What radio astronomy is realistic with a 3.8m dish, elevation-only, urban site?

61 views
Skip to first unread message

Tetsuhiko Kanari

unread,
Jun 4, 2026, 10:16:51 AM (yesterday) Jun 4
to sara...@googlegroups.com

Hi all,

I'm an undergraduate engineering student in Tokyo, and I'm looking into repurposing an old 3.8 m C-band dish on our campus for radio astronomy. Before I get too far into the project, I'd like to get a better idea of what's realistically possible with the constraints of the site and mount.

Current situation:

  • 3.8 m parabolic dish, originally used for C-band communications
  • f/D not yet measured
  • Surface accuracy unknown
  • Azimuth is fixed (pointing roughly southwest), elevation is adjustable only
  • Site is in central Tokyo, Japan, so I expect a fairly challenging RFI environment
  • My initial target is the 1420 MHz HI line

Planned backend:

  • RTL-SDR
  • Raspberry Pi 4 for unattended drift-scan observations

Current RF hardware:

  • S53MV-style cavity combline filter for 1420 MHz completed and tested
  • Feed not yet designed
  • LNA not yet selected

What I'm trying to figure out is:

  1. Besides detecting the HI line, what observations would be realistic with a dish of this size and an elevation-only mount?
  2. Are there any projects that are likely to be impractical from an urban site and not worth pursuing?
  3. Once I determine the dish geometry, are there particular feed designs or LNAs you would recommend for 1420 MHz work?
  4. Has anyone operated a similar transit-only system and found useful observing programs beyond simple HI detection?

My longer-term goal is to see whether the dish can support regular student-operated observations rather than just a one-time demonstration.

I've attached photos of the dish and the filter. Any comments, suggestions, or reality checks would be greatly appreciated.

Thanks,

Tetsu
Tokyo, Japan

92DC05B9-F310-441C-ABB6-D4EB9DA36A76.jpg
0519D68F-8E53-46F8-A98E-0B926D0EA0A2.jpg
6782007C-FE01-49F8-A7EF-2A6B45743835.jpg

Marcus D. Leech

unread,
Jun 4, 2026, 12:52:44 PM (yesterday) Jun 4
to sara...@googlegroups.com
You might find that 21cm is quite noisy in downtown Tokyo.  If the elevation axis is motorized, with precision position feedback, you can turn this into a sky-mapping
  engine.   Sure, start out at 21cm, but my rough guess is that this dish would work in the lower end of Ku-band or upper end of X-band.   An amateur-produced sky-map
  of as much as the sky as you can reasonably get to from there at X or Ku would be a challenging and satisfying project.  The "BUllseye" LNBF that is used for the amateur
  QO-100 payload reception operates out of the main piece of Ku-band, so can be used in that part of the band to do a sky-map.

If it's motorized AND you have good position feedback (the Level-developments USB inclinometer series are good enough for this), then you can use a technique called
  "driving on the meridian" to slowly build up a sky-map.  It sounds like you're offset from the meridian, so you'd need to do a little bit of coordinate transformation on
  the "raw" position data before you could use it in a sky-map.    In this technique, you just drive the elevation axis very slowly between its limits, 24x7, and record data
  tagged with the azimuth/elevation coordinates, and the LMST at the time of the observation.  You then coordinate-transform the data and bin/grid it into a celestial
  coordinate system, recording averaged power values.   Over time, you'll have a map of the sky at your desired wavelength.   


92DC05B9-F310-441C-ABB6-D4EB9DA36A76.jpg
0519D68F-8E53-46F8-A98E-0B926D0EA0A2.jpg
6782007C-FE01-49F8-A7EF-2A6B45743835.jpg
--
--
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 visit https://groups.google.com/d/msgid/sara-list/CAEQ-dRknXvh4GMtj3O%3D1d1NLw%2BxN2CbbbeA9VZzVc4zAe1CwYw%40mail.gmail.com.

Eduard Mol

unread,
Jun 4, 2026, 1:52:34 PM (yesterday) Jun 4
to sara...@googlegroups.com
a 3.8 metre dish can be a quite capable instrument, even without a motorized mount. I have a 3 metre dish at home which is also not motorized, i have to do driftscans for all my observations. Even with this limitation I can detect more than just the hydrogen line in our galaxy. For example last year I did a series of observations on the hydrogen line in the Andromeda galaxy (https://youtu.be/aqUufpIYFYM?si=_j_mgD_MLIhNOhF3). However, as Marcus already mentioned RFI may be a problem in the city, especially when looking for sources that are much weaker than galactic hydrogen. 
Personally I don't think lower Ku band is any good for amateur radio astronomy. First of all most of the "classical" radio sources like Cassiopeia A, Cygnus A and the galactic background are all very weak at Ku band. So you would be limited to observing the Sun and Moon, and maybe the methanol maser in W3(OH) as a distant third. Secondly, in my experience observations at Ku band have become near impossible in recent years due to interference from satellites. 


