Cavity Filters

[Nathaniel Butts]

Howdy Folks!

Hope all had a great Christmas holiday and celebrate jovially the New Year.  

The next Chinese New Year will be the Year of the Dragon, symbolized by Antares.  The Dragon is noble, honest, and prosperous.

My scientific question:  I have been researching alternative filter methods (alternative to SAW filters) in an effort to improve the cleanliness of my data, and provide scalability for future needs.  Sometimes I see spurious results that I am wondering if they are related to temperature or RFI.  I also wouldn't mind narrowing the bandwidth a bit.

Plus there's the 'cool factor'.....lol.

I am thinking of building a cavity filter, such as in the link below.  I'm looking to finish tuning my antenna soon so I can begin another full sky survey with the new equipment.

https://lea.hamradio.si/~s53mv/cavity/cavity.html

My question is: when using a 1m dish, would the benefits of a cavity filter be realized, particularly in an RF heavy zone like my neighborhood, when compared against  the Nooelec SAWBird+ H1?

Thanks,

Nathan Butts

Wannabe Astronomer

Bowling Green, KY

[C Bird]

Cavity filters are great when tuned correctly.

I've used them in broadcast TV to filter out very high power adjacent stations with great success.
You will probably need a VNA, a NanoVNA is good enough to tune filter and antenna.

I've used this site to design a few filters for Inmarsat and Iridium L band.
https://www.changpuak.ch/electronics/interdigital_bandpass_filter_designer.php

--
--
You received this message because you are subscribed to the Google
Groups "Society of Amateur Radio Astronomers" group.
To post to this group, send email to sara...@googlegroups.com
To unsubscribe from this group, send email to
sara-list-...@googlegroups.com
For more options, visit this group at
http://groups.google.com/group/sara-list?hl=en
---
You received this message because you are subscribed to the Google Groups "Society of Amateur Radio Astronomers" group.
To unsubscribe from this group and stop receiving emails from it, send an email to sara-list+...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/sara-list/f382434b-1e2a-4cf8-a31d-484ae909ff38n%40googlegroups.com.

--

.

[Marcus D. Leech]

It depends on how you're processing the signals in the back-end. For
example, if you're just doing total-power over the
 entire 65MHz bandwidth from the LNA, then another layer of filtering
after the LNA could help.

IF there are very strong signals within the 65MHz bandwidth, they could
cause issues with your receiver, even if you're
  only "paying attention" to the 21cm band itself.

If you're in a heavily "polluted" RFI area, you may need filtering *in
front* of your LNA, which requires a special low-loss
  filter, which means fairly wide-band.

[Nathaniel Butts]

I have a NanoVNA, and enough knowledge to use 1% of its capabilities, lol.  Still might give it a shot.

I'm starting to wonder if I'm not overloading the filter and causing these issues, then interpreting it as noise.....I don't see the same interference on my pyramid horn, and it's right next to the 1m.  My current noise traces relatively nicely with the galactic h1, almost like a harmonic.

I currently have antenna --> LaNA <--> Bias T --> Sawbird H1+ --> DC Block --> SDR

Guess I'll need to play around a bit.  Don't really need the LaNA as I'm just using 6-7m of RG316.  Was hoping to give my little antenna a better chance.  I do still like the idea of a more analog filter, though.

Thanks,

Nathan Butts

Wannabe Astronomer

Bowling Green, KY

[Marko Cebokli]

Hello Nathaniel!

Cavity versus SAW:

SAW filters can be very narrow and have very steep flanks, almost "brick wall" style, so if a (not too strong) interferer is close to your desired signal, a SAW filter might be the solution.

But SAW filters have considerable losses, and can not be used ahead of the LNA. Also, the far away attenuation is pretty miserable, at most 40dB, and it does not get better as you go further from their center frequency, even by hundreds of MHz, so strong signals, like transponders and DME from nearby aircraft can penetrate them.

Cavity filters have lower loss if reasonably wide, and their out of band attenuation tends to keep going lower the further you go, down to 100dB or more, if well made. Coaxial cavities will usually have a strong spurious response at triple frequency, but a simple lowpass can take care of that. So if you have very strong interfering signals, not too close in frequency, a cavity filter might save your day.

