Is it worth adding a second probe at 90 degrees to my cantenna waveguide on LRO-H2, my Solar Cooker Parabolic 150cm Radio Telescope?

August 28, 2025 Admin

Great question — this gets into radiometer theory and feed design.

If you add a second probe into your WR-650 feed at 1420 MHz, what happens depends on how you use that probe:

1. If the second probe is orthogonal (90° rotated)

You now have a dual-polarization feed (horizontal and vertical, or LHCP/RHCP if phased properly).
Each probe collects independent samples of the incoming noise power from the sky (assuming the source is unpolarized, like the H I line).
If you process both signals separately and then average them in power, the radiometer equation says sensitivity improves by a factor of √2 (≈ 1.41×).
- Equivalent to doubling integration time.

2. If the second probe is co-polarized (same orientation, just another pickup of the same mode)

Then both probes are not independent: they’re looking at the same polarization state of the field.
Their noise outputs will be highly correlated, so averaging them does not give √2 improvement.
At best you just duplicate the same measurement with slightly different noise pickup from the receiver chain — but the fundamental sensitivity won’t improve much.

3. Practical considerations

You’d need two LNAs, two receiver chains, and careful calibration.
The magic improvement only comes if the probes see independent noise modes — which is why dual-polarized feeds are used in big observatories (you get a √2 boost in sensitivity).
If you don’t need polarization information, you can simply average the two total power streams for extra sensitivity.

✅ So:

Yes, you can get √2 better sensitivity if you add a second probe in orthogonal polarization and average both streams.
No, if you just duplicate the same polarization you don’t gain much.

WR-650 Waveguide:

in WR-650 (rectangular) you can’t get two independent polarizations at 1420 MHz just by adding a 2nd probe. Only the TE₁₀ mode propagates there (TE₀₁ cuts on ≈1.82 GHz), so a second probe would still couple to the same mode → no √2 gain.

What to build instead (practical dual-pol you can DIY)

Add a short circular waveguide “can” (single-mode TE₁₁) in front of your WR-650, and put two crossed probes 90° apart in the circular section. Feed each probe with its own LNA/receiver and average the powers for the √2 sensitivity boost on unpolarized HI.

Dimensions at 1420.405 MHz (worked numbers)

Free-space λ ≈ 211.06 mm
Choose circular WG inner diameter D ≈ 135 mm → TE₁₁ cutoff ≈ 1.30 GHz, next modes TM₀₁ ≈ 1.70 GHz, TE₂₁ ≈ 2.16 GHz (so only TE₁₁ propagates at 1420).
Backshort (shorted end) distance to the probe plane: λg/4 ≈ 131 mm in this circular section (guide wavelength λg ≈ 523.8 mm at 1420 with fc=1.30 GHz).
Probe locations: two SMA/N feedthroughs 90° apart around the circumference, both at the same axial plane (the plane that is λg/4 from the short).
Probe length (starting value): ~0.18–0.22 λ ≈ 38–46 mm of inner conductor protruding into the guide (air). Start ~42 mm and trim for match.
Probe radial placement: centerline of each probe should land roughly at the circle’s mid-radius (i.e., the pin tip ends near the electric-field maximum of TE₁₁). Practically: mount the connector flush on the wall; don’t try to “aim” the pin—just straight in.

Stack-up / transitions

WR-650 → circular transition: a short taper or a simple square-to-circular plate set works; keep it ≤ ~0.5 λ long to avoid extra modes.
Circular section (D≈135 mm) of length ~ λg/4 + 10–20 mm margin to the shorting plate.
Shorting plate at the back end (good electrical contact all around).
Two orthogonal probes at the λg/4 plane as above.
Aperture end faces the dish as usual.

RF chain & combining

Treat the two probes as independent polarizations (X & Y). Use two identical LNAs, equal cable lengths, and calibrate gains.
For sensitivity boost (not polarization science), record total-power from each chain and average in power: Pavg=(PX+PY)/2P_{avg}=(P_X+P_Y)/2 → √2 improvement in rms noise.
If you want circular polarization (RHCP/LHCP), put a 90° RF hybrid after the LNAs to form (X±jY)/√2; also handy for isolation checks.

