azonenberg changed the topic of #scopehal to: libscopehal, libscopeprotocols, and glscopeclient development and testing | https://github.com/azonenberg/scopehal-cmake, https://github.com/azonenberg/scopehal-apps, https://github.com/azonenberg/scopehal | Logs: https://freenode.irclog.whitequark.org/scopehal
Degi_ has joined #scopehal
Degi has quit [Ping timeout: 265 seconds]
Degi_ is now known as Degi
bvernoux has joined #scopehal
<bvernoux> re
<bvernoux> I'm searching something better than my AOYUE 618 (old model especially the PCB holder is awfull)
<azonenberg> I use a Stickvise for smaller scale stuff
<azonenberg> i haven't done enough work with large boards to have anything good for bigger
<azonenberg> the currently planned BLONDEL mainboard is large enough that i may need to find something different
<azonenberg> (the other option is to use ribbon cables or something to connect them)
<azonenberg> since making the board excessively large just for a few connectors seems a bit sillly
<bvernoux> yes Stickvise seems great too
<bvernoux> also this one seems perfect for board not too big https://images-na.ssl-images-amazon.com/images/I/514vjXqFEyL._SL1001_.jpg
<bvernoux> but with COVID-19 delivery is expected in June ...
<bvernoux> same for Stickvise :(
<bvernoux> delivery in June
<azonenberg> oof
<bvernoux> also this microscope from Eakin seems very nice https://www.aliexpress.com/item/4000806290901.html?spm=a2g0s.8937460.0.0.47222e0erkp72l
<bvernoux> the price is lower than before
<bvernoux> with up to 2K 30FPS or 1080p 60FPS
<bvernoux> with 10x to 200x Zoom
<azonenberg> imo good stereo eyepieces and optical zoom are the only thing that matters for pcb work
<azonenberg> a camera is only important for taking pictures after you're done, or to shoot training videos
<azonenberg> the former can be done by holding a phone up to the eyepiece
<azonenberg> the non-stereo "cmos camera + magnifying lens" fake-microscopes are worthless
<bvernoux> yes I doubt it is nice to have just the camera for soldering ...
<bvernoux> or maybe both in one using AmScope + later adding Camera
<bvernoux> IIRC AmScope are nice for the price
<miek> 4tronix in the UK say they have stock of the stickvise, should be faster than June
<_whitenotifier-3> [starshipraider] azonenberg pushed 3 commits to master [+2/-0/±6] https://git.io/JvjmU
<_whitenotifier-3> [starshipraider] azonenberg 257868e - Added build for probe manual
<_whitenotifier-3> [starshipraider] azonenberg c08424e - Design files for v0.7 handheld-resistive-probe
<_whitenotifier-3> [starshipraider] azonenberg ca69389 - Updated final v0.7 PCB with soldermask and silkscreen tweaks
<bvernoux> All in one
<bvernoux> and quite cheap about 100Euros
<azonenberg> i think i have one of those preheaters. not a fan, i almost never use it
<azonenberg> nothing against the preheater in particular
<azonenberg> but i just find it doesn't add anything
<azonenberg> and working under a microscope with one is hard
<bvernoux> yes it is not easy to find good compromise ;)
<azonenberg> i find if i need more heat i just blast the area with a hot air rework wand from a distance
<bvernoux> anyway for short term I will buy the PCB glue
<bvernoux> as it is very nice to solder tricky connectors components with both side ...
<azonenberg> fwiw, my samtec SMA's friction fit very well onto a pcb without any glue or anything
<azonenberg> it takes a lot of effort to move one even with no glue or solder
<azonenberg> it seems like it might actually be a slight interference fit
<azonenberg> So i've never had any issues with them moving unintentionally
<azonenberg> the bigger problem is if i slide one on and it's misaligned they're annoying to slip into the right spot :p
<bvernoux> on my side there is no friction at all with about 0.1mm weak under so the connector move alone with hot air
<bvernoux> and with the fact there is less than 0.05mm margin for center pin it is tricky ;)
<bvernoux> so PCB glue will be a must for that
<bvernoux> but the system to lock PCB is awfull ;)
<bvernoux> need to change it with something better
<bvernoux> I will prefer something like that to fix the PCB => https://www.aoyue.eu/aoyue-328-working-platform-soldering-station.html
<bvernoux> but this one is not the best ...
<azonenberg> So it looks like the new probe shell in black nylon 12, with no glass infill, is $10 per unit
<azonenberg> this is using hp's Multi-Jet Fusion process which involves inkjet deposition of fusing agents followed by heating
<bvernoux> ha great
<azonenberg> shapeways's SLS process is unavailable because I have 0.6mm wall thickness around the probe holder
<azonenberg> the minimum is 0.7 mm
<bvernoux> ha bad just for 0.1mm !!
