and then add in the varactor/PLL when I get that built.

Then, build up a couple test assemblies with diodes on SMA bulkheads and an IF diplexer.  The plan is to have the diode assembly to be more flexible with dish/feed mounting.

Back from the void

Well, its been 3 years since my last posting - so its time for an update with new research into 78 GHz and inexpensive home built mixers.  Stay tuned.....

47 GHz part 2.5 - oh the joys of troubleshooting....

Well, I finished wiring up the 47G down converter and started turning up and testing using a gunn source for a signal and using an HT tuned to 912MHz WBFM as the IF.  And, nada.....
At first I suspected the 10MHz oscillator as sometimes it seemed a bit flakey but the 100MHz PLL was locked.  So then I checked the phase output of the 12GHz brick.  It was merrily sweeping through its lock range.  Well crap.
I decided to take the 100MHz pll off line and feed the brick with my HP synthesized oscillator set at 100MHz.  Still the same issue.
A note about this Brick - it does not have a cavity tuning screw but rather has an automatic sweep feature that engages when lock is lost.  I suspect this was used for the brick to quickly jump to a new frequency as the reference changed.  Anyway, I had a second brick and tested it at 100MHz - same problem.
So, I started flipping the thumb switches to 99MHz, 98, 97, etc.
At 99MHz, the test brick locked right up and I noticed the phase voltage was around10 volts.
As the reference went lower in frequency, the phase voltage decreased until 92 mhz and
went out of lock.  So what I found was that 100 MHz was right at the hairy edge of the
lock limit.  101 MHz would lock fine.  OK so the brick is good, the 10 MHz ref is good.
Then I saw the phase voltage drop to 0 and one of the tantalum caps on the regulator
board feeding the mixer multiplier and amp was smoking.  I quickly disconnected the voltage and decided that was enough fun for one night.  Guess I have to replace the tantalums now.  And, figure out
a new reference oscillator to drive the brick and the associated IF.....

24 GHz signal source

This is a 24GHz signal source (estimated 100mW out) used for testing.  It can also be used with a harmonic mixer for 47 GHz.

The design comprises of a brick oscillator tuned to 12.096GHz.  I am using a crystal oscillator
running at 96MHz for reference.  Note this oscillator has a heater for stability - plans are to
replace with a Direct Frequency Synthesis source fed by a 10MHz oscillator.

The Brick feeds a X2 multiplier (surplus) to generate 24.192GHz.  The assembly runs
off 24 Volts and uses regulators to supply voltage to the reference LO, Brick, and multiplier.

The multiplier is attached to an isolator and a WR42 output.

Now, for 47GHz, the output can be fed to a harmonic mixer - in this case, the homebrew
mixer I made from soldering chip diodes to the SMA connector discussed in an earlier
post.  This mixer acts as a half frequency harmonic mixer with a high side 1296 IF to
reach 47.088 GHz    (2x24192) yields 48.384 GHz -1.296 GHz = 47.088 GHz

This text mixer has only been tested in the shack but has detected signals from
a 47G signal source to confirm operation.

One option for this assembly was to include a PIC keyer to provide CW keying
for a beacon although this modification has not been pursued.  The keyer would
make/break the voltage line to the multiplier for keying.