Eduard Mol on one of the SARA meetings/online videos shows plot of W3ORS5 of water maser at 22 GHz. The plot shows LSR velocity km/s from -70 to -20 on x-axis. How does this convert into necessary bandwidth in frequency? I am ordering a cavity filter and it needs to have wide enough bandwidth.
For a water maser at 22.235 GHz (the H2O 6(16)-5(23) transition),
the conversion from LSR velocity to frequency bandwidth uses:
Delta f / f = v / c
where:
f = 22.23508 GHz
c = 299792 km/s
v = velocity relative to the Local Standard of Rest (LSR)
At 22 GHz, 1 km/s corresponds to:
Delta f = 22.235e9 * (1 / 299792)
= approximately 74.2 kHz per km/s
The SARA plot for W3 IRS5 shows:
Velocity range: -70 to -20 km/s
This is a range of:
50 km/s
Therefore the frequency coverage required is:
50 * 74.2 kHz = 3710 kHz
or approximately:
3.7 MHz
Minimum cavity filter bandwidth:
About 4 MHz
However, in practice, do not order a filter this narrow because
of tuning errors, frequency drift, and uncertainty in the source velocity.
Recommended cavity filter bandwidths:
5 MHz - probably adequate if accurately centred
10 MHz - recommended practical choice
20 MHz - very comfortable margin
The W3 IRS5 maser emission is around -40 km/s LSR.
The frequency shift is:
-40 * 74.2 kHz = -2968 kHz
or approximately:
-2.97 MHz
So the maser line is approximately at:
22235.08 MHz - 2.97 MHz
= 22232.11 MHz
A 10 MHz filter centred at 22.235 GHz would cover:
10 MHz / 74.2 kHz per km/s
= about 135 km/s
which is much wider than the W3 IRS5 velocity range.
For a practical 22 GHz amateur radio astronomy receiver,
a 10 MHz cavity filter centred on 22.235 GHz would be a good choice.
When I say a 10 MHz filter centred on 22.23508 GHz, I mean the total passband width is 10 MHz, not 10 MHz either side.
So:
Centre frequency: 22235.08 MHz
Bandwidth: 10 MHz total
Passband edges:
Lower edge: 22230.08 MHz
Upper edge: 22240.08 MHz
This corresponds to approximately:
10 MHz / 74.2 kHz per km/s = 135 km/s total velocity range
or about:
-67.5 km/s to +67.5 km/s around the centre frequency
If the supplier quotes a filter as “10 MHz bandwidth at 22.235 GHz”, this is normally what they mean.
If they instead quote “±10 MHz”, then the total bandwidth is 20 MHz:
Lower edge: 22225.08 MHz
Upper edge: 22245.08 MHz
For W3 IRS5, the maser range of about −70 to −20 km/s LSR corresponds to roughly 3.7 MHz, so a 5 MHz filter could work if centred correctly. A 10 MHz filter gives useful margin for tuning, Doppler corrections, and looking at other maser sources.
Would this 10 MHz provide enough room for all water masers at 22 GHz, and enough room to accommodate velocity shift due to rotation of Milky Way and rotation of Earth around Sun?
A 10 MHz total bandwidth filter at 22.235 GHz will cover many water maser sources, but it is not enough to cover every possible Galactic water maser velocity if you want a “set and forget” survey receiver.
Let’s look at the velocity equivalent:
At 22 GHz:
1 km/s = 74.2 kHz
10 MHz bandwidth = 10,000 kHz / 74.2
= about 135 km/s total
So a 10 MHz filter centred on the rest frequency covers:
22230.08 MHz to 22240.08 MHz
Velocity range:
approximately -67.5 km/s to +67.5 km/s
Galactic rotation effects
The rotation of the Milky Way produces much larger shifts depending on Galactic longitude.
Examples:
- Nearby star-forming regions: often within ±50 km/s
- Inner Galaxy masers: can easily be +100 to +200 km/s LSR
- Galactic centre region: velocities can exceed ±200 km/s
At 22 GHz:
100 km/s = 7.4 MHz
200 km/s = 14.8 MHz
So a 10 MHz filter will miss some inner Galaxy masers.
Earth’s orbital motion
The Earth’s motion around the Sun causes a Doppler shift up to about:
±30 km/s
At 22 GHz:
30 km/s × 74.2 kHz = 2.2 MHz
So this alone can move a source by ±2.2 MHz over the year.
Earth’s rotation
For a ground-based telescope this is much smaller:
up to about ±0.5 km/s
≈ ±37 kHz
(negligible compared with a MHz bandwidth).
Practical recommendation
For a general 22 GHz water maser receiver:
| Filter bandwidth | Velocity coverage | Comment |
|---|---|---|
| 5 MHz | 67 km/s | Too restrictive |
| 10 MHz | 135 km/s | Good for known sources |
| 20 MHz | 270 km/s | Good general-purpose choice |
| 30 MHz | 400 km/s | Better for Galactic surveys |
For your intended use (amateur radio astronomy, looking at sources such as W3 IRS5 and other known masers), I would choose:
20 MHz centred at 22.235 GHz
if the cavity filter cost and insertion loss are acceptable.
