Quote:
Originally Posted by beery
Hi Doug,
I can tell you that the CRT used in thec15" Baird T5, which is also from 1936, gives a much sharper picture and I'm sure this is down to the CRT design rather than anything else. The 15MW2 crt used in the T5 is a simple directly heated triode tube with magnetic focus.
Andy
|
I think that's the case.
Yesterday I tried my HackRF in transmit again and lo and behold, a recent
Windows 10 update has made it once again work reliably in transmit. It can now reliably generate a 20 MHz wide signal. I'm running it at a 12.5 MHz rate for convenience. The result is that I'm running at 52.08 Hz vertical scan rate so I can use Test Card C images designed for a 6 MHz pixel clock. This requires the minutest tweeks to scan controls. Starting from scratch you could of course run at 50 Hz.
All the test card images I have have hopelessly badly converted frequency test patches which look terrible in the data files. The set pretty much reproduces these correctly.
But I can computer generate test patterns with a 12.5 Mhz pixel clock and have the HackRF filter them digitally at baseband at cutoffs of 1.75, 2.5, 3.5, 5, 5.5, 6, 7, 8, 9, 10, 12 ... MHz. I'm using 3.5. This means I can make really good correctly synced signals with no artifacts or aliasing.
So I tried various patterns and find that, looking directly at the signal to the CRT with a scope (which adds a few pf to the intrinsic value of 26 pf),
the signal is flat to 2 MHz and drops smoothly to zero at 3 MHz, as it should, and as I thought using an old sweep generator. The signal is still
very clear on the scope at 2.7 MHz.
But the CRT is barely capable of, at the sharpest spots, 2 MHz, and over most of the screen more like 1.7 MHz. This is clearly defocus. Its not greatly dependant on brightness up to a fairly light gray or modulation amount so long as its not brighter than that. Above that brightness level the resolution drops very rapidly.
I modestly recommend the HackRF for this purpose. As a general purpose receiver at its absolutely abysmal due to low dynamic range. Most folks are using it at 1 to 6 GHz with Linux. But you have to write your own test files and play them using the command line program hackrf_transfer, a pain. Its NRFPT. Its also expensive.