Stephen Arbogast

unread,
Jun 4, 2026, 4:52:24 PM (yesterday) Jun 4
to Society of Amateur Radio Astronomers
In my humble opinion  ezRA  is a very  good  software suite to  post process  the  data  you have  collected  for  Hydrogen Line... as long as your  raw  data is formatted correctly in a  text file that  ezRA  can handle.  ezRA   will collect the  HI raw data,  ezCol,  then go beyond to  process it with  ezCon. 

Alex P

unread,
Jun 4, 2026, 5:12:04 PM (yesterday) Jun 4
to Society of Amateur Radio Astronomers
The software set       SDR# > IFaverage > HL3D > Rinearn 
can initially be used to quantify the performance and characterize the value of system changes/adjustments 

This is a 54 hour drift scan data set using the above process from a 12 meter Dish system .
NG01_vel_54Hrs_sm.jpg


Alex Pettit
Proj HLine3D

b alex pettit jr

unread,
Jun 4, 2026, 5:38:47 PM (yesterday) Jun 4
to 'Alex P' via Society of Amateur Radio Astronomers
Begin with SDR# > IFaverage > HL3D > Rinearn   for setup, characterization and adjustments of your system 

Inline image
then consider writing your own code using
=============================================
Virgo
or
========================================
 GNURadio


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


Alex Pettit


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




fasleitung3

unread,
3:15 AM (20 hours ago) 3:15 AM
to sara...@googlegroups.com
Hi Tetsu,
Obviously RFI will be a major challenge. One of the things to be concerned is out of band interference which might swamp your LNA and drive it into saturation. One remedy is to use a feed which offers some protection. You could use a "cantenna" to start with. This will not provide optimal illumination of the dish, but it would get you goinig. The advantage of a cantenna is that it has a lower cut-off frequency of around 1 GHz which will nicely attenuate all mobile phone bands below that frequency. It also offeres some protection at higher frequencies.
If that is not sufficient, you could put your filter between the cantenna and the LNA. This will degrade the noise figure but is the solution if out of band signals are overwhelmingly strong.

That should enable you to observe the hydrogen line in transit scans. You could also try to do a transit scan of a strong continuum source. I have not calculated it but I would assume that Cygnus A will transit your beam given your location and orientation of the dish.

Good luck,
Wolfgang
--

Ayushman Tripathi

unread,
3:39 AM (20 hours ago) 3:39 AM
to Society of Amateur Radio Astronomers

Hi Tetsuhiko,

That's a great dish to be starting with! I'm running H-line observations on a 2.4m dish at a remote site, fully automated on a Raspberry Pi 500 24/7. I previously tried observations from a high RFI urban area too.

What I saw: when you are close to a strong RFI source, the front end (LNA) gets overloaded (saturated), and the H-line just gets buried in noise. I also tried putting another SDR in a metal box to shield it, that helped a bit, but the results still were not consistent. The worst one was a solar inverter (not mine, a neighbour's) the moment it turned on around 6–7am, the whole H-line disappeared into the noise floor.

It should still work, the placement just matters a lot, mainly try to keep the dish away from things like monitors, switching power supplies, and solar inverters. It is more about distance from the noisy stuff than the city itself.

My earlier post on the urban RFI issues (it does not cover the 7am solar one):

https://groups.google.com/g/sara-list/c/HWHp3sojNhc/m/YLIzFhkQGgAJ

For other sources besides the H-line, I have not tried these yet, but these are possible:

- OH lines: 1612, 1665, 1667, 1720 MHz

- Pulsars (broadband, ~300–1500 MHz)

For the LNA, I use SAWbird+H1: https://www.amazon.co.jp/dp/B07XPV9RX2

LNA:

Screenshot From 2026-06-05 03-37-52.png

b alex pettit jr

unread,
7:19 AM (16 hours ago) 7:19 AM
to 'fasleitung3' via Society of Amateur Radio Astronomers
Hello Tetsu,

I can confirm Wolfgang's comments on the benefit of a Cantenna ( Tuned Cylindrical Waveguide )
I installed a small radio telescope at a local university and the only way I could remove the intense RFI from a nearby cell-phone tower was
by replacing the Loop feed with a Cantenna.


A guess at your dish .. it might be worth testing to see if a metal mesh screen around the perimeter helps block nearby interference
Inline image
========================================================================

Cantenna Info  


LNA needs to be installed at the Feed if possible .. 
 Passive Attenuation from cables or components adds to overall Tsys
 System Noise lowering sensitivity and this can Not Be Recovered by added post amplification 
Inline image


You can test a version without the choke ring, but it will better shape the beam for a shallow
dish thus helping to reduce blackbody ground noise spillover and perhaps some RFI


Inline image

photo and dimensions of the feed I constructed
Inline image

Response of a 1420 MHz Tuned Cylindrical Waveguide 
Inline image
Inline image


Here are a  Excellent Resources
=========================================================




Natalia K. Nikolova, Notes on Antenna Engineering 2018, McMaster University, Canada


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


            Regards,
           Alex Pettit
Inline image

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

On Friday, June 5, 2026 at 03:15:22 AM EDT, 'fasleitung3' via Society of Amateur Radio Astronomers <sara...@googlegroups.com> wrote:


Hi Tetsu,
Obviously RFI will be a major challenge. One of the things to be concerned is out of band interference which might swamp your LNA and drive it into saturation. 
One remedy is to use a feed which offers some protection.