Tuning & test (quick workflow)

With a VNA (or noise diode + SDR), tune one probe at a time:
- Trim the pin length for S11 dip at 1420 MHz.
- Fine-tune backshort distance a few mm if needed (slotted screw holes or shim gaskets help).
Install the second probe 90° around; verify minimal cross-coupling (S21) between probes (aim for ≤ −20 dB; the hybrid can improve effective isolation).
On sky: compare cold sky vs. absorber Y-factor on each arm; verify the √2 radiometer improvement by averaging powers.

If you must stay entirely in WR-650

Use an OMT (orthomode transducer) rather than a second same-mode probe. A Bøifot-style OMT (main guide + two side arms) will split the field into two orthogonal linear ports properly. Building one is more intricate than the circular can with crossed probes, but it keeps everything in rectangular.

Cantenna waveguide:

Yes — in a cantenna (simple circular-waveguide “can”) you can get two independent probes that behave as orthogonal polarizations if the can is sized so only the TE₁₁ mode propagates at 1420.405 MHz. Below is a practical, buildable recipe with numbers, tuning steps and pitfalls.

Basic design numbers (use these as starting targets)

Target frequency: 1420.405 MHz
Chosen TE₁₁ cutoff (fc): ≈ 1.30 GHz (keeps TE₁₁ below 1420 and next modes above it)
Can inner diameter (D): ≈ 135.2 mm (→ single-mode TE₁₁ at 1420 MHz).
Free-space wavelength (λ): ≈ 211.2 mm.
Guide wavelength in the circular section (λg): ≈ 524.2 mm.
Backshort plane (distance from probe plane to short): λg/4 ≈ 131.0 mm.
Can length (practical): λg/4 + 10–30 mm margin → ~141–161 mm overall from probe plane to short.
Probe tip insertion (start): ~0.18–0.22 λ ≈ 38–46 mm measured from the inner wall into the cavity (start ≈ 42 mm, trim to tune).
Probe radial placement: about mid-radius (i.e., tip ~ halfway between center and wall where TE₁₁ E-field is strong).
Probe angular separation: 90° around the circumference (crossed).

Mechanical layout (what to cut/drill)

Use a straight cylindrical can (aluminium or copper) with ID ≈ 135 mm.
Choose a plane inside the can (the probe plane) that will be λg/4 from the shorting plate. Drill two holes for SMA feedthroughs at that plane, separated by 90°. If you mark the can circumference, the two hole centers are opposite each other on the circle quarter-turn apart.
Fit SMA bulkhead/feedthroughs (prefer hermetic or soldered types if available). Use short rigid pin probes soldered to the center conductor of each SMA. The probe body should be insulated from the can.
Short plate: robust metal plate that mates to the can end and provides a solid short. You may want a small adjustable screw to fine-tune axial position by a few mm if possible.

RF chain & electrical notes

Each probe → identical LNA (same gain, noise figure) → equal length coax → receiver/SDR.
Two independent chains give two independent samples for unpolarized sky; averaging power of the two channels gives √2 improvement in sensitivity (≈ 1.414×) compared to a single polarization (assuming low correlation).
Aim for isolation between probes ≤ −20 dB; better is preferable. This is achievable with correct probe placement and a modest hybrid/isolation network if needed.
If you want circular polarization (RHCP/LHCP) or cleaner separation, put a 90° hybrid after matched LNAs (X±jY)/√2.

Tuning & measurement procedure (practical workflow)

Build mechanical can with feedthroughs and backshort. Leave probe tips long (≈ 42 mm) to start.
S11 tuning (VNA preferred): test one probe at a time with the other terminated in 50 Ω. Trim probe length to get S11 dip near 1420 MHz. Small adjustments (a few mm) have large effect.
Backshort fine tuning: if you can, vary short position a few mm and watch S11 — this lets you flatten/bandwidth-tune the match.
Cross-coupling check: measure S21 between probes (with both ports accessible). Aim for ≤ −20 dB. If S21 is high (> −15 dB) try small axial/pin adjustments or use a short coaxial isolator/hybrid.
On-sky or absorber Y-factor: measure cold-sky and absorber temperatures on each channel (or use a calibrated noise diode) to get system temperature and verify both arms comparable.
Radiometer test: record both power time series and compare noise: averaging the two independent powers should reduce RMS by ≈ √2 compared to single-arm (same integration time).

