<d1b2>
<dannas> Some more SCPI commands added. I can query for most properties now. I've pushed the parsing of the triggers and the actual waveforms in front of me.
<d1b2>
<dannas> So. Hopefully I can do some more hacking tomorrow evening. I really hope the Rigol MSO8000 won't need this much modifications. (If I ever get to do that).
<azonenberg>
dannas: are you caching all of the settings yet?
<azonenberg>
especially with these really slow scopes caching is critical so you don't constantly poll for the same configuration
<d1b2>
<dannas> @azonenberg I'm relying on the existing caching mechanism in place in the driver. Or rather: I'm relying on an assumption that the driver does caching. 🙂
<azonenberg>
I didn't write it, that was mubes. but it probably does
<d1b2>
<dannas> So far I've just added one SCPI command after another. Will zoom out and look at the overall architecture tomorrow. But there is a cache_mutex so I assume the driver does some kind of caching.
<azonenberg>
Yeah
<azonenberg>
OK so i'm starting to design a test fixture for these probe tip axial lead resistors. The Lecroy ones look to be cut to 7.5mm overall length
<azonenberg>
my vision is a pcb with a SMA at either end of a small cutout in the PCB, which i can solder the resistor across
<azonenberg>
So I'm thinking I should put a 6mm long slot in the board which would give me 0.75mm at either end attached
<electronic_eel>
can't you use two pt5 prototypes and bridge the series resistors? i guess the problem will be the cal standards. but that will be the same with a custom board with slot in it
<azonenberg>
bridging the PT5 resistors, adding ground ties between the two boards, and converting from SMPM to SMA and back would add a lot of uncertainty
<azonenberg>
if i have a board that terminates in a SMA and has a very short length of trace on the board, I can use ref plane extension in the VNA software to remove the electrical length of the transmission lines
<azonenberg>
i've already characterized the rosenberger sma launch so i know what it looks like
<azonenberg>
and the loss of the ~8mm of pcb trace should be pretty negligible
<electronic_eel>
yeah, you are right
<azonenberg>
also the pcb cost is going to be next to zero, it's under a square inch probably
<electronic_eel>
another idea: you got those launch test boards. could you mill a slot into one for the resistor?
<azonenberg>
and while the rosenberger SMAs are $$, i can desolder the center pin and reuse them
<azonenberg>
Yes, I could. But i want very repeatable geometry
<azonenberg>
my goal is to characterize rev A and rev E lecroy tip resistors, 100 ohm HML01s, and 450 ohm HML01s
<azonenberg>
I also made a test fixture for a Rosenberger 10 GHz BNC
<electronic_eel>
you could mill a slot into one board and solder one resistor after the other onto the same board. so the geometry would be the same. just the solder joint will be a bit different
<azonenberg>
I have no immediate need for it but it was cheap and enough test equipment uses BNCs that I want to have the option available in the future
<azonenberg>
the weird bit is, all of the >4 GHz BNCs I can find are thru hole, not edge launch
<azonenberg>
which seems very backward
<azonenberg>
you ~have to use a bottom side launch or you get horrible stub effects
<azonenberg>
and there is *still* stub effects wti ha bottom side launch but the null is far enough out of band that the edge of the dip you get is only like half a dB at 10 GHz in my early sims
<electronic_eel>
my guess re high freq bnc: they are often used by scopes. the scope frontends i've seen always use through hole
<electronic_eel>
... and the frontend circuit is on the back of the bnc, so bottom side launch fits perfectly
<azonenberg>
i guess if you have a vertical PCB that's sane, although i was targeting a single flat PCB with right angle BNC for my test
<azonenberg>
My plan for my future scope frontends is to use a 32K243-40ML5 and then top side frontend as close to the SMA as I can get it