Filter hacking for 2395 ISS TV

[radiogareth]

From a construction point of view a comb filter does not need a tap feed into the 'first digit' (of an interdigital) so that makes reliable repeatable replication (3 'R's'??) easy by virtue of using pieces of wire, tubing, brake pipe etc basically anything you can solder to a centre pin.
Slots would indeed be possible, but software simulation would surely be a better way these days?
Gareth

[G3YKI]

I may try adding additional screws in between the posts. That is supposed to adjust the coupling. Might be worth a try.

Colin.

Yes, I have seen many filters that use that technique and it is easy to do. I believe from experiments on a similar filter that it increases the coupling, but I could be wrong. ( I cannot see any intuitive reasoning one way or the other) I have also seen (and made) filters where the resonators are partly screened by dividing walls. This, I am sure reduces the coupling and would enable considerable reduction the overall length. It may not easy to do in in practice in your tube filter.

Having said all that, your filter is probably doing what you need. If it passes the signal you want, whether or not the rest of the passband has the text book flat top response becomes an academic exercise.

[g4eml]

Yes, from what I can find, tuning screws between the posts increase the coupling. Some texts recommend deliberately spacing the posts a little too wide and then using the coupling adjust to get the final result. It would probably be better if a design didn't need them as it makes the alignment more complicated.


Colin.

[g0mjw]

The spacing on my filter is much less. 19mm between the input/output rod and the first post, then 38mm between the posts, which are interdigitally arranged. The in/out posts span the waveguide so are 32mm but approx 25mm long at 10mm dia. The other posts are about 8mm dia. Hart to tell how long but it looks like about 3mm of gap at the end. The tuning screw goes into the post so it must have a hole in it.

Given it is shorter, it might be a better route size wise.

Mike

[g0mjw]

Many years ago, before NanoVNAs, I asked Bert Moddermann to make me a 13cm filter. It is fairly small, 110mmx40mmx20mm and excellently CNCd as usual from Bert. Here is the response.

13cm-5pole.jpg (93.16 KiB) Viewed 120 times

It has 5 tuning screws

Cavity width 31.8mm, length 103mm, depth 16mm, rods 6mm dia 27mm long, spacing 19mm, 21mm, 21mm, 19mm, tap 7mm up from ground. It uses M4 tuning screws, A2 stainless !!!

13_cm_filter_image.jpg (1.16 MiB) Viewed 120 times

The loss is low, so I assume the Q is also low. It is not what we want as the bandwidth is 150MHz or so. About right for covering the band, before we lost so much of it.

Mike

[g4eml]

I added some additional coupling adjustment screws between the 70mm spaced poles.

Winding in the two outer coupling screws effectively increased the coupling of the two outer pairs of resonators. So probably similar to an untuned spacing of 65 70 70 65.

The filter now tunes up with a reasonably flat passband and good return loss. The overall bandwidth is a bit wider at 22MHz as would be expected with higher coupling. Increasing all of the spacings should bring the bandwidth back down again but that would involve a lot of re-drilling.

IMG_20251110_154553.jpg (131.35 KiB) Viewed 89 times

Smaller cross section tube should reduce the resonator spacing and hence the total length.

I will now wait until I have some of the smaller tube and repeat the process based on what we have learnt.

Colin G4EML