Shade, glorious shade. (Costco in Kahului, Hawaii)

I agree, they're nicer. But TNC-F and BNC-F mate with N-M so perhaps N is the real threaded BNC. With only the spring force holding the contacts together, BNC are equally unreliable with gentle wiggling or real vibration. TNC nuts have holes for wiring them if you're installing them in an aircraft.

Two colleagues used Intelsat (901) for MeerKAT holography and pointing. Iirc they had a contact there for more accurate coordinates.

Has anyone ever made a Taptic paddle - rigid non-movjng paddles with strain gauge to measure force, and a tiny Linear Resonant Actuator to convey the "click" of having pressed hard enough. Like a MacBook touchpad or your phone keyboard - clicks with no moving parts.

the fine print: says * up to the start of UHF

The BladeRF reference input is a solid analog PLL www.nuand.com/bladeRF-2.0-... that disciplines the internal 38 MHz TCXO. I wonder if the square wave from the LBE somehow upsets the ADF4002 PLL? Maybe try a 10 MHz LPF filter there? Is the amplitude perfect for the PLL, about 3.3 / 0 V square?

Nice clean comparison, I'm glad the FRK is behaving so well. The LBE looks much worse than expected - we're seeing RMS of >0.5 Hz or 2E-10 but published ADEV is 3E-11 at 𝜏=10s. I wonder if something else is going on?

If you need only 10 MHz - I use a Rubidium oscillator as a reference (ok several, from eBay). I measure them as much better than plain GPS receiver up to 1000s; but equivalent to the Thunderbolt GPSDO. I've not tested an LBE, it seems to have a fairly small flywheel: www.eevblog.com/forum/metrol...

Oh I was just rambling about chasing down the source of the pattern distortion, for its own sake. I don't think it affects the peak gain much, and shpuldn't change pointing. (As long as the wire tension is mostly pre-tension, not gravity-tension, i.e fairly constant over elevation).

I think you're saying the same thing - a small systematic surface distortion with 3-way symmetry. Rather than just reflections from the wires themselves (if it wasn't clear, I meant the wires, going out to the rim, not the central rods). Chamging to plastic rope/rod, same tension, would be one test.

Photogrammetry would be quicker, trivial to get <1 mm. You may know someone that can process it. Hard to include the feed position in the whole analysis. Holography would be harder but more in the spirit of it, and you already have a reference antenna, two coherent receivers and accurate pointing!

At 4 cm wavelength metal rods will just create some <-20 dB diffraction spikes, probably too weak to see with a radiometer and a distant satellite. I think the dish surface is probably distorted slightly by the tension on the cables. 5-10 mm would be enough. So now you have a choice...

Perhaps useful: github.com/ska-sa/katpo... implements the pointing model, correcting for 8+ physical imperfections in the encoders and antenna geometry, also solves for these parameters given (many) observed offsets. Well documented, with references. Docs too katpoint.readthedocs.io/en/latest/ka...

I think SKA are rewriting all the software. This archive is everything that controls MeerKAT. Refined over a decade... so many subtle things: Sub-arcsecond pointing errors from single precision libraries. 1 metre Az/El axis offset effect on pulsar time of arrival. But not easy to get into the code.

I can't find a report showing results. But I see all the code is now open source! For making pointing observations: github.com/ska-sa/katsd... and for processing, finding source positions: github.com/ska-sa/katsd... class BeamPatternFit. A 1-page PDF would be the best demo, Sorry I can't find one.

Suggestion from my limited radio telescope experience: for locating sources (or calibrating your pointing) instead of a full raster, just do 3 slow scans across the target, record total intensity at many points. So you have 3x100 or so. Fit your beam to that. Scans could be 1-axis or even drift.

Mine arrived today. It's already been upgraded to a naked BK7231N module and four 3mm LEDs facing in all directions.

If you haven't already, do inhale the whole of galmon.eu including @berthubert.bsky.social 's 2019 reports about the failures and limitations of the ground segment. Lots of detail there, I'm not sure it discusses the thrusters.

Interesting it's linear. Your polarimeter is well-calibrated! I read up about Galileo SV attitude once. Z points towards earth. Y is the rotation axis of solar panels; the cross product of the Z-axis with the vector from satellite to Sun. So rotation around Z, and pol angle, is fixed by the sun.