<azonenberg>
you can see all the little nuances in the eye at 1.25 and 2.5 Gbps, at 5 it's still very decodeable but no longer as useful for signal integrity work
<azonenberg>
(the DUT is an FPGA board I made back in 2014, i'm impressed the eye is as open as it is - i did no simulation and didn't have any ground plane cutouts for connector launches etc)
<Degi>
How do you know that its not useful for SI work rather than the device having a bad eye diagram? The simulated bandwidth?
<azonenberg>
Measured BW with a VNA
<Degi>
I see
Degi_ has joined #scopehal
Degi has quit [Ping timeout: 252 seconds]
Degi_ is now known as Degi
<azonenberg>
So, it seems like Windows has a standard driver for USBTMC devices
<azonenberg>
that has a well defined IOCTL interface
<azonenberg>
i'm not sure how enumeration of devices using a given driver or class or something works on windows
<azonenberg>
Also, just had a nice chat with some folks from Aaronia, a company I hadn't heard of before. Looks to be a small German RF vendor making RTSAs and other gear
<azonenberg>
They're interested in developing a scopehal driver
<electronic_eel>
in the past aaronia was marketing their mobile sas especially to "electrosensitive" people, to find em emitters in your bedroom to sleep better and similar things
<electronic_eel>
don't know if they ceased that kind of marketing
<azonenberg>
i didnt see any evidence of it
<electronic_eel>
yeah, i don't see that stuff on their website anymore too. seems like they abandoned it. very good.
<monochroma>
am aware of the nutube, but is it really considered MEMS? hmm.
<d1b2>
<Catgirl Joyce> nay
<d1b2>
<Catgirl Joyce> it's not MEMS
<d1b2>
<xabean> oh that's the VFD tube!
<d1b2>
<xabean> @Catgirl Joyce and don't feel bad about using vacuum tubes :) I inherited a whole bunch of them from my Dad, and some day want to build a tube amp.
<monochroma>
you can abuse standard VFDs into triode duties too hehe
<d1b2>
<Catgirl Joyce> I know a lot of RF amplifiers using valves are usually tuned for a frequency band of interest
<d1b2>
<xabean> my Dad's 2KW HF amplifier I think has four set of 4-400b tubes in it. Well, mine now.
<d1b2>
<Catgirl Joyce> I suppose you're a ham 😛
<d1b2>
<xabean> woof the Heathkit Sb 220 is pricey on eBay
<d1b2>
<xabean> if only I had any interest in renewing my ham license, or shipping it somewhere (dear god the transformer in it has mass )
<d1b2>
<Catgirl Joyce> hear me out
<d1b2>
<Catgirl Joyce> valves powered by SMPS 😉
<d1b2>
<xabean> doable, with enough filtering right?
<d1b2>
<Catgirl Joyce> yeah
<d1b2>
<xabean> I have bunches and bunches of rectifier tubes though.
<d1b2>
<xabean> lets get those 250vDC at something like 15uA going though
<d1b2>
<Catgirl Joyce> I used to repair and service old test and measurement equipment that uses valves
<monochroma>
i know tektronix had some tens of GHz oscilloscopes back in 1960s (and for /SOME/ reason they had to come up with a radiation decontamination procedure for scopes that would come back pretty mangled from LLNL for repair)
<d1b2>
<Catgirl Joyce> one of my RF signal generators from Heathkit uses valves
<d1b2>
<xabean> @Catgirl Joyce oh man, I wish I had known this. I had a "laborator grade" vacuum tube tester, RCA WT-100A
<d1b2>
<xabean> *laboratory
<d1b2>
<xabean> literally gave it away to someone local-ish
<d1b2>
<Catgirl Joyce> if a large enough solar flare knocks out all our electronics at least we can rebuild our computers and other basic electronics
<d1b2>
<Catgirl Joyce> ooooh
<d1b2>
<xabean> yeah I spent far too much time and money getting that fab'd and put into a pelican case.
<d1b2>
<xabean> but it'll do all the most popular tube sockets, and follows the hickock TV-3/TV-7 wiring scheme
<d1b2>
<xabean> ok this isn't #vintage I'll stop now, heh
bvernoux has quit [Quit: Leaving]
<Bird|otherbox>
azonenberg: you can get into the hundreds of MHz to low GHz with *conventional* tube construction. beyond that, it's klystrons, magnetrons, TWTs, and other exotics.
<Degi>
azonenberg: Maybe for a wideband tube amplifier you could maybe use a coplanar waveguide with a hole poked into it and use that as a gate. That probably doesn't work at DC (since there is a lot of offset), but should work with a relatively large bandwidth, depending on beam velocity. Interelectrode capacitance could basically be 0, and after the beam passes through the gate it gets absorbed by a 50 ohm resistor with a
<Degi>
transmission line connected to it. That way you could have an input and output matched amplifier (though I am not sure how the capacitance of the electron beam effects the input, but I think this should work relatively well for low amplitudes). I think the bandwidth should depend on the ratio of beam velocity to waveguide thickness (but the beam width also gets smaller with a smaller thickness, so that probably limits the
<Degi>
power, but with 10 kV acceleration you could get a few tens of GHz maybe with a mm thick substrate, assuming a hole with 0 diameter though).
<Degi>
(Though not sure if this offers any benefit over using a RF transistor with 50 ohm biasing resistors)
<azonenberg>
Lol. Yeah it would be completely impractical IMO
<azonenberg>
more of a thing to do because you can
<Degi>
I think even at 100 GHz you can probably get some semiconductor amplifier
<Degi>
Hmm, if you can find a fast laser and photodiode, you could use an optical amplifier as a gain element (at least that doesn't need several kV and a vacuum)
<azonenberg>
lol EDFA based probe
<azonenberg>
you would have so many nonlinearities :p
<monochroma>
just have to characterize em out :P
<Degi>
Just put enough attenuators in front of it
<Degi>
I mean it would be useful if your DUT has a DC offset (the spicy end of a particle accelerator beam generator)
<azonenberg>
Lol
<azonenberg>
ok so, in all seriousness
<azonenberg>
degi: you have a rigol mso5000 right?
<Degi>
Maybe for the y axis we can only update it on the frontend until the user stops scrolling and then send the command to the scope to rescale on slow socpes.
<Degi>
*scopes
<Degi>
When I press on Start (single trigger) while it is running, it triggers a few more times and the waveform in glscopeclient updates twice
<Degi>
(But after that works as normal)
<Degi>
(And the context menu takes about half a second to open, fyi my computer is a bit slow)
<Degi>
How do I rescale the vertical axis on a clock jitter recovery?