BATC Forum

In my 'useful parts' pile I have a 13cm Diplexor (CommTACT CD2137/100SK-E apparently ex-Military so Google doesn't throw up any data)whose high band appears as per attached screenshot 2020 MHz-2310MHz (trace from Satsagen software running on a Pluto, truly amazing!!).
Each 'port' has 16 locked screws to adjust. The 'low' port passes 2020-2080 in a similar fashion.
I recall reading somewhere that you need patience and kit to do this correctly. I'd guess that unscrewing each screw say 45 degrees would move the passband up. Ripple and such can then be optimised as long as patience allows.......
Or, build a 'Stop-end pipe cap' filter(s) like these http://www.ko4bb.com/ham_radio/Pipe_Cap_Filters although as we are metricated I can get either 22 or 28mm i.d. caps for sensible money via eBay. I've plenty of hardline SMA patch leads I can sacrifice to build them. The larger '11/2"' start to cost silly money.

Anyone able to comment.....on either routes? Application is crud filter before big PA on 13cm.

Thanks

Gareth

Move it in small steps it probably the easiest way unless you have a vector network analyser. If you do there is a technique I have used in the past which involves detuning each cavity, e.g. removing all the screws, and then tuning each one in turn working from one end looking for a 180 degree phase shift in the return loss (s11) as each cavity comes into resonance - that gets you in the ballpark. It helps to use the smith chart so you can see the phase moving. Tune the input cavity first, work towards the output. There is also a time domain technique I have not tried. If you are lucky enough to have a full two port VNA, you probably already know how to do that.


Someone has made a video of a similar technique:

https://www.youtube.com/watch?v=Md6lFOKvJzI

Thanks for the pointers - and the video link. Its pretty much how I expected things to go, might try on a 70cms unit first with fewer cavities!

I have a NanoVNA but that is yet to be software expanded upwards beyond 900MHz although its performance at 2.4GHz might leave a bit to be desired. I think I'll have a crack at the low pass side first, as that is currently of no use on 13cm so if I mess it up its not a 'loss' as such. Being able to watch the PB live should provide the visual feedback I need. Got to remove the 'glue' from 16 screws first
Onward...
Gareth G4XAT

There is a link to using TDR to align filters on one of my old blog posts
http://g4guo.blogspot.com/2015/11/using ... ry-to.html

- Charles

Seems like Keysight have made this an email request only item now (Literature Number 5968-5328E)
however for those that are interested a Google for AN 1287-8 will find the document from a non Keysight address.

- C

Thanks Charles, a very interesting document. I wonder if I can do a 16 section tune up???

Gareth

It all looks good on paper but I found the resonators at the centre of the filter become very difficult to see.
I think that is obvious from the S11 and S22 plots on my blog.

My HP VNA didn't come fitted with a licence for TDR but there was a guy in the HP group that can create
a licence provided it is for non commercial use. I think that DG8SAQ VNWA 3 software supports TDR.
Even it you don't intend to use TDR it is as you say an interesting paper to read.

- Charles

Decided to experiment with some Pipe caps (28mm) and as I have an abundance of hardline links with SMAs each end I cut one up using an Ikea pipe cutter. First to length then -8mm of outer tube. Anyway, it works very well on 3.4GHz

For the next one I'll lock a couple of nuts on the end as per the guidance here http://www.ko4bb.com/ham_radio/Pipe_Cap_Filters/

At the moment I only have threaded steel rod, but the BW and insertion loss are what I would think is quite reasonable as-is. Before I turn down and thread some brass, will it make any difference?
Trace from Satsagen. Its great Gareth

Brass screws certainly make a difference on 10Ghz. The loss is much less.

I discovered that at 3.4 GHz. Makes a huge difference. I have been salvaging silver plated screws from otherwise useless filters.

Mike