<Bird|otherbox>
azonenberg: now that I have that sigc++ suppression enabled, do you want me to just PR what I have and let you sort out the warnings cppcheck spat out, go on a warning cleanup spree and tackle everything it spat out myself, or work on adding more analyzers?
Degi has quit [Ping timeout: 264 seconds]
Degi has joined #scopehal
<azonenberg>
Bird|otherbox: you have clang-analyzer and cppcheck working? that is good as is. Send a PR for it now
<Bird|otherbox>
just cppcheck
<azonenberg>
That's still a good start
<azonenberg>
Send a PR for cppcheck, i'll merge it
<Bird|otherbox>
I'll put a PR up then
<azonenberg>
your next task after that is to, in any order you choose
<azonenberg>
* add clang-analyzer support
<azonenberg>
* go through the cppcheck warnings and either add suppressions if they're false positives or file tickets/send PRs if they're legit
<azonenberg>
Bird|otherbox: also i plan to add you to the project on github officially. This will give you the ability to assign tickets to yourself etc which makes things easier to manage from my perspective
<Bird|otherbox>
ok, coolio
<azonenberg>
It will also give you push access however i ask that you not push anything to master without my signoff. Feel free to work in a feature branch on the actual repo though
<azonenberg>
in fact i'd prefer that as it lets me keep all of the draft code in one place, easier to know how stuff is going
<azonenberg>
Are you interested in tackling the installation related tickets too, as long as you're on a build system rampage?
<azonenberg>
basically make "make install" work, as well as adding CPack support so we can create binary releases in .deb format etc
<azonenberg>
i do not intend to actually *have* an official binary release until the project has matured a bit, but we need the infrastructure between now and then
<azonenberg>
invitations sent to scopehal, apps, docs
<azonenberg>
accept those and i'll assign the analysis ticket to you
<azonenberg>
also longer term i should probably make a github org for this project vs having it all under my account but that's a task for later
<Bird|otherbox>
azonenberg: yeah, makes sense re: not shoving stuff into master directly
<Bird|otherbox>
install...will likely come later. generating packages is a fiddly business
<azonenberg>
CPack does a lot of it for you. you dont have as much control as if you used the distro tooling but it's a good first step
<azonenberg>
basically once you get the cmake to the point that "make install" works, CPack should take 5 mins to configure
<azonenberg>
And it's not a priority but are you interested in working on it at some point?
<Bird|otherbox>
I probably could
<azonenberg>
(did you get the invitations?)
<_whitenotifier-f>
[scopehal-docs] azonenberg pushed 1 commit to master [+0/-0/±1] https://git.io/JTeEt
<_whitenotifier-f>
[scopehal-docs] azonenberg 95559e3 - Documented scrolling to zoom
<_whitenotifier-f>
[scopehal-apps] azonenberg pushed 1 commit to master [+0/-0/±2] https://git.io/JTeEm
<_whitenotifier-f>
[scopehal-apps] azonenberg b430de2 - Shift-scroll in a WaveformArea now pans horizontally. Fixes #29.
<_whitenotifier-f>
[scopehal-apps] azonenberg closed issue #29: Shift-scroll on viewport should pan horizontally - https://git.io/JvEtd
Nero__ has joined #scopehal
Nero_ has quit [Ping timeout: 272 seconds]
<_whitenotifier-f>
[scopehal-apps] azonenberg added user tarunik - https://git.io/JUgq6
<_whitenotifier-f>
[scopehal] azonenberg added user tarunik - https://git.io/JUgq6
electronic_eel has quit [Ping timeout: 240 seconds]
<_whitenotifier-f>
[scopehal-apps] tarunik commented on issue #180: Look into static analysis options - https://git.io/JTegi
<_whitenotifier-f>
[scopehal-apps] azonenberg commented on issue #180: Look into static analysis options - https://git.io/JTe23
<_whitenotifier-f>
[scopehal] azonenberg pushed 2 commits to master [+0/-0/±8] https://git.io/JTeVe
<_whitenotifier-f>
[scopehal] azonenberg 07fcec6 - SDCmdDecoder: fixed bus width command name
<_whitenotifier-f>
[scopehal] azonenberg d1a1fe7 - Oscilloscope: added Get/SetChannelDisplayName() rather than having m_displayname freely mutable. Allows for future sync of channel display names on hardware.
