azonenberg changed the topic of #scopehal to: libscopehal, libscopeprotocols, and glscopeclient development and testing | | Logs:
Construct has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
Construct has joined #scopehal
Construct has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
electronic_eel has quit [Ping timeout: 240 seconds]
electronic_eel has joined #scopehal
Construct has joined #scopehal
Degi_ has joined #scopehal
Degi has quit [Ping timeout: 265 seconds]
Degi_ is now known as Degi
Construct has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
<azonenberg> Analog Devices samples arrived
_whitelogger has joined #scopehal
futarisIRCcloud has quit [Quit: Connection closed for inactivity]
deltab has quit [Ping timeout: 240 seconds]
deltab has joined #scopehal
promach3 has quit [Ping timeout: 244 seconds]
jevinskie[m] has quit [Ping timeout: 246 seconds]
promach3 has joined #scopehal
promach3 has quit [Ping timeout: 265 seconds]
juli966 has joined #scopehal
jevinskie[m] has joined #scopehal
futarisIRCcloud has joined #scopehal
promach3 has joined #scopehal
bvernoux has quit [Read error: Connection reset by peer]
<azonenberg> So the new AKL-AD1 design should be reasonably flat to around 4 GHz
<azonenberg> and have approximately unity gain end to end
_whitelogger has joined #scopehal
Construct has joined #scopehal
<azonenberg> What do you folks think of using a 5-turn trimpot instead of a DAC for zeroing the output of the probe?
<azonenberg> at least in the first iteration it will greatly simplify things
<azonenberg> in particular, there don't seem to be any 16-bit DACs on digikey capable of running on a +/- 2.5V rail (i need bipolar because the output offset could be in either direction)
<azonenberg> So i'd need to run the DAC on +/- 5V which means two more power rails
<agg> it's pretty familiar to have a trimpot on a probe to tune it, doesn't seem like a big deal
<azonenberg> well i guess the bigger question is whether the amp offset is going to drift with temperature etc
<azonenberg> in which case dynamic tuning might be necessary
<azonenberg> For now i'll just do a static calibration and see
<monochroma> digi-pots?
<azonenberg> They only go out to about 8 bit resolution
<azonenberg> with a 5V span that's 19 mV per step
<azonenberg> Which is on the order of the offset i expect to be trying to calibrate out :p
<azonenberg> i need sub mV
<Bird|otherbox> azonenberg: take a single-supply 16bit DAC and run it off the +/-2.5V rails? you'd need a bit of digital level shifting but I don't see why it'd be particularly impractical
<azonenberg> Bird|otherbox: Yeah i'd need to do some messy shifting to move the LVCMOS25 MCU SPI bus down to be referenced to -2.5V instead of ground
<azonenberg> that doesn't seem any less work than using a bipolar +/- 5V DAC and throwing down two more LDOs
<azonenberg> the point is more that all of the obvious solutions are messy and add more components
<azonenberg> a pot is the simple option and as long as i can calibrate once and not touch it again, it seems the way to go
Construct has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
<azonenberg> Bird|otherbox: , monochroma, agg, lain: ^
<azonenberg> david.lenfesty: ping, just checking in on status of your probe stuff
<d1b2> <david.lenfesty> pcbs are ordered, I've got most of the drivers for the FUSB302 and USB PD stuff done, and have to flesh out the rest of the application.
<azonenberg> PCBs for both sides are ordered? Great
<d1b2> <david.lenfesty> yeah I was mostly trying to get them ordered before the lunar new year 😛
<azonenberg> let me update my spreadsheet to reflect that then
<_whitenotifier> [starshipraider] azonenberg closed pull request #5: Active probe testing PCB -
<_whitenotifier> [starshipraider] azonenberg pushed 5 commits to master [+16/-0/±14]
<_whitenotifier> [starshipraider] davidlenfesty 5686a68 - active-probe-test: finished base schematic TODO is part selection and layout
<_whitenotifier> [starshipraider] davidlenfesty ae7ca85 - active-host-test: switched 0u1 caps from 0603 metric to imperial
<_whitenotifier> [starshipraider] davidlenfesty 1ebb6b6 - active-probe-test: finish component selection Also fixed some issues with components on the host
<_whitenotifier> [starshipraider] ... and 2 more commits.