1dB less sensitivity versus totally swamped receiver, is not a bad bargain, so putting the filter before the LNA should not be an anathema.

Marko Cebokli

2023-12-27 15:40, je Nathaniel Butts napisal

[Marcus D. Leech]

On 27/12/2023 14:11, Marko Cebokli wrote:

Hello Nathaniel!

Cavity versus SAW:

SAW filters can be very narrow and have very steep flanks, almost "brick wall" style, so if a (not too strong) interferer is close to your desired signal, a SAW filter might be the solution.

But SAW filters have considerable losses, and can not be used ahead of the LNA. Also, the far away attenuation is pretty miserable, at most 40dB, and it does not get better as you go further from their center frequency, even by hundreds of MHz, so strong signals, like transponders and DME from nearby aircraft can penetrate them.

Cavity filters have lower loss if reasonably wide, and their out of band attenuation tends to keep going lower the further you go, down to 100dB or more, if well made. Coaxial cavities will usually have a strong spurious response at triple frequency, but a simple lowpass can take care of that. So if you have very strong interfering signals, not too close in frequency, a cavity filter might save your day.

1dB less sensitivity versus totally swamped receiver, is not a bad bargain, so putting the filter before the LNA should not be an anathema.

Marko Cebokli

Our "pre-filter" is ahead of the LNAs to keep some strong far-out-of-band signals from causing us non-linearity issues at the
  first gain stage.  Built for us by PE1RKI, it has very low loss (below 0.25db!) but is 200MHz wide at the 3dB points.

To view this discussion on the web visit https://groups.google.com/d/msgid/sara-list/64fd83431613bdbe5a8a14049f99333e%40hamradio.si.

[Alex P]

The LaNA may be  overdriving the SAWBird  with out of band RF from more  'spillover'  pickup in the  '1m'  vs horn.

>>>  Suggestion : Remove the LaNA from the chain ..

Alex

===========================================

Nathaniel Butts

1:42 PM (1 hour ago) 

to Society of Amateur Radio Astronomers

[Marcus D. Leech]

On 27/12/2023 14:32, 'Alex P' via Society of Amateur Radio Astronomers wrote:

The LaNA may be  overdriving the SAWBird  with out of band RF from more  'spillover'  pickup in the  '1m'  vs horn.

>>>  Suggestion : Remove the LaNA from the chain ..

Alex

Agreed.  You absolutely, positively, don't need that LaNA in front of the SawBird.  We use a single SawBird+ H1 at our observatory,
  driving about 12m of cable.   We have no issues observing with this device.

That LaNA is wide-as-a-barn-door, and is almost *certain* to be over-driving the first stage of the SawBird+ H1.

===========================================

Nathaniel Butts

1:42 PM (1 hour ago) 

to Society of Amateur Radio Astronomers

I have a NanoVNA, and enough knowledge to use 1% of its capabilities, lol.  Still might give it a shot.

I'm starting to wonder if I'm not overloading the filter and causing these issues, then interpreting it as noise.....I don't see the same interference on my pyramid horn, and it's right next to the 1m.  My current noise traces relatively nicely with the galactic h1, almost like a harmonic.

I currently have antenna --> LaNA <--> Bias T --> Sawbird H1+ --> DC Block --> SDR

--
--
You received this message because you are subscribed to the Google
Groups "Society of Amateur Radio Astronomers" group.
To post to this group, send email to sara...@googlegroups.com
To unsubscribe from this group, send email to
sara-list-...@googlegroups.com
For more options, visit this group at
http://groups.google.com/group/sara-list?hl=en
---
You received this message because you are subscribed to the Google Groups "Society of Amateur Radio Astronomers" group.
To unsubscribe from this group and stop receiving emails from it, send an email to sara-list+...@googlegroups.com.

To view this discussion on the web visit https://groups.google.com/d/msgid/sara-list/b7b3a96e-3d2e-4eef-8144-c54d703d9f50n%40googlegroups.com.

[Nathaniel Butts]

Yeah, after looking at the data a little bit closer I think that's probably the issue.  I'll rearrange it tonight when I get home.

