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.
I added hairpin probes, lengthened the coil and grounded one end. The results are pretty good other than the insertion loss.
I initially set up the filter with a eBay 'noise source' which has three stages of amplification after the actual noise source - so pretty high level. I measured about a -3dB loss which I felt was, for the passband (750kHz) entirely acceptable.
However, when doing further tests using much lower level signals such as might be arriving at an aerial, I found the loss was nearer 12dB! Yikes!
So, some questions....
1. Will any filter with such a narrow passband have a high insertion loss? (My understanding is yes, it will...)
2. Is there anything I can do to improve (reduce) the insertion loss? Its coil is silver-plated copper, the probes are just tinned copper.
3. Does a high level input cause some 'filter bypass' which might explain why my earlier tests showed only -3dB insertion loss.
4. I'm happy to build and machine stuff - does anyone have a favourite 'go-to' 70cms BPF design?
5. The pre-amps I'm using are small SMA-ended eBay 0-6GHz 20dB units, which do indeed work adequately at up to 6GHz. The filter loss is therefore reasonably acceptable, but I appreciate that distributed gain and sequential filters may better achieve my objectives.....
Pic shows the actual unit tested. Comments welcome.....
Look for a nice big mechanical filter - a couple of designs were published in CQ-TV, including a very nice one by Ian G3KKD, and summarised here: https://wiki.batc.org.uk/70cms_filters
I would start by tweaking for minimum loss on frequency using the Portsdown signal generator and a narrow band 70cms rig (FM is easiest for tweaking) or the Portsdown power meter and a source.
Always run the signal source at a lower level as possible to avoid leakage - you may need a couple of external attenuators to get the level low enough to use a NB rxr (-120 dBm or lower).
Then check for overall shape with Satsagen and go back to tweaking for minimum insertion loss and then Satsagen etc - you can spend many happy hours doing this to get a good result.
Aim for ~1 dB insertion loss if before the pre-amp and 3 dB loss after the pre-amp but a very good shape.
https://wiki.batc.org.uk/images/5/5d/DA ... t_2021.pdf
The coupling is not ideal in your experimental filter. The loops need to be in the same plane as the resonating element; you can't achieve that with a helical resonator. Even if you could get around that, the enclosure is too small to have a high enough Q to do what you want.
Try to find a commercial UHF cavity filter similar to this: I have one from Procom that's 300mm tall and 200mm diameter. It has adjustable loops that can alter the trade-off between insertion loss and bandwidth. Typically, 1db insertion loss will give a 3dB bandwidth of 500kHz. If 2.5dB insertion loss can be tolerated the bandwidth comes down to 250kHz.
Our repeater group has some boxy ones that are 290mm tall and about 250mm square that you might com across.