I just recently acquired an Agilent Acqiris U1067A-001 8 bit, 150 MHz, 500 MS/s PCI digitizer. It's a mid 2000's 1CH oscilloscope card. The price was "OK" (TEA GAS justification), I needed a fast sampling digitizer for a future project. The card itself was probably taken out of a workstation PC. I suppose some companies throw their PCs away and don't check the insides for treasures...
The card certainly looks well designed: gold plated BNC connectors, large heat spreader and a dedicated fan. The board is EMC shielded in order to deal with noise and emissions inside of a PC housing. The input stage is connected to a mezzanine board, just in case you fry the input attenuator and need a replacement.
Well, after initial inspection, I had to find a suitable PC with a PCI slot to power it up. My victim was a dust-covered Dell Precision T5400. Although it has 2 PCI slots inside, there is almost no room for two heat generating oscilloscope cards. The NI PCI-5105 had to be relocated ;-)
The setup was straight forward: installed the card, powered up the PC, installed the drivers (provided by Keysight), rebooted and "We're in like Flynn". The drivers haven't been updated since 2010 but they still work on a Win10 x64 system. There are drivers availible for Windows, Linux, VxWorks and other systems. The drivers also support LabVIEW, MATLAB and C/C++ programming languages so it should be possible to use a Python wrapper in order to get the data acquisition running.
A quick test was rather sobering: the acquisition software AcqirisLive was very primitive. The waveform windows were fixed and not resizeable, the scaling of the graphs was not possible. The UI wasn't great but it did the job. It felt like it was "built for your Win98 machine just to be viewed on a 800x600 14" CRT monitor". Nevertheless, one can write his own acquisition routines so there is no need for a scope software.
I performed few simple tests with standard waveforms (80 MHz sine, 1 MHz square wave). Everything was fine, nothing unusual. I couldn't test the performance of the card yet because of lack of time/equipment (who would have thought -> I need more Test Equipment!). I hope some experiments will follow in 2022... So many projects (another TEA GUS justification here...)