Thanks for helping a wannabe out!

Thanks,

Nathan Butts

Wannabe Astronomer

Bowling Green, KY

You received this message because you are subscribed to a topic in the Google Groups "Society of Amateur Radio Astronomers" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/sara-list/ek16mlZjy9U/unsubscribe.
To unsubscribe from this group and all its topics, send an email to sara-list+...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/sara-list/fda534b0-4946-4231-81e1-72d9e60f61ec%40gmail.com.

[Nathaniel Butts]

Yep.

 

The weird rf is gone in the doppler graph.

Raw antenna average power is much cleaner as well.  I'm actually getting better power measurements from only using 1 LNA. 

Lesson learned.

Thanks,

Nathan Butts

Wannabe Astronomer

Bowling Green, KY

[b alex pettit jr]

1)  Non-auto-scaled data is beneficial for comparing/adjusting/improving the performance of your system,

             ( dB relative to Cold Sky ) & Plot via Excel , etc.

2) 18.25hrs*60mins_per_hr / 2489samples = 26 second avgs. ..   4 minute averages will improve S/N

Alex

====================================================

On Thursday, December 28, 2023 at 01:57:39 PM EST, Nathaniel Butts <nathani...@gmail.com> wrote:

Yep.

 

The weird rf is gone in the doppler graph.

Raw antenna average power is much cleaner as well.  I'm actually getting better power measurements from only using 1 LNA. 

Lesson learned.

Thanks,

Nathan Butts

Wannabe Astronomer

Bowling Green, KY

.

[Douglas Decker]

I wanted to respond to Nathan, but have not had the time. So maybe a day late.

Here goes. 

I have had a lot of years in RF and Telecommunications. Even with 50 years experience I still feel like am Armature. So here goes.

I tied to insert some documents but it will only let me add them. 

So the fist document is a block diagram of my current system. A 1 meter dish.

The new system in the second document is how I intend on putting it together. It will be close to a 2 meter dish.

Receivers can be desenced by any close or adjacent RF signal. 

So the elimination of the unwanted signals needs to be done before the first RF amplifier in a receiver or before a Preamplifier or LNA.

in the third document some cut and paste on the different kinds of filters.

I have used a lot of cavities building RF Repeaters. 

Cavities isolate the transmitter from the receiver so that one antenna can be used. 

They also are used to block adjacent transmitters and other signals.

The secret is a thing called the "Q" factor or the selectivity of the filter. The higher the Q the narrower  the filter.

So Take a look at the Picture and the Filters document and see if any of this helps a day late but hey better then never.

Doug  K5WMT

[Douglas Decker]

forgot the attachments.   Here they are.

I am a victim of C R S...

[Marcus D. Leech]

On 29/12/2023 14:22, Douglas Decker wrote:

Receivers can be desenced by any close or adjacent RF signal. 

So the elimination of the unwanted signals needs to be done before the first RF amplifier in a receiver or before a Preamplifier or LNA.

In general, the mantra is "filter early, filter often".  BUT, when dealing with weak-signals, or radio astronomy, any loss
  in front of the first LNA/Pre-amp can be somewhat "fatal", because it adds directly to the overall noise figure of the system.
  So if you have a sexy super-low-noise LNA with a NF of 0.1dB, and then "prefix" it with a filter with a 0.8dB loss, your Tsys
  goes up a LOT.   Which is why, IN GENERAL, we avoid putting filters IN FRONT of the first LNA.  There are specific situations
  where it's unavoidable, and you generally in those cases select filters that are as low-loss as practical.  This often means
  the resulting filter is wide-band, because, in general, the narrower the filter, the higher the loss.

Professional radio observatories *very rarely* have filters in front of the first LNA, because it would take their lovely
  low-noise, cryogenic amplifiers from a noise temperature of let's say 6K to perhaps an order of magnitude worse.
  Which is why they locate their observatories far away from "civilization".