<azonenberg> They also have gray SLA (photopolymer) resin, which produces much smoother surface finishes that in some cases approach injection-molded quality
<azonenberg> and can go down to 0.4mm walls
<azonenberg> But this is $33.87
<azonenberg> I'm contemplating offering this as an additional, premium enclosure option on the high end editions of the probes with the traceable cal
<azonenberg> since, at the markup i'm offering, i can afford more expensive material
<azonenberg> Ordering a single prototype of each to evaluate
<azonenberg> The MJF glass-filled nylon process i used for the first-run prototypes has an 0.75mm minimum
<azonenberg> so that's no longer an option with the new probe body design
<bvernoux> When I think you can have premium anodized aluminum for less thann10USD ...
<bvernoux> but that requires to pay at least 500USD for a mould so it is only interesting if you are planning to do 500units or more
<azonenberg> i don't think this part can be made out of aluminum easily
<azonenberg> long, skinny, with a high aspect ratio hole through the middle
<bvernoux> I have a partner which do everything ;)
<azonenberg> it would have to be extruded then machined or something
<bvernoux> I provide 3D file and they do
<bvernoux> including a nice logo engraving
<_whitenotifier-3> [starshipraider] azonenberg pushed 1 commit to master [+1/-0/±0] https://git.io/Jvjm5
<_whitenotifier-3> [starshipraider] azonenberg 787e92d - Added new probe shell
<azonenberg> If you want to submit a quote just for kicks, go export the solvespace file i just pushed to step or stl or whatever they want
<bvernoux> yes the requires step/stl
<azonenberg> i wont be ready to do a volume order until the 3d printed prototypes are back, for sure
<bvernoux> for quotation
<azonenberg> but it would be nice to have a price just to know
<bvernoux> I can show you results on my custom cases
<bvernoux> it is done in 1 piece
<bvernoux> fully custom
<bvernoux> it is interesting to compare anyway as it cost tooling before to start anything...
<azonenberg> i cannot do a prototype with shapeways aluminum either
<azonenberg> they have a 1mm minimum wall thickness for SLM production
<bvernoux> it is why shapeway is better for test parts as price is a bit expensive but it include all ...
<bvernoux> ha yes I shall check the minimum wall thickness
<azonenberg> same with steel
<bvernoux> ha even with steel it is crazy
<azonenberg> yeah it's actually not pure steel
<azonenberg> it's 60% steel / 40% bronze
<azonenberg> They can do lost wax copper down to 0.6 mm
<azonenberg> it's expensive, $193 for a single prototype of my shell
<bvernoux> Do you have tried 3D metal printing service ?
<azonenberg> No
<azonenberg> i've only done SLS and MJF nylon
<bvernoux> IIRc there are expensive but they can do anything ...
<azonenberg> Bronze can go down to 0.6 mm wall thickness too
<bvernoux> I'm waiting affordable 3D metal printer ;)
<azonenberg> and is cheaper, $137
<bvernoux> it will be amazing if they could be cheap ;)
<azonenberg> that looks like the most affordable option if i want to do a conductive prototype to test EM performance on my design
<azonenberg> Since i'd be worried about coupling noise into it or changing impedances etc
<azonenberg> a conductive test shell is a must before i'd consider using one in prod
<bvernoux> yes
<bvernoux> other option is SLA with conductive paint too
<bvernoux> it works very nicely IIRC
<bvernoux> some do horn antenna with that
<bvernoux> and the caracteristics are nice
<azonenberg> honestly, my plan for the moment is to use MJF nylon for basic small volume, SLA acrylic for high-end low volume
<azonenberg> apply a spray-on antistatic finish to both
<azonenberg> then if i get the demand, switch to injection molding of carbon-doped plastic for higher volume
<bvernoux> yes it is great
<bvernoux> for clean finish you can also apply some special solvent to have smooth surface
<bvernoux> with same material (for higher end version for example)
<bvernoux> as for the cheapest version I think we do not care as that add work/time just to have something which look nicer
<azonenberg> yeah
<azonenberg> if you really want high end
<azonenberg> for $204 shapeways will do 18k electroplated gold on brass :P
<azonenberg> although that won't work with this shell because it has an 0.8mm minimum thickness
<bvernoux> hehe
<bvernoux> the minimum thikness is bad !!
