<azonenberg>
Work has invited me to do a blog article on glscopeclient for our lab blog
<azonenberg>
With emphasis on applications for hardware security work
<azonenberg>
but it should bring some useful attention to the project
Nero_ has joined #scopehal
Nero_ is now known as NeroTHz
<_whitenotifier-f>
[scopehal-apps] azonenberg commented on issue #130: Digital channels display protocol decodes in the right margin area where analog channels have Y axis text - https://git.io/JUCRI
<_whitenotifier-f>
[scopehal-apps] azonenberg closed issue #130: Digital channels display protocol decodes in the right margin area where analog channels have Y axis text - https://git.io/JJqpo
<azonenberg>
pepijndevos: so i was looking at your rigol scope trace
<azonenberg>
re the "extra data on the other channels" problem
<azonenberg>
one possible explanation there is that if the channel is disabled in hardware, and the UI doesn't know that
<azonenberg>
that the memory and ADC capacity is likely used for other stuff
<azonenberg>
the DS1000Z series for example uses the HMCAD1511 ADC which is four 250 Msps converters that interleave to 2x 500 or 1x 1000 Msps when some channels arent in use
<azonenberg>
so if you have CH2 running in 1 Gsps mode with 1 and 3/4 disabled
<azonenberg>
and you try to view one of the other channels, it's very likely you'd see every 4th sample from CH2 or something like that
<_whitenotifier-f>
[scopehal-apps] azonenberg closed issue #101: Show current voltage where a vertical cursor touches the waveform - https://git.io/Jfzpb
<_whitenotifier-f>
[scopehal-apps] azonenberg commented on issue #101: Show current voltage where a vertical cursor touches the waveform - https://git.io/JUCR8
_whitelogger has joined #scopehal
_whitelogger has joined #scopehal
juli965 has joined #scopehal
<NeroTHz>
oh for the love of god, some of the industry partners I have to work with...
<NeroTHz>
¨We put 0 V on the AM Modulator pin of your IC, and no RF power comes out, what is wrong? ¨
<NeroTHz>
and then ¨We then linearly increased the DC voltage on the AM modulator pin, and the RF power increased linearly with it, but we don´t understand why¨
* NeroTHz
bangs head into desk, repeatedly
<monochroma>
XD
<monochroma>
have they done RF before, or is the documentation lacking
<NeroTHz>
It also turns out that their supply for the VCO is just straight form a switching powersupply and it has >50 mV peak-to-peak ripple
<monochroma>
so they got the intern to do it
<NeroTHz>
no, these are guys from optics using millimeter-wave chips for the first time
<monochroma>
ah
<NeroTHz>
so they are clueless, and didn´t listen to use when we told them what to do
<NeroTHz>
because they still don´t get that we as research institude don´t make production-ready ICs
<NeroTHz>
our ICs have like 15 supply lines and 10 bias lines because we don´t want to spend months designing on-chip LDOs and DACs, both because we have better things to spend our time on, and we want the flexibility to play around with those voltages during testing and characterization
<NeroTHz>
but these guys are used to taking some samtec or whatever photo-diode receiver that has a 3.3V supply and some digital IO and that is it
<NeroTHz>
they also didn´t get that when we say that our VCO needs to be biased at 0.438 V, we mean 0.438 V and not 0.4 or 0.5
<NeroTHz>
but they don´t listen, applied 0.5 V, and boom, the LO is 20 GHz offset from design
<monochroma>
lol
<NeroTHz>
sigh
<NeroTHz>
and they don´t get fundamental performance limits either
<NeroTHz>
they seem to think they can just say ´Hey this 25 GHz bandwidth is great, not just make it 50 for the next version, thanks´
<NeroTHz>
´Now just make your square-law diode detector linear, thanks´
<monochroma>
one company i worked for, their first light ASIC bringup PCBs were littered with beefy "high current" DACs so all voltage rails into the textool socket could be easily programmed
<NeroTHz>
our bias/supply PCBs are similar. most of our supplies are like 10-100 mA, so it´s doable with DAC or DAC+high-power opamp
<NeroTHz>
that said, one of my (many) side-projects is looking at getting some standard IP blocks for on-chip biases and LDOs
<NeroTHz>
because our pad-couts are getting out of control and it makes wirebonding a pain, so we need to go stggered-row flip-chip and stuff
<NeroTHz>
and for RF probing, we pretty much need to be able to wirebond so flip-chip is not possible
<monochroma>
how many pads ? :O
<NeroTHz>
depends, but some of the IQ modulators >80 I think. which is not *that* much, but it´s not trivial to bond, esp not if you also need to squeeze in two-three RF GSG probes
<NeroTHz>
because we usually also need to keep some bondwires short, because they need to carry >40 GHz basband signals
<NeroTHz>
anyho, lunch, be back in 20
<NeroTHz>
what the
<NeroTHz>
azonenberg, I ordered solderspheres from easyspheres
<NeroTHz>
we got an empty vial
<NeroTHz>
(I think - a colleague collected the shipment and sent me a picture of an empty vial with the caption ¨they shipped us air¨)