<_whitenotifier-f>
[scopehal-apps] azonenberg pushed 1 commit to master [+0/-0/±10] https://git.io/JTeVf
<_whitenotifier-f>
[scopehal-apps] azonenberg b2f0ac4 - Channel display name is now accessed via accessors rather than directly
_whitelogger has joined #scopehal
<_whitenotifier-f>
[scopehal] azonenberg pushed 2 commits to master [+0/-0/±6] https://git.io/JTeSD
<_whitenotifier-f>
[scopehal] azonenberg closed issue #267: Add APIs for displaying channel names on instruments that support this - https://git.io/JUamY
<_whitenotifier-f>
[scopehal-apps] azonenberg pushed 1 commit to master [+0/-0/±1] https://git.io/JTeSj
<_whitenotifier-f>
[scopehal] azonenberg pushed 1 commit to master [+0/-0/±1] https://git.io/JTeNv
<_whitenotifier-f>
[scopehal] azonenberg 2ea0539 - MDIODecoder: correctly handle back-to-back transactions and >32 bit preambles
<azonenberg>
Sooo i'm looking at scopehal:#112, the manchester line code filter
<azonenberg>
What should the output format be?
<azonenberg>
recovered bits and clock as two separate streams? should i deserialize and output bytes? should i support non-byte-sized data streams?
<_whitenotifier-f>
[scopehal-apps] azonenberg pushed 1 commit to master [+0/-0/±3] https://git.io/JTvJJ
<_whitenotifier-f>
[scopehal-apps] azonenberg 361221f - Changed window manager hint in HistoryWindow/ProtocolAnalyzerWindow. See #215.
<_whitenotifier-f>
[scopehal-apps] azonenberg commented on issue #215: Window title of protocol analyzer windows not shown or imcomplete - https://git.io/JTvJk
<azonenberg>
bluecmd[m]: when you get a chance please see if 361221f fixes the missing titlebar bug
<_whitenotifier-f>
[scopehal-apps] azonenberg commented on issue #209: glscopeclient Fullscreen issue with dialog box - https://git.io/JTvJZ
<_whitenotifier-f>
[scopehal-apps] azonenberg commented on issue #140: cmake doesn't check for presence of yaml-cpp - https://git.io/JTvJ8
<_whitenotifier-f>
[scopehal-apps] azonenberg pushed 1 commit to master [+0/-0/±3] https://git.io/JTvUW
<_whitenotifier-f>
[scopehal-apps] azonenberg e2efbc9 - Non-overlay filters are now displayed in the "add" menu. Fixes #154.
<_whitenotifier-f>
[scopehal-apps] azonenberg closed issue #154: Allow hidden filters to be added to display - https://git.io/JJaD6
jevinskie[m] has quit [Quit: killed]
promach3 has quit [Quit: killed]
bluecmd[m] has quit [Quit: killed]
jevinskie[m] has joined #scopehal
pepijndevos has quit [Ping timeout: 260 seconds]
pepijndevos has joined #scopehal
pepijndevos has quit [Max SendQ exceeded]
pepijndevos has joined #scopehal
_whitelogger has joined #scopehal
<pepijndevos>
You mentioned +/-5V and 50 Ohm termination, but the first stage of attenuation and filtering would not be on-chip right? So the IC would get -/+2.5V and should have high-impedance input?
deltab has joined #scopehal
deltab has quit [Max SendQ exceeded]
deltab has joined #scopehal
bluecmd[m] has joined #scopehal
promach3 has joined #scopehal
<pepijndevos>
I'm a bit confused about the interaction between U1 and U18. U18 adds the 2.5 ref and the offset voltage, but then the offset voltage is also applied to U1
<pepijndevos>
ah uh... does the common mode of U1 create an offset on VIN_filtered, that has to be canceled out by U18?