<azonenberg> they can make it out of gold for $3549 :p
<bvernoux> haha
<azonenberg> or platinum for $8843
<bvernoux> crazy
<azonenberg> or sandstone for $11.64
<azonenberg> because who doesn't want their probe made out of rock? :p
<bvernoux> it is not a probe but a jewel ;)
<bvernoux> sandstone could be nice but will break ;)
<azonenberg> ook so let's think. probe pcb sent to fab for quote, probe shell ordered, stencil ordered
<azonenberg> AFE prototype components/stencil here, PCB arriving early next week
<azonenberg> ADC board components/stencil here, PCB still at fab
<azonenberg> scope enclosure and front panel LCD here so i can tinker with board floorplans etc
<azonenberg> VNA ordered
<bvernoux> nice lot of things in //
<bvernoux> I have found a small error in latest probe
<bvernoux> starshipraider\boards\probes\probe-characterization
<azonenberg> TCP stack development was pasued for the past few days but in progress
<azonenberg> ?
<azonenberg> oh, the characterization board? that's en route from fab now, it shipped a couple of days ago
<azonenberg> so too late to fix whatever you found :p
<bvernoux> 3D view display error => Unable to find segment with an endpoint of (29.95mm, 110.9mm)
<bvernoux> it is just minor for the 3D display ;)
<bvernoux> it work even with this warning in fact
<_whitenotifier-3> [starshipraider] azonenberg pushed 4 commits to master [+2/-0/±2] https://git.io/JvjY8
<_whitenotifier-3> [starshipraider] azonenberg 6332964 - Updated ignore properties
<_whitenotifier-3> [starshipraider] azonenberg 9e03867 - Added tip impedance matching simulation
<_whitenotifier-3> [starshipraider] azonenberg d5c04f7 - Updated tip matching simulation
<_whitenotifier-3> [starshipraider] azonenberg 0972f20 - Added preliminary concept floorplan for BLONDEL enclosure/boards
<bvernoux> so latest probe have 6 resistors v0.7
<azonenberg> Yes
<azonenberg> that gave the flattest frequency response
<azonenberg> there is still a small amount of ringing due to ground lead inductance that i'm experimenting with different tip/ground combos to reduce
<bvernoux> yes it is a tricky part
<azonenberg> this was the best result we've had to date
<azonenberg> thin ground blade, short tip needle, on lain and monochroma's 5 GHz lecroy
<bvernoux> yes it is great
<azonenberg> that is the closest approximation to the final board setup we've tested so far
<bvernoux> do you have comparison with other probe like Pico .. ?
<azonenberg> So i have no concerns about calling this a 2 GHz probe
<bvernoux> I'm pretty sure it will be better that my 18GHz Probes ;)
<azonenberg> https://www.antikernel.net/temp/pico921.png this is a picoconnect 921 6 GHz probe
<azonenberg> faster risetime but rings a bit
<bvernoux> -that+than
<azonenberg> (some of that may be the signal source too? need to try other pulse sources)
<azonenberg> But not significantly worse
<azonenberg> and this is a $1k probe we're comparing to
<bvernoux> yes the difference is not huge
<bvernoux> especially for the price
<azonenberg> if i can get even *close* to that result with mine, i'll be very competitive
<bvernoux> it is not far to be 6GHz compliant ;)
<azonenberg> well again these results were on a reworked v0.6 board
<azonenberg> with two actual 0402 footprints, the probe tip bridging a wrong-sized tip footprint and one of the 0402 footprints
<azonenberg> and four more 0402 footprints created with a scalpel or mill
<azonenberg> and the ground in the wrong place compared to where it ultimately will live
<bvernoux> ha ok so v0.7 shall bring some improvements too ?
<azonenberg> That is the hope? shouldn't be worse for sure
<azonenberg> and ergonomics of the tip will be improved greatly
<bvernoux> do you have 3D picture of latest case ?
<azonenberg> (16:03:21) azonenberg: https://www.antikernel.net/temp/probe-shell2.png
<azonenberg> (16:03:23) azonenberg: https://www.antikernel.net/temp/probe-shell3.png
<azonenberg> (16:03:25) azonenberg: https://www.antikernel.net/temp/probe-shell4.png
<azonenberg> (16:15:05) azonenberg: Get your nicotine fix in 50Ω matched clarity with the new 2 GHz vape pen from Antikernel Labs!
