(LONG) Re: [NLRS] Homebrew gear to test at 10 GHz?

[Gerald]

On Thu, 2005-05-05 at 22:51 +0100, Donn Baker wrote:
> 
> Hi Scott,
> Glad Chris pointed out Peter's web page... unfortunately, thats about all I
> know about.  There MUST be others, but I don't know where.
> 
> As Gerald also points out, homebrew test equipment for 10GHz is not easy,
> but it can be done.  There are, however, a few things that you need that
> you really can't build.  My list is:
> 
> 1) HP 415E SWR meter.  This is THE basic tool.  Its actually a 1000 Hz
> amplifier and a precision 60 dB attenuator.  Its use to measure SWR,
> attenuation, and other stuff.  E-bay will usually have one or two; expect
> to pay $50-$75 of one promised to work.  There are three out there today:
> one for $49 claimed to be fully functional; one for $65 from England ; and
> (the joke of the day!) one for $1,499 with eight pieces of waveguide and
> slotted line parts.  Yeah, useful, but not for $1,500 !!!

Actually any of the 415 family work decently, even the tube types and
give different numbers than a DVM would. The 415 meter curve is adjusted
to follow the square law response of the point contact microwave diode
(1N21 and 1N23) at low levels.

> 
> You CAN use a standard voltmeter for some measurements, but the 415E really
> needs to be on your bench.  A digital voltmeter works, but an analog meter
> is better.  Not only is it easier to small changes, but usually, at least
> one scale will be calibrated in dB so you don't have to do all the
> arithmetic to convert voltages to dB.
> 
> 2) Detector(s)  A 1N34 diode won't work at 10GHz.  You need to spend some
> money to get at least one or two that are good.  It is possible to buy
> diodes and make your own for most things.  You still need at least one for
> the "lab standard."  HP 423A is top of the line, HP 420s only a little
> below that.  Wiltron/Anritsu make similar ones.  You need ones that match
> the connector series you use (i.e., N, SMA, etc.).  The outputs are almost
> always BNC.
> 
> 3) Attenuator(s)  At least one each 3dB, 10dB, 20dB high quality
> attenuators.  Make sure they're rated to at least 10GHz; 18 GHz (if using
> N-connectors) is better.  Use these only for calibration of other, cheaper
> attenuators.  Don't use'em for run of the mill stuff.

N connectors might have a waveguide mode at 18 GHz? Might be that you
need SMA that high in frequency.


> 
> 4) Dummy load.  At least one high quality 50 ohm load.  Again, use ONLY for
> calibrating other, cheaper loads.

I've made good loads from a connector and a resistor. I start with a
connector made for RG-8. I take a short piece of RG-8 and pull out the
center conductor and drill the dielectric to take my 51.1 ohm metal film
precision resistor. I then taper that dielectric so its gone to a point
at the back end of the resistor and I squeeze the braid to fit that
tapered dielectric. Then I mount resistor and cable pieces into an N
connector as if it was coax. The theory is that the characteristic
impedance of the braid and resistor should be the same as the resistance
to ground. Which properly made makes a curved taper, but a straight one
works better than none at all.

> 
> 5) Signal generator w/1000 Hz modulation.  Obviously, you can use your
> transverter with a 1000 Hz tone being transmitted (AM or SSB.  Not FM)
> 
> Once you have these, you're on the way.  To measure SWR (or Return Loss),
> you also need a slotted line (lab-quality measurements), or Directional
> couplers (usual measurements).  Directional couples can be home-built...
> the Radio Society of Great Britain (RSGB) Microwave Handbood, or their VHF
> Handbook have details for making "cross couplers" for 10GHz with scraps of
> WR90 waveguide.  You can also buy directional couplers for 10GHz at
> reasonable prices... $10 to $30 depending on the exact type, etc.  20 to
> 30dB coupling are the most common.  You probably don't want anything below
> 10dB, and I'd even stay away from 10dB ones, too.  You don't want anything
> over 40dB, either, without a 415E to use with it.  You have to be careful
> about what connectors are being used.  Waveguide couplers are cheapest, but
> you'll need transitions to get to/from the waveguide.  N-connectors are
> common on older stuff that's not waveguide and SMAs are used on newer
> equipment.  (Transitions can be homebrew, by the way.)

Lots of directional couplers show up on epay at low prices. Many with
odd part numbers not found in manufacturer's catalogs. They may give
good directivity outside their rated frequency range, but the coupling
value may not be as labeled.

> 
> Hook up the generator (with 1000Hz tone on) to the directional coupler to
> the unit (antenna) you're testing so that it measures in the forward
> direction.  Connect the detector to the coupled port.  Note the voltage,
> etc. on the 415E or voltmeter.  Reverse the directional coupler so that it
> measures the energy being reflected from the antenna.  Note the dB or
> voltage.  Convert the voltages (both of'em) to dBs.  The difference in dBs
> is the Return Loss (RL) of the antenna.  You'd like to see a >20 or 25 dB
> difference.  If you "diddle" with the antenna, you need to verify the
> forward measurement to get an accurate RL, as changing the RL can affect
> how much power is put out by the generator.
> 
> 73 Donn
> WA2VOI/0

-- 
73, Jerry, K0CQ
Dr. Gerald N. Johnson, electrical engineer
All content copyright, Dr. Gerald N. Johnson