juli965 has joined #scopehal
<bluecmd[m]>
azonenberg: titlebar looks fine for me now on the protocol decoder
<bluecmd[m]>
I can move it around and it has X buttons
<bluecmd[m]>
azonenberg: does glscopeclient support "magnitude / phase plots"? It is something some users are a bit sad that Gaw (our current viewer) does not do
<bluecmd[m]>
I guess you have FFT viewers so that should be magnitude bode plots, but this is all beyond me tbh
<Degi>
pepijndevos: U18 is to cancel out any DC offset at the input caused by U1
<Degi>
Basically U1 causes some current to the input which otherwise would cause it to be not 0 V when unconnected, U18 causes it to be 0 V again
<pepijndevos>
right
<Degi>
Thus the input offset is I think < 1 mV
<Degi>
Basically we don't want the scope to source current into the probe
<pepijndevos>
yea, it took me a while to realize the fully differential opamp created an offset at the input
<Degi>
Yeah
<Degi>
Hm, maybe in the next revision we should have two U1s, where the first one has Vcm at 0 V and VIN_OFFSET at 0 V too and the second one does offset and CM shift
<pepijndevos>
a billion revisions from now maybe it'll use my custom designed IC hehe
<Degi>
Hm we could try DIY ICs
<pepijndevos>
I'm trying to design a frontend in Sky130, then PDK google has open sourced
<pepijndevos>
just as a challenge basically
<Degi>
How do you even get something MFGed there?
<Degi>
I guess its not like JLCPCB?
<pepijndevos>
Well, for now google is sponsoring a shuttle program.
<pepijndevos>
Not clear to me what it'd cost to do it without the shuttle program
<pepijndevos>
So the one thing I'm not super clear about in terms of requirements is the input range and impedance.
<Degi>
50 Ohm, 5 V would be good, or 1 M and 400 V
<monochroma>
for larger node sizes, depending on design size, you can get MPW designs done for <$100k
<Degi>
MPW?
<pepijndevos>
multi porject wafer
<pepijndevos>
So you share the mask costs with a bunch of projects and get a few dozen chips
<pepijndevos>
Degi, azonenberg sent me the schematic of the input characterization board, and it seems that there is already some filtering, attenuation, and termination.
<Degi>
Ah, you want to keep that with a custom IC though?
<pepijndevos>
well, on-chip inductors suck, so I assume you keep the filter external.
<pepijndevos>
so yea the question is what the actual requirements on the IC are, and what is better taken care of by discrete components
<Degi>
Guess u could use RC filters
<pepijndevos>
6th order RC filter hehe
katharina has joined #scopehal
katharina has quit [Quit: Lost terminal]
<_whitenotifier-f>
[scopehal-apps] nshcat commented on issue #215: Window title of protocol analyzer windows not shown or imcomplete - https://git.io/JTvwG
<bvernoux>
hi
<Degi>
Heya
<bvernoux>
so I plan to finally buy Rigol MSO5074 + Sensepeek 4016 PCBite 2x SP200 Kit ;)
<bvernoux>
at batronix
<Degi>
Oh thats neat
<bvernoux>
I will buy later the MSO64B ;)
<Degi>
Can u attach rigol probes to it?
<bvernoux>
yes
<bvernoux>
it is the plan to hack it ;)
<bvernoux>
to check how to attach better probes
<Degi>
Now I wonder if it is possible to build a 2 GHz scope for < 1000 € per channel
<bvernoux>
ha yes to hack the Rigol to push it to 2GHz BW ;)
<Degi>
Heh
<bvernoux>
it is potentially a project I would love ;)
<bvernoux>
I will check
<Degi>
I mean I found an ADC for like 200 € with 2 GHz BW and 5 GS/s... Though its kinda hard to get.
<bvernoux>
as it is the aim because 350MHz BW is not enough ;)
<bvernoux>
the MOS5074 have a 10GS/s ADC in fact ;)
<bvernoux>
they have limited it to 8GS/s ...
<Degi>
Lol
<Degi>
Is it a custom design by rigol?
<bvernoux>
yes
<bvernoux>
it is the same chipset used on bigger version MSO7000/8000 ...
<Degi>
lol
<bvernoux>
until I can buy a real scope with 50GS/s ;)
<bvernoux>
I wil start a project (in background) to hack it to unlock > 350MHz BW ;)
<Degi>
We could build one... Maybe for less than 50 k...
<Degi>
*5k
<bvernoux>
so far I'm not sure it is a HW limitation but it is possible
<bvernoux>
and the aim is to have it stable to send data over ethernet Gb
<bvernoux>
Degi, do you have tried again your python script to reproduce better the issue ?
