So this weekend I attempted a high power low pass filter for my 2m amp project. I followed the schematic of W6PQL but it failed. I came to the conclusion that there was some coupling of the board that I made to the ground plane and after removing the ground plane that was below the inductors it started to resemble a low pass filter on the VNA. I was able to get it close with an input SWR of 1.2 but the insertion loss is pretty high. Think I will make another board and give it a go. I have some small Air Variable capacitors here that I am thinking I can use to get the capacitance to the correct value then order some high power silver Mica capacitors.

I originally went with 1kv surface mount caps, but I used an application and found that at resonance the caps would only see around 300 volts. so I think if I go with 500v rated caps I may be ok. I am planning on running this setup at a max of around 1.2kw. I do not claim to know everything, so figured I would ask and see if anyone more knowledgeable on this subject has any pointers. I would post pictures of the original design but it looks like garbage right now after soldering on numerous sets of inductors on it.

I have no particular skill at doing that, so I have no real wisdom to impart, and you can probably ignore the following, but...

Does the W6PQL design suggest using special Rogers PCB substrates or is FR4 OK? A photo that I've seen that is evidently a W6PQL 2m LPF board (at http://www.g4ztr.co.uk/amateur-radio/construction/2m-low-pass-filter/) seems to show a light or white board surface, which makes me wonder whether it may be using some kind of Rogers material. I have a few sheets of different Rogers PCB materials, but I have not yet tried using them, thinking that I may want to use them for future high UHF and microwave projects.

That photo also shows the use of chip capacitors (perhaps from Johanson?). Are you using chip capacitors or capacitors with wire leads? I guess that you mentioned using surface-mount caps (and maybe trying some small air variable caps) - are they low-inductance ones specifically geared toward rf? I think that PQL calls them 'metal mica capacitors' in this description (but I haven't looked up the parts in his BOM):

https://www.w6pql.com/filter_+_coupler_+_detectors.htm

Is your board from PQL (including from a kit) or is it home-made? Does your board resemble the one in either of those articles?

Some Johanson links of possible interest (but I have not checked extensively for appropriate voltage ratings):

https://www.johansontechnology.com/rf-capacitors

https://www.johansontechnology.com/understanding-chip-capacitors

https://www.johansontechnology.com/rf-capacitor-current-power

https://www.johansondielectrics.com/high-voltage-mlc-capacitors

https://www.johansontechnology.com/srf-prf-for-rf-capacitors

It is possible that the capacitors may exhibit series or parallel resonances that may be causing your problems, for example.


ETA: Here it mentions using teflon dielectric for the high-power LPF boards (and the photo appears to show multiple SMD caps in parallel, as opposed to the G4ZTR article showing the chip capacitors that I was thinking of - the board in the G4ZTR article has a 2012 date on it):

https://www.w6pql.com/parts_i_can_provide.htm#2m

Yea the chip capacitors were designed for RF in the VHF UHF frequencies. Its the board material. I found some of the issues today with the values of inductors. I got it tuned for the resonant frequency but the roll off above 2m was not that great. I have built them for HF but you see issues as you get into the VHF/UHF/SHF frequencies. I got the inductor values right and used some RG-400 and built out my own caps. It put the frequency spot on with relatively low SWR. But the 2nd and 3rd fundamental was only like 12db down with 2db of insertion loss. Board is now in the trash.

So I will either go with a coaxial stub filter system or buy the high power W6PQL board and be done with it. I have some more RG-400 and some BNC connectors so I may try to make the coaxial stub filter. Looks like I will be pushing the limit of the BNC connectors but for testing I think it will work. I really need to get a spectrum analyzer.