SESSION 1 · 08/26 · PARTLY WORKED
First light — and a cable that lied to me
Got the SV205 talking to AstroDMx and the Moon onto the sensor. Burned most of the session on two failures that both presented as camera limitations and were neither.
What I set out to do
Point the 76 mm at the Moon, get the SV205 recognised, and record something. That was the whole ambition. It took far longer than it should have.
The camera "couldn't" do full resolution
AstroDMx offered me a max resolution well below what the SV205 is rated for. The obvious read is that the camera is overstated, or that Linux support is partial. It was neither.
Uncompressed YUYV at 3264x2160 and 15 fps needs about 212 MB/s. That does not fit through USB 2.0 — and rather than offer the mode and fail, the camera simply does not advertise it at all on a 2.0 link. So there is no error, no warning, nothing to search for. The mode is just absent, and it looks exactly like a camera limitation.
Blue port and a USB 3.0 cable. Then:
cat /sys/bus/usb/devices/*/speed # must read 5000
After that: full resolution, 14.9 fps, no other change.
Then the exposure settings wouldn't stick
Second hole in the evening. I'd write an exposure value, the write would succeed, and nothing would change. No error at any point.
exposure_time_absolute reports flags=inactive while auto_exposure is on, and the
UVC driver accepts a write to an inactive control and silently discards it. Same
story for white balance. The fix is ordering — autos off first, then the dependent
values, then read back and compare. And the read-back matters, because that is the only
thing that would have caught it.
And there is no gain control
Went looking for a gain slider for a while. There isn't one. The full v4l2 control list
has no gain under any name. Exposure time is the only brightness lever this camera has:
exposure_time_absolute, units of 100 µs, range 1–10000, so 0.1 ms to 1 s.
Where it ended
Frames on disk, the Moon recognisably the Moon, and three lessons that each cost more than they should have. Every one of them looked like a hardware limit and was a software or cable problem.
Nothing here is a complaint about the gear. It's a note that "the device says no" is not the same as "the device can't" — and I will assume that much faster next time.