<bvernoux>
as it is clearly something to report to Rigol as they shall fix it
<bvernoux>
-shall+must ;)
<Degi>
Hm didnt play with it any further
<Degi>
It seems to work pretty reliably, but maybe I can reduce it somehwat
<Degi>
Are telephone cables even rated for 1 kV continuous...
katharina has quit [Quit: Lost terminal]
<_whitenotifier-f>
[scopehal-apps] azonenberg closed issue #215: Window title of protocol analyzer windows not shown or imcomplete - https://git.io/JUbF2
<_whitenotifier-f>
[scopehal-apps] azonenberg commented on issue #215: Window title of protocol analyzer windows not shown or imcomplete - https://git.io/JTv5E
<azonenberg>
bluecmd[m]: as of now there isnt a way to view fft phase
<azonenberg>
bvernoux: yeah lecroy has a 3 year warranty even on their factory refurb scopes lol
<bvernoux>
azonenberg, yes like others Rigol, Keysight ...
<bvernoux>
azonenberg, it is clearly a things to be fixed @Tek without additional cost
<bvernoux>
as warranty extension is very expensive
<bvernoux>
azonenberg, I have someone which contacted me about LeCroy scope ;)
<bvernoux>
I told them I want a product with same specification as Tek 6 MSO64B for max 15KUSD ;)
<bvernoux>
50GS/s, 1GHz extensible by SW to 8GHz ;)
<bvernoux>
MSO64B are even up to 10GHz now all is by SW
<bvernoux>
azonenberg, it is possible in future near or far I will be full time on my own project ...
<bvernoux>
azonenberg, so far it is very hard to do something on embedded/hardware stuff with a full time job ...
<bvernoux>
azonenberg, I do not know how you are doing as IIRC you are working full time in your job too ?
<azonenberg>
bvernoux: yes i am. I was up working on scopehal until 3am last night
<azonenberg>
:p
<bvernoux>
yes it is crazy on log term to do not burn
<bvernoux>
azonenberg, do you have bought the Sensepeek 4016 PCBite 2x SP200 Kit ?
<azonenberg>
No. It's on my list of things to play with. But not yet
<bvernoux>
I plan to buy it to check if it can be hacked to use GHz probes ;)
<azonenberg>
did you see my tweet from yesterday?
<bvernoux>
no ?
<azonenberg>
i was just talking about designing a transmission line probe for it lol
<azonenberg>
saying i wanted to buy one and design a 5 GHz head
<azonenberg>
I do however have the cheap fx2 LA we talked about. It came in a day or so ago and i havent had time to do anything with it yet
<bvernoux>
ha yes nice ;)
<bvernoux>
the challenge is to find a good tip with spring like their probes
<bvernoux>
which reach multi GHz
<bvernoux>
yes the nice stuff is we can replace their probes ;)
<azonenberg>
PMK has pogo tips that fit the same socket i used in the AKL-PT1
<azonenberg>
i have one that came with the ZS1500. havent tried it or measured performance yet
<azonenberg>
only used the rigid ones
<bvernoux>
ha yes will be very interesting ;)
<bvernoux>
I will take also their LA probes ;)
<bvernoux>
it is nice to probe SPI, UART ...
<azonenberg>
bvernoux: also i'm just getting the paperwork rolling to pick up two active 4 GHz differential probes from LeCroy
<azonenberg>
one is a D400A-AT which is a handheld browser module with diff amplifier and two probe tips and a thumbscrew to adjust spacing
<azonenberg>
the other is a D420 which has a socket that can mate to one of several tips, three of which are included
<azonenberg>
the SI tip (solder in, two semi-permanently-attached damping resistors)
<bvernoux>
ha yes very interesting
<bvernoux>
to be used as reference probes ;)
<azonenberg>
the QC tip (quick connect, as far as i can tell it comes with some special carbon resistors you solder to the DUT then has a connector that mates with them, you can have several sets and move the probe around)
<azonenberg>
and the SP tip (square pin, fits 2 pin 2.54mm header, limited to 3 GHz)
<bvernoux>
ha great
<azonenberg>
The D400A-AT comes with a WL-PBUS2 which is an 8 GHz capable ProBus2 adapter cable that also works with ProBus instrments up to 4 GHz (like my WaveRunner)
<azonenberg>
The D420 does not come with an adapter cable but i made an offer from an ebay seller, which was accepted, for a WL-PBUS which is the 4 GHz max ProBus predecessor to the WL-PBUS2
<azonenberg>
So basically once everything is here I'll have one 4 GHz solder in active diff probe and one 4 GHz handheld browser probe
<bvernoux>
yes very nice
<bvernoux>
for a total price >4KUSD I imagine ?