A 20 MHz filter gives:
22225.08 MHz to 22245.08 MHz
approximately -135 to +135 km/s LSR
and leaves room for Earth’s orbital velocity correction plus most Galactic star-forming regions.
If your aim is surveying unknown masers across the Milky Way, even 20 MHz may be restrictive; you would ideally want a wider preselector (e.g. 50 MHz) followed by SDR filtering.
Eduard’s plot of W49 water maser covers -20 km/s to 30 km/s. What bandwidth covers that all year regardless of rotation of Earth and Milky Way?
For W49 water maser, Eduard’s velocity range is:
-20 km/s to +30 km/s LSR
That is a span of:
50 km/s
At 22.235 GHz:
1 km/s = 74.2 kHz
50 km/s = 3.71 MHz
So the maser emission itself occupies only about 3.7 MHz.
Now add the motions you want to allow for.
1. Earth’s orbit around the Sun
The maximum barycentric correction is approximately:
+/-30 km/s
At 22 GHz:
30 x 74.2 kHz = 2.23 MHz
So allow:
+/-2.2 MHz
2. Earth’s rotation
For a fixed ground station:
up to about +/-0.5 km/s
which is:
+/-37 kHz
This is negligible compared with MHz bandwidths.
3. Milky Way rotation
This is already included in the quoted LSR velocity of W49. The LSR correction removes the Sun’s motion relative to the local standard of rest, so you do not add Galactic rotation again to the same source.
The important point is: different sources have different LSR velocities, because of Galactic rotation. W49 is around +10 km/s LSR, but another maser could be hundreds of km/s away.
For W49 specifically:
Intrinsic velocity range:
-20 to +30 km/s
= 50 km/s
= 3.7 MHz
Allow Earth orbital motion:
+/-30 km/s
= +/-2.2 MHz
Total required:
3.7 + 4.4 = 8.1 MHz
Therefore:
- 10 MHz filter: just about sufficient for W49 all year, assuming it is centred correctly.
- 15 MHz filter: more comfortable.
- 20 MHz filter: gives plenty of margin.
For W49 I would centre the filter not exactly at 22.23508 GHz, but around the expected observed frequency. The mean velocity is about +5 km/s LSR, so the line is shifted by:
5 x 74.2 kHz = 371 kHz
towards lower frequency (radio convention), giving a centre around:
22,234.7 MHz
A 20 MHz cavity filter centred around 22.235 GHz would be a very good general-purpose choice for W49 and other Galactic water masers. A 10 MHz filter is the minimum I would consider.
Here is a list of bright water masers. Again, what bandwidth will cover all of these, regardless of which time of year I observe them?
G#
G208.993-19.385
G43.167+0.011
G25.649+1.050
G133.949+1.064
G92.670+3.071
G133.715+1.215
G12.681-0.182
G81.877+0.782
G49.488-0.387
G109.871+2.114
G43.795-0.127
G205.109-14.109
G10.623-0.383
G31.243-0.111
G27.184-0.082
G65.780-2.612
G269.270+37.196
G16.868-2.158
G105.370+9.841
G34.257+0.153
G63.115+0.340
G33.145-0.417
G133.694+1.216
G34.403+0.232
G81.711+0.563
G213.705-12.597
G133.749+1.198
G16.927+0.960
G25.826-0.178
G35.131-0.746
G173.481+2.446
G95.053+3.972
G106.797+5.312
G310.357+67.896
G97.312+3.282
G17.638+0.157
G69.540-0.976
G75.773+0.343
G37.821+0.412
G10.472+0.027
G59.783+0.064
G15.033-0.674
G97.525+3.182
G35.197-0.743
G158.040-21.410
G17.551-0.126
G188.796+1.033
G111.542+0.769
G173.853-13.744
G12.209-0.106
G56.370-0.634
G188.946+0.887
G28.862+0.066
G24.790+0.083
G196.454-1.677
G17.016-2.400
G168.063+0.820
G192.599-0.048
G108.595+0.492
G19.609-0.234
G107.300+5.640
G74.036-1.712
G79.977+0.816
G209.015-19.405
G210.062-19.594
G78.887+0.709
G43.237-0.046
G48.606+0.024
G31.412+0.307
G345.698-0.090
G208.898-20.050
G174.198-0.076
G154.310+21.517
G24.942+0.075
G341.218-0.212
G359.9338-17.8541
G94.603-1.796
G333.607-0.215
G105.405+9.877
G78.871+2.762
G43.816-0.117
G203.316+2.055
G0.5464-0.8511
G135.278+2.797
G30.817-0.057
G78.122+3.633
G336.994-0.027
G206.565-16.361
G208.816-19.239
G305.208+0.206
G158.347-20.556
G111.255-0.770
G300.504-0.176
G339.621-0.121
G318.022+32.811
G271.034+18.612
G188.715-2.492
G52.097+1.042
G311.643-0.380
G34.278+69.213
G81.518+0.194
G210.435-19.766
G213.879-11.829
G5.885-0.392
G23.010-0.411
G10.841-2.592
G158.768-21.577
G28.397+0.078
G35.200-1.737
G8.663+22.181
G218.054-0.116