In back-yard radio astronomy, it may be unavoidable that there are strong signals outside your notional pass-band that
  will "de-sense" your first gain stage--pushing hard on your p1dB limits, etc.    Fortunately, to a certain extent, the
  other "radio using industries" have similar problems, and over the last decade, LNAs have emerged that have both
  more-than-adequate noise figures AND high linearity (which means that they are less sensitive to "de-sense").  It can also
  help to choose feed architectures that have a strong filtering component to them, which means that the strong out-of-band
  signals will be attenuated by 10-20dB or even more.  The goal at the first stage isn't to eliminate them entirely, but just
  to keep the first stage operating linearly.  After the first stage, you can aggressively filter, and there's almost no
  Tsys penalty for doing so.

in the third document some cut and paste on the different kinds of filters.

I have used a lot of cavities building RF Repeaters. 

Cavities isolate the transmitter from the receiver so that one antenna can be used. 

They also are used to block adjacent transmitters and other signals.

The secret is a thing called the "Q" factor or the selectivity of the filter. The higher the Q the narrower  the filter.

So Take a look at the Picture and the Filters document and see if any of this helps a day late but hey better then never.

Doug  K5WMT

On Thursday, December 28, 2023 at 2:26:17 PM UTC-6 b alex pettit jr wrote:

1)  Non-auto-scaled data is beneficial for comparing/adjusting/improving the performance of your system,

             ( dB relative to Cold Sky ) & Plot via Excel , etc.

2) 18.25hrs*60mins_per_hr / 2489samples = 26 second avgs. ..   4 minute averages will improve S/N

Alex

====================================================

On Thursday, December 28, 2023 at 01:57:39 PM EST, Nathaniel Butts <nathani...@gmail.com> wrote:

Yep.

 

The weird rf is gone in the doppler graph.

Raw antenna average power is much cleaner as well.  I'm actually getting better power measurements from only using 1 LNA. 

Lesson learned.

Thanks,

Nathan Butts

Wannabe Astronomer

Bowling Green, KY

.

--

--
You received this message because you are subscribed to the Google
Groups "Society of Amateur Radio Astronomers" group.
To post to this group, send email to sara...@googlegroups.com
To unsubscribe from this group, send email to
sara-list-...@googlegroups.com
For more options, visit this group at
http://groups.google.com/group/sara-list?hl=en
---
You received this message because you are subscribed to the Google Groups "Society of Amateur Radio Astronomers" group.
To unsubscribe from this group and stop receiving emails from it, send an email to sara-list+...@googlegroups.com.

To view this discussion on the web visit https://groups.google.com/d/msgid/sara-list/e0f8b23b-a0f8-45e1-8d51-840409191aa3n%40googlegroups.com.

[Douglas Decker]

I will agree.  Where I live , In a City\Urban area sometimes the fm stereo and the tv stations are over whelming near here. The dish and its set up helps to block some out. The gain and beam width of the 1 meter dish is low. A 2 meter dish should do better and will have a different feedhorn. All good information. Surprised I do as well as I do with what I have and where I am at...

Tks.

To view this discussion on the web visit https://groups.google.com/d/msgid/sara-list/17242f10-cfdc-457e-b7c4-e95aa1cafe2e%40gmail.com.

[b alex pettit jr]

Hello Douglas,

Instead of  a second SAWbird,    consider upgrading the SDR

Alex KK4VB

1m system @  Dec+40

==================================================

[Martin Pepe]

Team, 

How do these types of filters work on harmonics ? It’s been my experience that they tend to pass even harmonics. Any one have some extended freq. range ( out of passband) testing results ?

Sincerely,

Martin Pepe

 Cloudy skies ? Switch to a longer wavelength! 

Sent from my iPhone

On Dec 27, 2023, at 2:11 PM, Marko Cebokli <s57...@hamradio.si> wrote:



To view this discussion on the web visit https://groups.google.com/d/msgid/sara-list/64fd83431613bdbe5a8a14049f99333e%40hamradio.si.

[b alex pettit jr]

Attenuation prior to first Amplification Stage :

Unfortunately, it is not just   " 1dB less sensitivity ".

=================================================

H Line data is in the 5K - 60K range ..

===============================================================

With a Low Noise Figure  LNA, of say 0.2 dB, the equivalent system noise temperature is  13.5K

But, add a Filter with  '1dB of attenuation'  such that the input noise figure is now

1.2 dB and you have an equivalent system noise temperature of 92.5 K  .