<azonenberg>
The D400A-AT plus WL-PBUS2 is $3840.13 including tax and shipping
<azonenberg>
The WL-PBUS2 was $620 asking price on ebay, I got it for $500 plus tax and shipping from Tel Aviv. Ended up being $590ish total
<bvernoux>
with a MSRP of >=8KUSD I imagine for the D400A-AT plus WL-PBUS2
<azonenberg>
Yes. General rule is that lecroy refurbs are sold at around a 50% discount
<azonenberg>
from MSRP
<azonenberg>
So it's probably $7500-8000 MSRP
<bvernoux>
do you have a warranty on it ?
<azonenberg>
The D420 is $1760 asking price from lecroy refurb, plus tax, shipping, and calibration to match the WL-PBUS which i'm not yet sure if is included or not
<azonenberg>
Lecroy refurbs have the same warranty and accessories as if you had bought it new
<azonenberg>
They're in mint condition
<azonenberg>
i believe it's 3 years on scopes and one on probes
<bvernoux>
ha ok it is refurb
<azonenberg>
Yeah. The WL-PBUS is not but it's a passive cable so i'm not too worried about anything being wrong with it
<bvernoux>
it is fun I have received also some RF stuff from TEl Aviv ;)
<azonenberg>
it's just a 2-3 foot piece of wire with a probus connector at one end and a wavelink on the other
<azonenberg>
so basically if it worked when new and isn't mechanically damaged it's fine
<bvernoux>
it seems it is the 2nd country (after USA) where there is high end / rare RF stuff ;)
<bvernoux>
azonenberg, yes the main issue could be mechanical damage
<azonenberg>
It's from a 100% rated seller who says it's in good shape, and the photos showed no sign of damage to the connectors
<azonenberg>
So i think it's OK
<bvernoux>
yes it is often good I confirm I had only 2 bad things on Ebay ;)
<bvernoux>
where I have bought tons of RF stuff
<bvernoux>
it is a good lesson to do not buy cheap connectors on Ebay anyway ;)
<bvernoux>
except if you want to loose lot of time to test them one by one with microscope inspection ..
<bvernoux>
especially those crappy SouthWest microwave connectors which shall be soldered to have correct performance what a shame
<bvernoux>
even the ones I have checked with microscope which was clearly new/never used they are not good if not soldered
<pepijndevos>
azonenberg, I looked at your schematic for the frontend, and while input is +/-5V, 50 Ohm, the actual input after antialiasing to the IC would be +/-2.5, high impedance, or not?
<azonenberg>
pepijndevos: To what IC?
<pepijndevos>
my theoretical frontend IC that I'm designing
<azonenberg>
pepijndevos: the output to the DAC is +/- 1V differential with a 900 mV common mode range. So basically each leg of the final output swings from 400 to 1400 mV
<Degi>
DAC?
<azonenberg>
to the ADC*
<azonenberg>
The initial attenuator is a 2:1 divider, then the antialiasing filter is a 50 ohm matched path
<azonenberg>
after that you need a 50 ohm terminator to ground either in the frontend IC or as a discrete resistor. on die termination isn't crucial at these frequencies
<azonenberg>
but from a transmission line perspective it's 50 ohm all the way to the first amplifier
<pepijndevos>
ahhh, okay, so the IC still needs 50 Ohm, internal or external
<azonenberg>
Correct
<azonenberg>
it's a low impedance input
<pepijndevos>
and it's +/-2.5V at that point
<azonenberg>
Yeah
<pepijndevos>
got it
<pepijndevos>
I've been looking at high-speed opamp designs all day
<azonenberg>
the reason for that initial attenuation is to allow us to use the offset stage to shift it up to a 5V unipolar rail
<azonenberg>
only the first offset amp is bipolar, the gain stage runs differentially with a 2.5V common mode
<azonenberg>
then the final stage shifts the common mode down to 900 mV which is what the HMCAD1520 expects
<azonenberg>
VoutP - VoutN should be in [-1V, 1V]
<azonenberg>
and (VoutP + VoutN)/2 should always be 0.9V
<pepijndevos>
So there is no reason the IC would have to be +/-2.5V right? It could also be the full +/-5V, or even more attenuated
<azonenberg>
Correct, the initial 6 dB attenuator is just for convenience and making the input protection easier
<azonenberg>
if you look at my frontend circuit i think i specified a +/- 12V non-destruct voltage or something like that?
