azonenberg changed the topic of #scopehal to: libscopehal, libscopeprotocols, and glscopeclient development and testing |,, | Logs:
m4ssi has joined #scopehal
futarisIRCcloud has joined #scopehal
futarisIRCcloud has quit [Quit: Connection closed for inactivity]
m4ssi has quit [Remote host closed the connection]
azonenberg_work has quit [Remote host closed the connection]
azonenberg_work has joined #scopehal
bvernoux has joined #scopehal
<bvernoux> hello
<azonenberg> o/
<azonenberg> bvernoux: digikey was out of the MMCXes so i ordered them from mouser, they'll be here thursday
<bvernoux> ha ok good
<azonenberg> that's the last holdup i think until i send you the stuff
<bvernoux> perfect
<azonenberg> I have to make up a whole probe tip and everything too
<azonenberg> i assume you do not want the probe pcb glued into the enclosure?
<bvernoux> do you have found what is the issue with the probe tip ?
<azonenberg> for easier testing best to be separate?
<azonenberg> I have not
<bvernoux> as it seems mainly related to it
<azonenberg> What i did do was swap the tip with an edge launch SMA
<bvernoux> if possible do not solder it on my probe
<azonenberg> You want the tip not installed?
<bvernoux> I will try with and without with measurements and accurate signal generator to check
<azonenberg> OK
<azonenberg> Maybe best if i send you one premade and one bare then
<bvernoux> I will also check the tip alone
<bvernoux> ha yes it will be very nice
<azonenberg> So the issue is the tip is two separate conductors
<azonenberg> the ground and the signal
<azonenberg> it will be hard to test the tip in isolation
<bvernoux> yes I need to check how I will test it
<azonenberg> best option would be to de-embed by measuring the probe+tip and probe only, i think?
<bvernoux> yes
<bvernoux> to characterize both
<azonenberg> I have two PCBs so i'll send you one with resistor+MMCX and one with resistor+MMCX+tip?
<bvernoux> with S11 measurement & smith chart
<azonenberg> i can include a second set of tip parts loose as well
<bvernoux> yes perfect
<azonenberg> Speaking of smith charts i need to do more reading to understand how to interpret them
<azonenberg> i know of them, and i'd recognize one if i saw it
<azonenberg> but i dont know how to read them yet
<bvernoux> most important is S11 for that
<bvernoux> and also check how it work on real circuit attenuation
<bvernoux> ...
<azonenberg> also, s11? not s21?
<azonenberg> for probe signal quality i'd think S21 is what you want
<bvernoux> also S21
<bvernoux> both
<azonenberg> You should ideally see -26 dB S21 from DC to infinity
<azonenberg> if the resistor is purely resistive and all impedances are perfectly matched
<azonenberg> of course we know that isnt the case :)
<bvernoux> yes
<bvernoux> interesting point is we can deduce the Inductance & Capacitance of different elements
<azonenberg> The other thing i want to try is adding a 9mm long open stub just behind the resistor, early qucs simulations suggest that this should null out some of the peaking i've been seeing
<bvernoux> not deduce but compute ;)
<azonenberg> but that was using an idealized model of the probe tip which doesn't map to experimental results as well as i'd like
<bvernoux> interesting information on smith-charts
<bvernoux> to understand how it works
<bvernoux> I like smith charts for antenna tuning in fact
<bvernoux> visually you see quickly if it is ok
<azonenberg> hopefully my probe isnt an antenna
<azonenberg> buuut you never know
<bvernoux> also important point is to check VSWR for the probe
<bvernoux> at least from 1Mhz to 2GHz or more
<bvernoux> in few days I could do measurement up to 3GHz
<azonenberg> You should have that capability by the time it gets there
<bvernoux> yes
<bvernoux> anyway I can also characterize probe only with signal generator
<bvernoux> it is longer to do ;)
<bvernoux> with freq meter power meter ;)
<azonenberg> Yeah thats what i did with the last version and got the really bizarre S21 chart that i'm not super confident in
<bvernoux> yes it is why VNA is a must for that
<bvernoux> with a very good accurate clock too
<bvernoux> I shall receive GPSDO soon coupled to a 4 Chan OCXO/GPSDO in order to synchronize all instruments ;)
<azonenberg> ooh
<bvernoux> LeoBodnar Mini Precision GPS Reference Clock is one of the best for the price
<azonenberg> i want to build a 1U clock distribution board based around (tentatively) a LMK04806 and a GPSDO
<azonenberg> take in a 10 MHz input, phase lock to it, then spit out arbitrary synced clocks
<azonenberg> so for example if one instrument wants a 100 MHz ref you can do that
<bvernoux> mine is based on 100USD GPSDO + Clock Distribution(with OCXO at 10ppb which can be also connected to GPSDO) for 99USD too ;)
<azonenberg> or you can phase shift the outputs
<bvernoux> so for a total of about 200USD it is not too bad
<bvernoux> ha yes playing with phase is even better
<bvernoux> I have done that on SI5351C ;)
<bvernoux> it is clearly amazing for the price
<bvernoux> what I do not know is the difference between GPS Antenna
<bvernoux> vs
<bvernoux> mainly to put on my roof so 10m cable is enough
<bvernoux> I have also lot of UpDownConverter_moRFeus ;)
<bvernoux> as they are very cheap (I have 3 IIRC°
<bvernoux> and it is fun to do special test and chain them ...
<bvernoux> my 4 Output 10MHz Freq Standard =>
<bvernoux> I will check it ;)
<bvernoux> very cheap with an OCXO with better than 10ppb stability when GPSDO is not connected ;)
<bvernoux> it use recycled OCXO NDK ENE3311
bvernoux has quit [Quit: Leaving]