<azonenberg> also just realized i never merged that
<azonenberg> So your current host board takes +/- 7V from a bench supply?
<azonenberg> no regulation onboard?
<azonenberg> I'm thinking what i might want to do once your firmware etc is done for this test board is to build a standalone host box
<azonenberg> i.e. not to be integrated into BLONDEL
<azonenberg> probably doable with a slightly different firmware for the host side MCU
<azonenberg> basically i want to be able to build a box that runs off a wallwart and has usb-c ports for up to four probes
<azonenberg> providing power and no control, or maybe applying static hard-coded config or something
<azonenberg> the idea being to allow the AKL-AD1 to be powered over usb while used with a COTS scope
<azonenberg> make sense?
<azonenberg> and then obviously the companion project would be doing a variant of the AKL-AD1 that's closer to production form factor with USB power instead of bench supply etc
<azonenberg> My tentative thought is to have the form factor be L-shaped
<azonenberg> in that you have the probe input on the right and USB+SMA on the top
<azonenberg> The thought being that the LeCroy probes are kinda awkward in the positioner with the cable being jammed against the bench and fed in at a weird angle when it's got the input pointing up
<azonenberg> so if the signal path made a 90 degree bend we wouldn't have that issue
<azonenberg> basically i'm proposing the SMA and USB connectors about where the pink light is on this probe
<azonenberg> and then the "tail" end would be flat, no wires
<azonenberg> I also think a more rectangular form factor makes more sense for the amplifier module
<azonenberg> rather than round
<azonenberg> it would allow you to pack them in close together on a ganged holder if you so desired
<azonenberg> and the amp isnt something you'd be hand holding anyway
<d1b2> <david.lenfesty> Currently my host design has full on board regulation. Takes input from a 12V barrel jack. I'm still not super comfortable with designing switching supplies but worst case I've got lots of testpoints and I can bring everything up w/o them. For a standalone host box yeah sounds good. Should only require pretty trivial modifications to FW the way I'm trying to write it. Might make sense to expose another USB port and just have a dumb CDC-ACM
<d1b2> serial port for changing settings, especially because the F103 has USB already.
<azonenberg> oh great
<azonenberg> and yeah i'm not sure on details but i figure if i can have probes that are usable standalone
<azonenberg> it would be a nice thing i could start assembling and selling to fund R&D on the scopes
<azonenberg> And an open hardware 4 GHz active differential probe i think is something that would be pretty attractive
<d1b2> <david.lenfesty> I glance over at my 100MHz Rigol
<d1b2> <david.lenfesty> wait 50, I still haven't gotten around to patching it
<azonenberg> There's really two reasons i am targeting midrange first instead of low end
<azonenberg> The first is that it's easier to compete than with the likes of rigol which is cost optimized to hell and back
<azonenberg> but the second is that it's more popular with business type customers that have more budget
<azonenberg> Which means i can sell them and use funds from the sales to subsidize R&D on the entry level models
<azonenberg> it'd be a lot harder to fund a high end design off sales to hobbyists
Construct has joined #scopehal
<d1b2> <david.lenfesty> Makes sense, I'm just amused by the difference in capabilities between what I have, and the functionality I'm trying to enable with this project 😛
<azonenberg> well hey, maybe we can have an active probe for your rigol soon
<azonenberg> the current ones are focused on bandwidth but a multi-megohm impedance active probe would be nice for looking at crystal oscillators among other things
Construct has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
<Bird|otherbox> yeah, no doubt
<azonenberg> it's just that my dev efforts are shaped by my needs
<azonenberg> And i spend a lot of time looking at fast 50/100 ohm lines
<azonenberg> and relatively little time looking at high-z low speed stuff
<azonenberg> in fact, most of my designs don't have a single crystal on them
<azonenberg> i tend to use oscillators
<azonenberg> so while there is a crystal internally, the high-z node is not exposed anywhere i could possibly probe
<azonenberg> and if i do need to probe one high-z signal i have my ZS1500
<azonenberg> on the other hand, i only have two active diff probes total
<azonenberg> and it's very likely i would want to probe >2 high speed signals
<azonenberg> If anybody wants to work on a high-z active probe with 100+ MHz bandwidth though, i'm all for it
<Bird|otherbox> yeah, the unfortunate part is that I don't believe that canned 32kHz oscs are a thing?