<pepijndevos>
Except I guess the more you attenuate the more noise you get
<azonenberg>
Correct
<azonenberg>
It's a tradeoff
<azonenberg>
See page 3 of the scheamtic re input protection
<pepijndevos>
And you actually have a negative supply right... which will be funny to design with. Normally the bulk voltage is at the lowest voltage, but I think they have isolated fets specifically for this reason
<azonenberg>
Basically, the series resistor in the pi attenuator limits current through the ESD clamp diode if the input is overloaded
<azonenberg>
But it's only rated for that overcurrent for a short time before overheating
<azonenberg>
so the relay opens up when a fault is detected to protect the frontend
<pepijndevos>
Yea I saw you have a relais that's supposed to shut it down
<azonenberg>
there's a discrete 74xx latch to provide "passive" (well i mean active components, but no software in the loop) safety
<azonenberg>
even if the MCU totally locks up or something
<azonenberg>
overcurrent protection still works
<Degi>
Maybe we could add a fuse too
<pepijndevos>
hmmm so if you do away with the 2:1 attenuation, you have to be careful what happens if you get 20V on the IC suddenly
<azonenberg>
pepijndevos: Correct
<azonenberg>
I think 20V would probably still kill my frontend
<azonenberg>
i forget the exact design voltage limit but it was at least +/- 10V
<azonenberg>
I generally target 12V non-destruct at probe inputs because that's the highest voltage you see in typical wallwart/battery powered hardware
<azonenberg>
i.e. it's OK to malfunction or shut down but there should be no permanent damage
<pepijndevos>
cool, hmmm, lots of thinking and math and simulation to do for the chip design...
<azonenberg>
note also that using attenuating probes provides additional protection
<azonenberg>
if you use a 10x probe you need 50V at the probe tip to see 5V at the scope
<azonenberg>
(although the AKL-PT1 is only rated for about 11V before *its* attenuator will overheat assuming a 5 ohm load at the scope)
<pepijndevos>
but... 10x probes are generally very high impedance or not?
<pepijndevos>
I guess not...
<pepijndevos>
Ok, bedtime here. Thanks for the additional info :)))
<azonenberg>
pepijndevos: The AKL-PT1 is a 450 ohm resistor at the tip then 50 ohm coax to the scope
<azonenberg>
so you have a 50 ohm matched path with a 10:1 division at the front
<azonenberg>
designed to reflect 90% of the amplitude back into the DUT and take 10% for the scope
<pepijndevos>
Yea, i said that and then realised there is no reason why it should be the case. just that my probes are high impedance
<pepijndevos>
What's funny about the rigol RE is that they split the AC and DC path
<pepijndevos>
so they AC couple the high speed part, and then they use an opamp with a very low bandwidth to do low frequencies and offset. Kinda funny approach. Not sure if it's any good...
<pepijndevos>
Like... how flat is the response in the transfer region
<azonenberg>
pepijndevos: i actually considered splitting ac and dc paths in the frontend for the original starshipraider board, now known as CLARKE (more precisely the probes for it, which i never finished)
<azonenberg>
but that was a logic analyzer
<azonenberg>
so flat response wasn't critical
<pepijndevos>
oh, any other interesting trade-offs that made you not do that in the end, or just flat response?
<Degi>
Maybe they do gain control in the ADC
<pepijndevos>
No actually what they do is modulat the tail current in the diffpair
<sorear>
the other question is, if you accidentally hook it up to a 19V laptop charger rail, how expensive is the part you just blew and how much time will you spend replacing it
<pepijndevos>
Yea hopefully just those diodes...
