Preface
I've put together this document for two reasons.
First, to give a real-world example of how cheap hamfest connectors
can ruin your day. Second, to give me an excuse for trying out
the digital camera that Santa brought me for Christmas. You will
probably see that my photography skills leave a lot to be desired.
I spent my belated New Years vacation today working on a repeater
project. One of the things I needed for this project is a pass/reject
cavity to use to keep a link transmitter clean to prevent from
desensing another UHF receiver on the tower. So, I robbed a cavity
off a Decibel DB4076Z duplexer, put it on the bench, and hooked
it up to the service monitor for a quick tune-up (click
here for info on Decibel UHF duplexers). I soon realized
that there was something very wrong with the cavity. I could
only get about 18 dB of notch depth, and I couldn't get the notch
to move more than about 2 MHz away from the pass frequency. Having
used these cavities before, I knew the notch should be at least
twice that deep and the cavity should tune much wider than a 2
MHz split.
<Picture of bad
cavity response sweep - 640x480 - 66K>
The most common failure with pass/reject cavities that use small piston trimmer capacitors used to tune the notch frequency is arcing and subsequent burning or welding of the concentric "plates" inside the capacitor. This is usually caused by lightning, excessive power input, or mechanical trauma causing the plates to not properly align. Figuring that this had happened to this particular cavity, I unscrewed the upper threaded half of the capacitor to take a look. It didn't appear to be damaged. A quick check with an ohmmeter also confirmed that it wasn't shorted.
Cavities are pretty simple devices; there isn't much to go
wrong with them. Unfortunately there isn't a whole lot of repair
examination that can be done from the outside of the cavity.
Figuring I had nothing to lose, I proceeded to drill out the rivets
that hold the bottom cap on the cavity so I could take a look
inside.
<Picture of inside
of DB4076Z cavity - 640x480 - 79k>
Everything looked normal. No signs of arcing inside the cavity.
The loop and solder joints appeared to be OK. So I removed the
loop assembly to examine it further.
<Picture 1 of DB4076Z
loop - 640x480 - 27k>
<Picture 2 of DB4076Z
loop - 640x480 - 46k>
<Picture 3 of DB4076Z
loop - 640x480 - 41k>
Again, everything looked fine. Ohmmeter checks came out fine. A check of the capacitor with an LCR meter showed it tuned properly from near zero pF to about 12 pF which is typical for this type of UHF loop design.
After reinstalling the loop and putting the cavity back together, I again swept it to verify I wasn't losing my mind. Sure enough, it still showed only about 18 dB of rejection. I tried swapping cables between the cavity and the service monitor with no change in response. There was only one thing left, and that was the type N tee adapter. I rummaged through the box o' adapters on the bench and found another one, a nice silver-plated job.
Lo and behold, the tee connector was the culprit! The notch
depth and tunability was fine with the new tee connector, about
38 dB of notch depth at 5 MHz spacing.
<Picture of both
adapters, good one on left - 640x480 - 65k>
<Picture of proper
cavity response sweep - 640x480 - 61K>
<Picture of bad
cavity response sweep again for comparison - 640x480 - 56K>
I learned many years ago that a lot of the adapters sold at hamfest bargain tables are junk. My first experience with a bad adapter was a UHF elbow adapter that degraded a transmitter's output from 50 watts to 35. Dissassmebling that UHF elblow connector revealted the connection between the two ends was made with a steel spring akin to what you'd find in an old-style ball-point click pen, a both lossy as well as highly inductive connection that played havoc on UHF RF. Being both aggrevated and inquisitive, I decided to try to take apart this tee to find out why it was ruining the performance of the cavity.
You can remove the center pin of the male side of most UHF
or type N tee connectors by grabbing it with a pair of needle-nose
pliers and unscrewing it from the center conductor that runs between
the center pins of the two female connectors. So that's what
I tried. I turned it and turned it and turned it, but it didn't
appear to be unscrewing. So I gave it a good tug with the pliers,
and it slid right out! It wasn't threaded into anything! Closer
examination revealed that the flat bottom end of the center pin
simply contacts the round center conductor running between the
two female ends of the tee. Not only is this kind of poor tangential
contact very prone to problems to begin with, but there was nothing
captivating the pin to keep mechanical pressure on the connection.
<Picture of connector
and removed center pin - 640x480 - 66K>
Apparently the male center pin was not making solid contact with the center "thru" conductor inside the body of the tee adapter. The DC-open connection was effectively a capacitor in series with the cavity loop. No wonder the cavity wasn't tuning right.
I don't know where this bad tee came from. I know I didn't buy it; it probably came off some other piece of equipment. If it doesn't say Kings or Amphenol on it, I usually throw it away but somehow this one made it onto the bench.
OK, here's the bottom line. I took the time to write this up. You took the time to read it. Save yourself time and aggrevation in the future. Round up all of those adapters in your toolbox, in the lid of your service monitor, in the box on your bench, in your medicine cabinet, under your couch cushions, etc. and take a good hard look at them. Throw away the ones that you can't find a brand name on, and make a mental note to never again buy cheap hamfest adapters. Pay the extra buck or two and get the real deal.
Oh yeah, I had a loop from a lowband Decibel pass cavity on
the bench so I took a picture of that as well for the heck of
it (one end is propped up with a GE ICOM to keep it level - ignore
that).
<Lowband pass loop
and UHF pass/reject loop - 640x480 - 59k>
The digital images were shot using a Nikon Coolpix 950, an Ikea halogen floorlamp, and a $5 Home Depot work lamp with 50 watt incandescent bulb. Needless to say, the color rendition, focus problems, and shadows can be attributed to my low-tech makeshift bench-studio lighting combined with my lack of photography skills. The images are JPEGs, compressed heavily to speed up download time.
End note. Please let me know if you found this worthwhile reading. I'm always reluctant to write up stuff like this because I don't know if anyone a) actually reads it, b) finds it useful, or c) is able to make sense of my ramblings. I'd appreciate your feedback. Send email comments, good or bad, to Jeff DePolo
73 de WN3A
Visit the N3KZ repeater systems' home page at https://w3abt.seas.upenn.edu/repeater-system/