 You could use a "cantenna" to start with. This will not provide optimal illumination of the dish, but it would get you going. 

Andrew Thornett

unread,
8:21 AM (15 hours ago) 8:21 AM
to sara...@googlegroups.com
Just adding a couple of comments, as someone who enjoys hydrogen line, although I dont have Alex's or Wolfgang's expertise and experience. 

Alex has suggested adding some mesh around the edge of the dish. At the hydrogen line meeting last Monday, one attendee discussed the enormously beneficial effects of doing just this - personally, I have been remiss and not done it on my own dish, in spite of Alex's strong recommendations to me to do so. So now, I have purchased, and took receipt of l, large roll of aluminium mesh yesterday and will be urgently adding this extension myself. Doubtless, Alex will again be shown to be correct - and I will wish I had done it earlier.

With regards to placing the LNA as close to feed horn as possible, I cannot agree more. Again, my personal experience was that even a few metres of coaxial cable between the feed and the LNA/filter dramatically reduced the signal level detected by the SDR. Even using expensive LMR-400 coaxial cable did not solve this issue- but moving the LNA to immediately next to the cantenna feed was magical in its effect!

Andy



From: 'b alex pettit jr' via Society of Amateur Radio Astronomers <sara...@googlegroups.com>
Sent: Friday, 05 June 2026 12:18:40
To: 'fasleitung3' via Society of Amateur Radio Astronomers <sara...@googlegroups.com>
Subject: Re: [SARA] What radio astronomy is realistic with a 3.8m dish, elevation-only, urban site?
 
--
--
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.

B Lowell

unread,
10:17 AM (13 hours ago) 10:17 AM
to sara...@googlegroups.com
For choice of a project and with a good clock (GPS) and a suitably chosen set of radio stars and your antenna fixed in elevation (might the change in elevation be modified from manual to motor/computer drive to increase flexibility) you could time the passage of stars and, perhaps, obtain meaningful measurements of the rotation rate of the earth. Best, Bartley (KD1KG)

Marcus D. Leech

unread,
11:28 AM (12 hours ago) 11:28 AM
to sara...@googlegroups.com
On 2026-06-05 10:10, B Lowell wrote:
For choice of a project and with a good clock (GPS) and a suitably chosen set of radio stars and your antenna fixed in elevation (might the change in elevation be modified from manual to motor/computer drive to increase flexibility) you could time the passage of stars and, perhaps, obtain meaningful measurements of the rotation rate of the earth. Best, Bartley (KD1KG)
Something that our students do during our summer camp is to take a 24-hour drift-scan with the radio telescopes that they build, and then plot a spectrogram in
  heat-map form for those 24 hours.   In many locations in the sky, you'll see what I call the "doppler ghost" of Earth rotation.    The "local bubble" of H1 around us has
  a small-enough doppler distribution that it's almost like a noisy signal generator at 21cm.   What happens is that as we rotate, there's a coupling in doppler space
  between our *rotation* and our *orbit* -- we turn both towards our orbital direction, and then away from it over 24 hours.



This was from one of the student telescopes (small horn antenna).

We've also been engaged in a long-term study of earth-orbit parameters, using both J0332+5434 and a bright H1 source to get precise Doppler measurements once
  a day.   This year, we're observing M8 daily, to measure the absorption feature in M8, and extract Doppler information from it.   Both the pulsar observations and
  now the M8 observations are precise enough that when you curve-fit a multi-parameter doppler curve, and compute the residual, the lunar influence on our orbit
  becomes visible (about 11 m/sec) as a 28-day sub-cycle.

Cheers
Marcus

Dave Typinski

unread,
12:48 PM (11 hours ago) 12:48 PM
to sara...@googlegroups.com
On 6/5/26 11:28, Marcus D. Leech wrote:
>
> Something that our students do during our summer camp is to take a 24-hour
> drift-scan with the radio telescopes that they build, and then plot a spectrogram in
> heat-map form for those 24 hours. In many locations in the sky, you'll see
> what I call the "doppler ghost" of Earth rotation. The "local bubble" of H1
> around us has
> a small-enough doppler distribution that it's almost like a noisy signal
> generator at 21cm. What happens is that as we rotate, there's a coupling in
> doppler space
> between our *rotation* and our *orbit* -- we turn both towards our orbital
> direction, and then away from it over 24 hours.
>
> We've also been engaged in a long-term study of earth-orbit parameters, using
> both J0332+5434 and a bright H1 source to get precise Doppler measurements once
> a day. This year, we're observing M8 daily, to measure the absorption
> feature in M8, and extract Doppler information from it. Both the pulsar
> observations and
> now the M8 observations are precise enough that when you curve-fit a
> multi-parameter doppler curve, and compute the residual, the lunar influence on
> our orbit
> becomes visible (about 11 m/sec) as a 28-day sub-cycle.

That's *WAY* cool, Marcus!
--
Dave

Reply all
Reply to author
Forward
0 new messages