<sorear>
a valid approach is to make the answers to those questions “not very” and “not much”
<pepijndevos>
but you'd like to have it fail in a way that stops it from failing further
<pepijndevos>
So better to blow the series resistor than the ESD diodes
<azonenberg>
pepijndevos: I never finished building it
<azonenberg>
the design i actually built did split
<azonenberg>
but i never debugged it completely
<azonenberg>
and yeah if you blow the diode you will overvolt the offset stage amplifier
* sorear
has been wondering if anything has functionally unnecessary multi-kV caps as an anti-RE feature
<azonenberg>
The resistor will overheat first
<azonenberg>
The dynamic behavior of the protection circuit, including things like time domain transient response to overloads at the amplifier, has not yet been studied
<pepijndevos>
Degi, their gains stage is basically a operational transconductance amplifier. The limitation is that they have a relatively small linear range
<azonenberg>
my plan was to do more extensive measurements of flatness etc
<azonenberg>
and then basically test to destruction
<azonenberg>
apply 10V pulses, if it doesnt die go up to 15V, etc
<azonenberg>
all with a high-Z scope probe after the protection circuit to see what the spike looks like and what the peak voltage is, how fast it's clamped, etc
juli965 has joined #scopehal
<_whitenotifier-f>
[scopehal-apps] electroniceel commented on issue #215: Window title of protocol analyzer windows not shown or imcomplete - https://git.io/JTfIb
<azonenberg>
Degi: speaking of gain control in the ADC
<azonenberg>
i havent played with fine gain on the HMCAD yet
<Degi>
Hmh
<Degi>
I think it could be useful for calibrating gain offsets
<azonenberg>
yes. what i want to do with it in particular is reducing interleaving spurs on ZENNECK
<Degi>
Hm, does the HMCAD have fine phase adjust (in 10 ps range)?
<azonenberg>
I dont think so. But ZENNECK is going to probably use a LMK04806 per ADC as a clock generator
<azonenberg>
So i can tweak phase of each of the four hmcad's to 25ps
<Degi>
For example the EV10AQ190A is the 10 bit version but is much pricier (cheapest 518 $ vs 189 $)
<Degi>
Huh, apparently the EV12AQ605 is 819 $ for QTY 1, it has 12 bit and 6.4 GS/s
<Degi>
Has 4.5 / 6.5 GHz BW, +- 4.5 ps phase adjust with 17 fs step
<Degi>
But it requires a good serdes since it has 12.5 Gbit/s data
Nero__ has quit [Read error: Connection reset by peer]
<azonenberg>
Degi: any high end ADC like that is gonna be jesd204 or similar
<azonenberg>
we're going to need lots of kintex7/ultrascale's to process the waveform data
<Degi>
Hmh, the 5 GS/s one would even be able to be interfaced with an ECP5 over DDR
<Degi>
With lots of lanes haha
<electronic_eel>
azonenberg: about the sense peek probe holder - do you know these dial indicator holders used on lathes and milling machines? have you considered using one of these as probe holders?
<electronic_eel>
I have two of them on my bench and use them for holding cables and similar stuff for soldering
<electronic_eel>
but I was also playing with using them as probe holders
<electronic_eel>
I have a thick and heavy chunk of metal as base and use the magnetic base of the dial indicators to keep them fixed, this works really well
<electronic_eel>
the only thing that could be improved is that they should create some downward pressure even if the lock is screwed tight, to keep the probe tip pressing on the pcb
<electronic_eel>
this wouldn't be necessary if the probe tip were springy
<azonenberg>
Hmm, interesting idea but the holder i ahve for mine is way too big
<electronic_eel>
you mean the thing were you mount the dial indicator in the holder?
<azonenberg>
the holder i mean. the end is too big, i cant imagine trying to put a lot of them in close proximity
<azonenberg>
they'd hit
<electronic_eel>
ah, yeah
<azonenberg>
I think there is potential for some custom holders in the future, and probably adapting third party stuff
<azonenberg>
the sensepeek stuff seems to have potential
<azonenberg>
but i wont know for sure until i get one in my lab to play with
<electronic_eel>
the dial indicator holders basically have a screw at the end, where a 6mm diameter holder is screwed in, that is where you normally put the dial indicator in. I guess when you want to put probes properly into there you would design a 3d printed probe holder that would screw into there
<azonenberg>
oh ok. Not the same design as mine
<azonenberg>
mine has a screw clamp that pinches the outer shaft of the indicator
<azonenberg>
(not the sliding part, the outer shell it slides into)
<electronic_eel>
and that wouldn't need to be directly at the (thicker) end of the holder. it could be longer so many holders in close proximity don't collide
<electronic_eel>
mine has this screw clamp too. but the screw clamp mechanism itself is screwed into the end of the flexible arm
<azonenberg>
Oh ok
<azonenberg>
i didnt look at how mine attached
<azonenberg>
I think there are a lot of positioning gizmos of various sorts, not necessarily sold for EE purposes, that have potential
<azonenberg>
at some point we should do a more formal survey of what's out there
<electronic_eel>
yeah. was just venting ideas around