Practical tips & pitfalls

Mode control: if D is too large the can will support higher order modes → outputs will mix and probes won’t be independent. Stick close to the ~135 mm diameter.
Length/resonances: avoid very long cans that create cavity resonances — keep it close to λg/4 region plus small margin.
Probe mounting: keep probes rigid and repeatably placed; loose probes cause frequency shifts and unpredictable coupling.
Cable lengths: keep RX cable lengths equal to avoid systematic phase/gain differences (important if you later hybridize for circular pol).
Grounding & contacts: ensure good electrical contact at the shorting plate and can seams — poor contact causes spurious resonances.
Thermal stability: thermal expansion changes tuning — if you expect outdoor use, plan to re-check tuning across temperature range or use mechanical adjusters.
Balancing LNAs: unequal LNA gains reduce the √2 benefit; equalize gains or normalize channels in software before averaging powers.

If isolation is poor — quick fixes

Add small metal septum or balancing ring inside (a tuned OMT-like feature) — more work.
Use a 90° hybrid and form circular pols; each hybrid output then has better isolation for RHCP/LHCP.
Add a small absorptive strip placed at a strategic location inside the can to damp unwanted cavity modes (use sparingly).

Summary (one-line)

Yes — with a properly sized cantenna (≈ 135 mm diameter), two orthogonally placed probes at the λg/4 plane, two matched LNAs and careful tuning, you’ll get two largely independent polarizations and ≈√2 sensitivity improvement by averaging their powers for unpolarized HI.

I have used a 150cm diameter cantenna for my waveguide on my solar cooker radio telescope, as this is a commonly available coffee tin size found in the UK:

A 150 mm can will still work. It still supports only the TE₁₁ mode at 1420.405 MHz, so two orthogonal probes can give you two largely independent polarizations and the ≈√2 sensitivity gain when you average powers.

Below are the concrete numbers and practical notes (so you can go straight to building/tuning).

Mode cutoffs (150 mm ID)

TE₁₁ cutoff ≈ 1.171 GHz (so TE₁₁ propagates at 1.420 GHz).
Next higher mode (TM₀₁) cutoff ≈ 1.530 GHz (so TM₀₁ does not propagate at 1.420 GHz).
→ Single-mode (TE₁₁ only) at 1420 MHz, which is what you want for clean orthogonal probes.

Guide / geometry numbers (use these for drilling & tuning)

Free-space wavelength, λ₀ ≈ 211.06 mm.
Circular-guide wavelength, λg ≈ 373.1 mm.
Quarter-guide wavelength (backshort distance from the probe plane): λg/4 ≈ 93.3 mm.
Start probe tip insertion (from inner wall into cavity): ~0.18–0.22 λ₀ ≈ 38–46 mm (start ~40–42 mm, trim for match).
Probe angular separation: 90° (crossed).
Probe radial placement: near mid-radius → tip roughly ~37–38 mm from centre (radius = 75 mm; mid-radius ≈ 37.5 mm).

Practical implications of the 150 mm size

Good: wider can is a bit easier to machine and still single-mode at 1420 MHz.
Caveat: the TM₀₁ cutoff is only ~110 MHz above 1420 MHz, so you have less margin to the next mode than with a smaller can (e.g., 135 mm). That means:
- Be careful with irregularities (seams, holes, mismatched shorting plate) — they can excite unwanted cavity behavior.
- Keep the can reasonably short (don’t create long cavity resonances) and ensure solid electrical contacts at seams and short.
Isolation between probes should still be achievable (target S21 ≤ −20 dB). If isolation is poor, you can use a hybrid or small internal tuning features.

Tuning & test checklist (exact steps)

Mount two SMA feedthroughs at the probe plane (λg/4 from planned short).
Fit probes long (~42 mm) initially. Tune each probe individually with the other terminated. Trim for S11 minimum near 1420.405 MHz.
Adjust the short position (if adjustable) by a few mm to widen/broaden the match.
Measure S21 between probes; aim ≤ −20 dB. If worse, try small axial or angular adjustments or use post-LNA hybrid/isolation.
Confirm on-sky/absorber Y-factor and equalize LNA gains—then average powers to get the √2 sensitivity improvement.

astronomy.network The Website of Lichfield Radio Observatory (LRO)