<azonenberg> (you can't tell me it doesn't look like one)
<Degi> lol
<bvernoux> very nice
<azonenberg> i may add some smoothing and transitioning to the tip in future versions rather than having it just cut straight across
<azonenberg> some kind of diagonal may be added
<azonenberg> but for now, this is what i will put on beta units
<bvernoux> probe-shell4.png look like a gun ;)
<azonenberg> lol
<azonenberg> For scale the pcb is 1.6mm thick
<bvernoux> yes so the case width shall be less than 2cm
<bvernoux> need to check with the size of PCB in KiCad
<azonenberg> the rectangular center area of the probe is 8x10 mm, then there's an 8mm diameter semicircle at each end
<azonenberg> so i believe 18x8 mm is the bounding box of the probe
<bvernoux> ha very nice and small
<bvernoux> I see PCB is less than 14mm width
<azonenberg> i could make the shell a little bit thinner if needed
<azonenberg> but i'll try this out and see how it feels
<bvernoux> and less than 80mm length
<bvernoux> I was afraid it is too big/long but it is perfect
<azonenberg> yes it's fairly close dimensionally to commercial active probe handpieces
<azonenberg> Like the tetris
<bvernoux> yes they are very nice too
<miek> is there enough leeway to thin out one side and centre the probe tips?
<azonenberg> there is a little bit of play but i don't think quite that much. Right now the centroid of the PCB+component bounding box is centered
<azonenberg> the reason for this is to balance out wall thickness
<bvernoux> TETRIS Probe cost is > 1200 pounds !!
<azonenberg> Right now i have 0.6 mm minimum on either side where the positioner grabs it
<azonenberg> the minimum supported wall thickness for this process is 0.4 mm
<azonenberg> so i could, i guess, move the centerline 200 μm lower
<azonenberg> but that's a hard limit and i don't see it being necessary/helpful
<azonenberg> the off-center-ness of the probe tip is a fraction of a millimeter
<azonenberg> but again this is the sort of feedback i will be looking for from beta users
<azonenberg> ergonomics are going to be tweaked
<bvernoux> Out of topic: Interesting comparison of Qucstudio vs ADS2016 => https://csc-rf-electronics.blogspot.com/2020/03/qucs-studio-vs-ads-2016-performances.html
<Famine-> morning folks
<Degi> hi
<Degi> Whoops my laptop OS crapped itself after an update
<Degi> From 5.5.3 to 5.6.3 kernel...
<Famine-> Degi, how about a trade, i'll fix your laptop if you read 200 pages of tax law for me lol
<Degi> Nvm fixed, probably cause no lvm2 hook in mkinitcpio
futarisIRCcloud has quit [Quit: Connection closed for inactivity]
<lain> arch?
<Degi> Yeh
<Degi> Somehow the mkinitcpio had no lvm2, now it works though
<Degi> Huh there's yellow spots all over the balcony from a chemical fuckup a while ago... I think its iron chloride leftovers.
<lain> o.o
bvernoux has quit [Quit: Leaving]
Famine- has quit [Read error: Connection reset by peer]
<Degi> In retrospect, I should've put a ATMega32u4 instead of 328 onto the PCIe adapter. Would work as USB to Serial bridge
<Degi> Hm I have a I think well preserved HDD head here... Maybe I could try to design a magnetic probe around that, that'd be fun heh
<azonenberg> it may not be as well preserved as you think
<azonenberg> from what i hear hdd heads are SUPER esd sensitive
<Degi> huh
<azonenberg> like hundreds of volts esd will kill them
<azonenberg> vs modern ICs are typically esd protected to high single or low double digit kV
<Degi> Hm I didnt extract it yet its attached to the arm
<Degi> I think the ECP5 SERDES also only has like 250 or 500 V lol
futarisIRCcloud has joined #scopehal
<lain> fun fact: many (most?) autofocus camera lenses have magnetically-encoded position data along the focusing group barrel, which is read by a GMR sensor (same as a hdd head) to detect the focus position
<Degi> Is it possible to buy RF GMR sensors somewhere?
<lain> this is important to know when repairing lenses, because touching that sensor is death to it, and they aren't easy to find from what I hear
<Degi> Hm ok this one is probs dead too
<azonenberg> so it looks like perpendicular-to-plane GMR sensors can survive "much higher" ESD voltages
<azonenberg> i.e. 160V
<azonenberg> while 48V ESD will kill a current-in-plane sensor
<azonenberg> apparently ESD is the main source of yield loss
<Degi> huh
<azonenberg> 48V ESD is so tiny you wont even get *close* to feeling it
<lain> yeah wow lol
<monochroma> 12V ESD zap zap :3
<lain> 1V ESD good luck powering it :<
<azonenberg> lain: so i think i am going to spec the probe for 0-95C absolute max
<azonenberg> and 15-45C recommended
<lain> that sounds very reasonable
<azonenberg> then specify tolerances for gain, resistance, etc based on the 15-45C range
<azonenberg> with the cal certificate issued at about 23C or so