23cm Filter
Forum rules
This forum is run by the BATC (British Amateur Television Club), it is service made freely available to all interested parties, please do not abuse this privilege.
Thank you
This forum is run by the BATC (British Amateur Television Club), it is service made freely available to all interested parties, please do not abuse this privilege.
Thank you
23cm Filter
Not the most exciting of subjects but here goes: I built a filter based on a design by W6PQL. (23cms Filters BATC WiKi).I made the enclosure from double sided PCB board and the resonators were cut from 6mm brass rod ,drilled and tapped. The "capacitors" were simply 6mm brass screws held in place by 2 brass nuts soldered on both sides of the PCB.I should add that this was the 2 resonator version. I measured the forward loss at about 3dB which is about twice that which the author measured.On another matter,I am unsure as to what purpose is served by the large screw in the centre and in practice it appeared to do nothing very useful. I have 2 questions; Has anyone tried making the resonators from 6mm threaded rod (easier to fix than drilling and tapping plain rod) and can anyone explain the purpose of the screws used both in the 2 and 3 resonator versions. I have used these types of filter before in the workplace but cannot remember seeing any screws which were not associated with resonators.I would be tempted to try making another version using UHF capacitors were they not so expensive, ie £10 on Ebay. Feedback on these filers would be welcome. Thanks,Gordon.
Re: 23cm Filter
The screws are there to alter the aperture size between the tuned elements. That allows you to control the degree of coupling between the tuned rods which alters the shape of the filter.
Insertion loss will increase as you reduce the aperture and bandwidth will reduce.
Your insertion loss does seem higher than it should be. Brass is more lossy than the aluminium used by W6WQL, but I wouldn't have thought that was the sole cause. I wouldn't bother to replace the tuning screws. Seam soldering the PCB might also contribute to the losses if the solder is in a high current path. Threaded poles would probably increase losses; none of the commercial or amateur market manufacturers use them.
Clive G3GJA
Insertion loss will increase as you reduce the aperture and bandwidth will reduce.
Your insertion loss does seem higher than it should be. Brass is more lossy than the aluminium used by W6WQL, but I wouldn't have thought that was the sole cause. I wouldn't bother to replace the tuning screws. Seam soldering the PCB might also contribute to the losses if the solder is in a high current path. Threaded poles would probably increase losses; none of the commercial or amateur market manufacturers use them.
Clive G3GJA
Re: 23cm Filter
Hi Gordon,
The centre screw would be intended to provide adjustment of the coupling between the two resonators. Ideally you should get a flat top on the bandpass response and further adjustment would start to introduce a dip at the centre of the passband. If you are seeing a single, rounded peak in the response it is possible your coupling is insufficient and that could account for the excess loss that you see. Are you able to measure the return loss? That will tell you if the filter is working properly. Power can only come out of the output if it is not reflected back to the source!
If you cannot get to the stage of seeing the flat top or double peak response, you need to reduce the size (area) of the coupling loops slightly, maybe possible just by bending them. Try to ensure that the two remain equal in size in order to maintain a good match.
I would expect the loss of your filter to be as good as the designer's as your copper box could be considerably better than a diecast box and the brass and aluminium rods are probably similar in performance. The lid would need to be in contact with the inside surface of the box, not just the outside, or rf would get into the middle of the pc board with unpredictable results. (Extra loss, leakage, extra resonances...) I am not convinced that diecast boxes make a good enclosure for filters like this.
The filter could still work with threaded rod but the loss resistance of the rod will be doubled. (A metric thread has 60 degree angle so the current actually has to travel twice as far.)
Regards, Ken
(TCPD Brookmans Park 1982-88)
The centre screw would be intended to provide adjustment of the coupling between the two resonators. Ideally you should get a flat top on the bandpass response and further adjustment would start to introduce a dip at the centre of the passband. If you are seeing a single, rounded peak in the response it is possible your coupling is insufficient and that could account for the excess loss that you see. Are you able to measure the return loss? That will tell you if the filter is working properly. Power can only come out of the output if it is not reflected back to the source!
If you cannot get to the stage of seeing the flat top or double peak response, you need to reduce the size (area) of the coupling loops slightly, maybe possible just by bending them. Try to ensure that the two remain equal in size in order to maintain a good match.
I would expect the loss of your filter to be as good as the designer's as your copper box could be considerably better than a diecast box and the brass and aluminium rods are probably similar in performance. The lid would need to be in contact with the inside surface of the box, not just the outside, or rf would get into the middle of the pc board with unpredictable results. (Extra loss, leakage, extra resonances...) I am not convinced that diecast boxes make a good enclosure for filters like this.
The filter could still work with threaded rod but the loss resistance of the rod will be doubled. (A metric thread has 60 degree angle so the current actually has to travel twice as far.)
Regards, Ken
(TCPD Brookmans Park 1982-88)
Re: 23cm Filter
Thanks for the suggestions; I intend to rebuild the filter using a diecast box but will line the inside with copper sheet;hopefully I will see a decrease in loss. Power at 23 cms is expensive and I have no wish to throw away half of it. As the proud owner of a VNA (not a real one but one of the cheap Chinese jobs) I can measure S11 and it reads about 20dB which is quite reasonable. I will forget the idea of threading the rods and stick to drilling and tapping but I have found that ,using hand tools, it is quite difficult getting a surface which is entirely flat and at right angles to the rod.
On another matter entirely I tried to apply for a NOV for 2 metres but was informed that Ofcom are thinking about it.
Gordon
On another matter entirely I tried to apply for a NOV for 2 metres but was informed that Ofcom are thinking about it.
Gordon
Re: 23cm Filter
Gordon,
If you want me to prepare some rods in the lathe for you, that can be done.
Ken
If you want me to prepare some rods in the lathe for you, that can be done.
Ken
Re: 23cm Filter
Hi ken,
That is a very kind offer and if I cannot get the sort of performance which I am looking for I might take you up on it. I have been looking into the relative conductivity of various metals; I had assumed,incorrectly , that Brass was almost as good as copper but that does not appear to be true.
Gordon
That is a very kind offer and if I cannot get the sort of performance which I am looking for I might take you up on it. I have been looking into the relative conductivity of various metals; I had assumed,incorrectly , that Brass was almost as good as copper but that does not appear to be true.
Gordon