G103.152+23.178
G45.071+0.132
G139.912+0.199
G99.982+4.170
G71.522-0.385
G14.852-0.990
G59.471-0.183
G269.153-1.128
G38.917-15.560
G320.906-0.293
G329.183-0.314
G0.1644-0.4425
G189.778+0.345
G305.363+0.213
G38.300+8.676
G108.471-2.818
G337.404-0.402
G337.419-0.160
G331.132-0.244
G311.230-0.032
G173.580+2.442
G352.158+0.403
G77.928+0.871
G108.426+0.893
G31.280+0.062
G183.721-3.664
G314.408+0.048
G170.657-0.269
G349.722+0.124
G189.053+1.069
G176.517+0.182
G264.289+1.469
G82.583+0.201
G81.897+0.812
G10.632-0.512
G100.377-3.579
G323.740-0.263
G165.472-9.059
G44.310+0.041
G328.307+0.430
G351.581-0.353
G327.402+0.444
G81.767+0.596
G357.967-0.162
G98.856+2.934
G43.124+0.033
G108.485-2.805
G353.814+0.315
G141.081-1.060
G37.554+0.201
G14.166-0.060
G31.581+0.077
G73.652+0.195
G40.494+2.541
G71.313+0.828
G36.388+0.786
G12.889+0.490
G30.547+4.721
G123.066-6.309
G331.278-0.188
G26.110-0.094
G245.438-0.147
G6.566-3.638
G41.952+0.454
G352.894+16.847
G299.013+0.128
G351.446+0.660
G333.127-0.434
G11.936-0.151
G53.184+0.159
G11.499-1.485
G37.498+0.530
G25.382-0.181
G111.618+0.830
G173.719+2.698
G21.880+0.014
G211.567-19.294
G311.719+0.240
G349.799+0.108
G109.854+2.104
G45.465+0.053
G355.130-0.302
G359.6161-0.2480
G35.578-0.030
G80.867+0.420
G30.473+5.101
G13.657-0.600
G229.570+0.150
G23.955+0.151
G32.744-0.076
G278.588+52.479
G111.524+0.801
G117.686+4.104
G48.709-4.778
G78.937-2.717
G31.584+5.377
G337.860+0.271
Data from MaserDB https://maserdb.net/
For this list, the answer is quite different from the W49/W3 examples because you are no longer asking about one source — you are asking for a single fixed cavity filter that can observe any bright Galactic water maser at any time of year.
The key quantity is the LSR velocity range of these masers. The Galactic longitude/latitude names alone do not give the velocity; we need the measured maser velocities from MaserDB. However, from the distribution of these well-known Galactic H2O masers we can estimate the required range.
Contributions to the required bandwidth
At 22.235 GHz:
1 km/s = 74.2 kHz
1. Galactic maser velocities
The bright masers in your list include:
- Outer Galaxy sources (negative LSR velocities, often -100 to -20 km/s)
- Local arm sources near 0 km/s
- Inner Galaxy sources (positive velocities, sometimes +100 to +200 km/s)
A realistic envelope for bright Galactic star-forming water masers is approximately:
LSR velocity:
-150 km/s to +200 km/s
That is:
350 km/s total
Frequency span:
350 × 74.2 kHz
= 26 MHz
2. Earth’s orbital velocity
The Earth’s orbit changes the observed velocity by:
+/-30 km/s
Equivalent bandwidth:
30 × 74.2 kHz
= 2.2 MHz each side
= 4.4 MHz total
3. Earth’s rotation
Maximum effect:
<1 km/s
<75 kHz
Negligible.
Total required bandwidth
A realistic calculation:
Galactic maser velocity range:
350 km/s = 26 MHz
Earth orbit allowance:
60 km/s = 4.5 MHz
Total:
410 km/s = about 30.5 MHz
So:
| Filter bandwidth | Suitability |
|---|---|
| 10 MHz | No – only covers a subset |
| 20 MHz | Covers many, but misses extreme velocities |
| 30 MHz | Probably adequate for your list |
| 40 MHz | Comfortable all-sky Galactic coverage |
| 50 MHz | Very safe survey bandwidth |
My recommendation for your 22 GHz system
If the cavity filter is intended mainly for:
- W49
- W3 IRS5
- Orion KL
- other famous bright masers
then:
20 MHz is sufficient.
If the goal is to use your dish as a survey instrument and observe any source from MaserDB, I would choose:
40 MHz bandwidth centred at 22.235 GHz.
That gives:
22.215 GHz to 22.255 GHz
Velocity coverage:
approximately +/-270 km/s
which covers essentially all known Galactic star-forming masers and leaves room for the Earth’s annual Doppler shift.
One further point: at 22 GHz the cavity filter does not need to be extremely narrow because the SDR (for example an SDRplay/ADALM-Pluto/USRP) can do the final channel filtering. The cavity filter’s main job is usually front-end protection from out-of-band signals and improving LNA dynamic range, so a wider, lower-loss filter is often preferable.