Filter hacking for 2395 ISS TV

S-Band DVB-S from the International Space Station
radiogareth
Posts: 1421
Joined: Wed Jan 06, 2016 9:46 am

Re: Filter hacking for 2395 ISS TV

Post by radiogareth » Fri Nov 14, 2025 7:12 am

So a quick search on google to try and understand my NanoVNA (This I found helpful https://www.youtube.com/watch?v=S8MCcY6dHZ4) and some time tweaking this morning produced the following.
The insertion loss matches the filter designers figures (0.7dB) and I THINK the yellow trace shows return loss(5dB/div), the blue the passband, the green a Smith Chart (?? AFAIK 'in the middle' is good for 50R) and the magenta looks like phase/resonance??
Anything I should try to change?
Gareth (still learning ;-))
Nano.jpg
Nano.jpg (246.06 KiB) Viewed 125 times

g0mjw
Posts: 2689
Joined: Sat Sep 20, 2014 9:15 am

Re: Filter hacking for 2395 ISS TV

Post by g0mjw » Fri Nov 14, 2025 10:16 am

All good - it looks properly tuned - but probably best to not display all the traces once. You only really need the blue and yellow. Otherwise it gets confusing.

The blue is S21 - the transmission through the filter received at port 2 from port 1. Seems you have a loss of about 0.7dB That's at 2397MHz which is very close to the ISS frequency. It is also the centre of the pass band. You need to decide if that's where you want it centred, which will probably depend on what side of the wanted band the interference is strongest. It if is coming from 5G at 2390MHz, you might want to offset the pass-band a little higher. Conversely if it is coming from the Wifi, the opposite. If in doubt, leave it where it is. You want as little ripple on the flat part as possible, but there will usually be some. There is a trade off between ripple and how steep the cut off is.

The yellow trace S11 is the return loss, what port 1 gets back as a reflection of what it sent. This needs to be as little as possible to have a good match in band, but it won't be perfect and -17 dB will be good enough. That's 1/50th of the power reflected. Outside the pass band the power should be almost all reflected. That's how a filter works. You will find out that you can get a very good match at some frequencies but anything better than 10 dB is likely to be good enough for most cases. It's equivalent to a 2:1 VSWR, which most PAs can cope with.

The purple trace is the phase. That looks right. In the pass band you want it to be changing linearly with frequency, (i.e. a straight line, not a curved line) which it is. OK in the middle it looks like a big transient but that's just an artefact going round the circle. These filters tend to be less linear near the edges of the pass band. I am not sure if you are looking at the transmitted or reflected phase response, I think it is reflected. I will have to find my NanoVNA to be sure. You might also want to look at the group delay.

The green trace is the smith chart. It shows the impedance of the filter. The chart is showing the resistive and reactive part of the impedance. The centre of the chart is 50 ohms resistive, (assuming you are using 50 ohms). Horizontal line through the centre represents the resistance from 0 ohms (short circuit) to infinite (open circuit). If the measurement is not on that line, it has a reactive element. If the measurement is above the horizontal axis, the match is inductive - i.e. it looks like a resistor with an inductor, if it is below that line it is capacitive. That's not much help here but it would be useful information if you were say tuning an antenna. Now is probably not the time to explain Smith charts. This is great resource. https://www.microwaves101.com/encyclope ... art-basics

G3YKI
Posts: 25
Joined: Tue Jul 19, 2022 3:07 pm

Re: Filter hacking for 2395 ISS TV

Post by G3YKI » Fri Nov 14, 2025 10:26 am

radiogareth wrote:
Fri Nov 14, 2025 7:12 am
So a quick search on google to try and understand my NanoVNA (This I found helpful https://www.youtube.com/watch?v=S8MCcY6dHZ4) and some time tweaking this morning produced the following.
The insertion loss matches the filter designers figures (0.7dB) and I THINK the yellow trace shows return loss(5dB/div), the blue the passband, the green a Smith Chart (?? AFAIK 'in the middle' is good for 50R) and the magenta looks like phase/resonance??
Anything I should try to change?
Gareth (still learning ;-))
Nano.jpg
Very good Gareth.
My saying when it comes to filters is "symmetry is your friend". The symmetry is most clearly illustrated in the green (reflection coefficient) trace. It Shows the resonators are correctly tuned. In textbook case it would look something like this:
export.png
export.png (8.18 KiB) Viewed 100 times
The line passes exactly through the centre of the chart (zero reflection) three times corresponding to the three notches that would be seen on the return loss v frequency plot. I am not suggesting you need to do any more to your filter; for practical purposes it is as good as it will get. The comment is just for learning purposes! Departure from the the theoretical will be due to inevitable slight departure from perfectly symmetrical construction.

radiogareth
Posts: 1421
Joined: Wed Jan 06, 2016 9:46 am

Re: Filter hacking for 2395 ISS TV

Post by radiogareth » Fri Nov 14, 2025 11:35 am

Thanks both for the explanations/ clarifications :-)
In other good news I've just replaced the input device on my DDK 13cm pre-amp and its now functional. A tad over 40dB of low noise gain!!
Think I can stand some filter loss with that!!
Gareth

g0mjw
Posts: 2689
Joined: Sat Sep 20, 2014 9:15 am

Re: Filter hacking for 2395 ISS TV

Post by g0mjw » Fri Nov 14, 2025 2:59 pm

My aluminium tube arrived today.

Mike

Post Reply

Return to “ARISS HamTV”