Alex KK4VB

============================================================

[Richard Flagg]

Perhaps one should include a few tens of degrees of spillover temperature in this calculation.

Also I believe that a good bandpass filter can be obtained with perhaps 0.3 or 0.4 dB of attenuation.

If you are stuck in a noisy urban environment (RFI Hell)  you may have no choice but to put a filter in front of the LNA to prevent overload and desense.

Richard

--
--
You received this message because you are subscribed to the Google
Groups "Society of Amateur Radio Astronomers" group.
To post to this group, send email to sara...@googlegroups.com
To unsubscribe from this group, send email to
sara-list-...@googlegroups.com
For more options, visit this group at
http://groups.google.com/group/sara-list?hl=en
---
You received this message because you are subscribed to the Google Groups "Society of Amateur Radio Astronomers" group.
To unsubscribe from this group and stop receiving emails from it, send an email to sara-list+...@googlegroups.com.

To view this discussion on the web visit https://groups.google.com/d/msgid/sara-list/647855699.3809151.1703973082345%40mail.yahoo.com.

[b alex pettit jr]

A tuned cylindrical waveguide ( Cantenna ) provides ~ 40dB atten at CellTower freqs with minimal effect at 1.42GHz.

Alex

========================================================

[Marcus D. Leech]

On 30/12/2023 17:51, Richard Flagg wrote:

Perhaps one should include a few tens of degrees of spillover temperature in this calculation.

Depends on your illumination profile, which can be hard for amateurs to get good data on, for their particular setup.
   But, yes, when estimating Tsys, I usually throw in about 20-30K for spillover, although with more-aggressive edge
   taper, you can get down to 10K or better.

Paul Shuch wrote a good article on this a couple of decades ago about optimizing your dish antenna in both the
  gain and spillover dimensions, and showed that increasing the edge-taper has significant benefits compared to
  a couple more dB total antenna gain.

Also I believe that a good bandpass filter can be obtained with perhaps 0.3 or 0.4 dB of attenuation.

Ours (200MHz wide) is under 0.2dB loss.

If you are stuck in a noisy urban environment (RFI Hell)  you may have no choice but to put a filter in front of the LNA to prevent overload and desense.

We have a cell tower about 1km from the observatory.   We didn't have enough mechanical clearance to put in a "can"
  feed, so using an air-spaced patch instead.   Patch antennas don't generally have the nice steep low-frequency
  cut-off compared to "cans".

To view this discussion on the web visit https://groups.google.com/d/msgid/sara-list/8531dbb8-69f4-479d-b420-78dbfe3f8da2%40hawaii.rr.com.

[Marcus D. Leech]

On 30/12/2023 18:48, 'b alex pettit jr' via Society of Amateur Radio Astronomers wrote:

A tuned cylindrical waveguide ( Cantenna ) provides ~ 40dB atten at CellTower freqs with minimal effect at 1.42GHz.

One has to keep in mind that while waveguide feeds *DO* have a quite aggressive cut-off below the design
  frequency, their filtering performance above the notional design frequency isn't nearly as impressive.

Still, whenever possible, I recommend them.

[b alex pettit jr]

Agreed. The intent was to illustrate that atten prior to the first gain stage is more significant than solely its  "dB loss"  value.

Spillover noise can be large  .

Originally, long sides = f/D 0.35  &  narrow sides =  f/D 0.6

Overall effective f/D now =   0.27

Cheers,

Alex

=====================================================================

On Saturday, December 30, 2023 at 05:51:44 PM EST, Richard Flagg <r...@hawaii.rr.com> wrote:

Perhaps one should include a few tens of degrees of spillover temperature in this calculation.

Also I believe that a good bandpass filter can be obtained with perhaps 0.3 or 0.4 dB of attenuation.

If you are stuck in a noisy urban environment (RFI Hell)  you may have no choice but to put a filter in front of the LNA to prevent overload and desense.

Richard

========================================================================

On 12/30/2023 11:51 AM, 'b alex pettit jr' via Society of Amateur Radio Astronomers wrote: