<azonenberg>
I'll look after my stream is over, however long it runs
<azonenberg>
i'm willing to merge pretty alpha driver code in situations like this where the pre-existing functionality is nonexistent or hopelessly broken
<d1b2>
<mubes> I've just come to bed and then realised I've left a debug define switched on (note to self:don't commit code as the last job of an evening) so there is certainly at least that to fix...after that it's just stuff to make it play nice with the rest of the suite I think.
m4ssi has quit [Remote host closed the connection]
juli969 has joined #scopehal
<d1b2>
<theorbtwo> Slightly randomly: why did the stream that you just did use the Lacroix vendor software instead of glscopeclient? If it's missing features, are there wishlist bugs for them yet?
<d1b2>
<theorbtwo> On an even less on-topic note, I find myself pondering trying to make a sort of diy probe kit, very much the other end of the market from your probes. Why are the series resistors in your transmission line design on the tip end instead of the connector end?
<d1b2>
<theorbtwo> Am pondering doing a little adapter that goes on the scope end that makes a direct connection into something closer to a 10:1 probe.
pepijndevos has quit [Ping timeout: 246 seconds]
pepijndevos has joined #scopehal
<azonenberg>
theorbtwo: I used the vendor software because the "zero offset" feature for differential probes, and the SERDES trigger, are not currently implemented in glscopeclient
<_whitenotifier-5>
[starshipraider] azonenberg pushed 1 commit to master [+4/-0/±0] https://git.io/JYyDq
<_whitenotifier-5>
[starshipraider] azonenberg 0548f65 - Initial version of SMA test board
<azonenberg>
theorbtwo: as far as series resistors, you want them as close to the tip as possible
<azonenberg>
The cable is matched to 50 ohm impedance and can be made arbitrarily long with no ill effects other than propagation delay and gradually increasing loss
<azonenberg>
but the stub from the DUT to the resistor creates a reflection
<azonenberg>
you want to minimize the length of that reflection
<azonenberg>
if you had say a meter of 50 ohm line from DUT to the resistor, now your signal on the DUT will come down that meter of cable, hit the resistor
<azonenberg>
a bit will go through to the scope, the rest will reflect back to the DUT
<azonenberg>
With a two meter round trip delay
<azonenberg>
this will give you a quarter wave null in the frequency response at about 40 MHz if my mental math is right
<azonenberg>
which is enough to cause serious problems with even fairly low bandwidth measurements
<azonenberg>
And yes there are wishlist items for both of those
<azonenberg>
But also, my point was mostly to show off probing techniques more so than glscopeclient
<d1b2>
<theorbtwo> Hm, thanks much.
<azonenberg>
An ideal probe would have infinite input impedance at zero distance from the tip of the needle
<azonenberg>
and thus present no DC load on the circuit, and also not have any stub to introduce reflections
<azonenberg>
This is of course impossible in practice, but minimizing stub length and maximizing impedance seen by the DUT are both desirable properties of a probing system
<d1b2>
<theorbtwo> Edge launch coax connectors seem to be obscenely expensive. (Farnell's cheapest is 5 GBP each for SMA for the jack, but they don't offer a matching plug.)
<d1b2>
<theorbtwo> Ah. I want right-angle, not edge, apparently.
<azonenberg>
Edge launch is much easier to get good performance from
<azonenberg>
And 5 GBP is cheap for RF connectors :p
<azonenberg>
that's 7 USD?
<azonenberg>
the one I use on the AKL-PT1/PT2 right now is closer to 12 USD
<azonenberg>
and that's still not a high end connector
<d1b2>
<theorbtwo> Yeah, my priorities go cheap, and acceptable performance. (I'm aiming more toward "I can clearly see what that signal is supposed to be", and not neccessarily that the waveform seen by my scope is the same as the waveform that the other components on the board see.)
<d1b2>
<theorbtwo> Also, the design I have in my head now really wants to be made from adapters, as were, so the availability of both genders of connector is a goodness.
<d1b2>
<theorbtwo> ...and I think BNC is probably better then SMA for me. AIUI, not looked it up properly, BNC can be cycled more, and is designed to be the correct tightness when hand-tightened with zero tools.
<azonenberg>
Other connectors to look at are push-on style like MCX/MMCX
<azonenberg>
or SMP/SMPM but those tend to cost a lot more
<d1b2>
<david.lenfesty> Ooooh idea for active probes (stolen from lecroy via macrofab slack): rgb led on the probe to indicate channel
<azonenberg>
Already planned as a possibility
<azonenberg>
but for the MEAD logic pods i went all the way with an LCD to show channel name for each input
<azonenberg>
Because colors don't work when you have 96 inputs :p
<d1b2>
<david.lenfesty> I kind of assumed you knew about it already but I hadnt seen anything about it :)
<azonenberg>
Yeah i have a bunch of wavelink probes with the led indicators on them
<d1b2>
<theorbtwo> Slightly randomly, because it's clearly out of scope for my first set of ideas: using ethernet magnetics to make a differential probe -- how stupid an idea is it?
<azonenberg>
if you're ok with being AC coupled and only having 200 MHz or so of bandwidth?
<azonenberg>
might not be toooo stupid
<azonenberg>
but opamps with that kind of performance that can run down to DC are pretty cheap
<d1b2>
<theorbtwo> I mean, way past 200 MHz if you use gigabit ethernet magnetics, surely? But 200 MHz is plenty, I'm targeting 100 MHz scopes.
<azonenberg>
and no
<azonenberg>
gig-e is 250 Mbps per pair * 4 pairs
<d1b2>
<theorbtwo> D'oh.
<azonenberg>
and they use PAM-5 so the 250 Mbps is only i think about 125 Mbaud
<azonenberg>
the spectral efficiency of baseT is actually quite decent in terms of bits/sec/Hz
<azonenberg>
typical cat5e cable is tested to 100 MHz or so iirc, cat6 is 250 MHz?
<azonenberg>
although most is better than that, i think my 5e is specced at 350 MHz
bvernoux has joined #scopehal
bvernoux has quit [Read error: Connection reset by peer]
<_whitenotifier-5>
[scopehal] mubes synchronize pull request #402: Changes to support SDS2000+ scope - https://git.io/JYiAM