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: 250 seconds]
Degi_ is now known as Degi
marcos has joined #scopehal
marcos is now known as Guest48020
futarisIRCcloud has quit [Quit: Connection closed for inactivity]
Degi has quit [Ping timeout: 240 seconds]
Guest48020 has quit [Quit: Konversation terminated!]
Degi has joined #scopehal
balrog has quit [Quit: Bye]
balrog has joined #scopehal
zkms has joined #scopehal
juli963 has joined #scopehal
futarisIRCcloud has joined #scopehal
juli963 has quit [Quit: Verlassend]
juli963 has joined #scopehal
_whitelogger has joined #scopehal
hlzr has quit [Quit: Connection closed for inactivity]
balrog has quit [Quit: Bye]
balrog has joined #scopehal
bvernoux has joined #scopehal
<bvernoux> hi
<azonenberg> o/ bvernoux
<bvernoux> hehe fun with latest episde of TSP https://www.youtube.com/watch?v=AdVlSvVIGwM
<azonenberg> i now have a nist traceably cal'd bench DMM
<azonenberg> And i just cal'd one of my PSUs to it https://www.antikernel.net/temp/IMG_20200424_122513.jpg
<bvernoux> the famous cheap amazing Preselector is hard to find on Ebay
<bvernoux> ha great
<bvernoux> I suspect they have just verified everything pass and nothing was changed ?
<azonenberg> Yes
<azonenberg> i bought the PSU 3 years ago, never cal'd it
<azonenberg> and it's still *well* within tolerance
<azonenberg> R&S makes good stuff that doesn't drift much
<bvernoux> as I really doubt PSU need calibration each year especially in stable environment with 20 to 25°C ;)
<bvernoux> But it is a buisness
<azonenberg> This is my own internal cal
<bvernoux> ha ok
<azonenberg> the DMM had enough significant digits i wanted to get a proper cal on it
<bvernoux> you have a good decade box ultra accurate for that ?
<bvernoux> with some voltage ref ...
<azonenberg> no i used the DMM as a reference standard for the PSU
<azonenberg> run the PSU at constant voltage/current, measure actual output with the DMM
<azonenberg> then compare to what the display on the PSU says
<bvernoux> yes nice and cheap compared to do that by R&S ;)
<bvernoux> I have received my silicon carpet for soldering ;)
<bvernoux> very good for the price
<azonenberg> also VNA is on a delivery truck a few blocks from me
<azonenberg> expecting it in the immediate future
<bvernoux> ha great
<bvernoux> I'm interested by a review of it
<bvernoux> do you have some reference passband, low pass, high pass or even attenuator to check it ?
<bvernoux> the ones from Mini Circuits are very good with SxP files and not too expensive for the quality
<azonenberg> I have a few filters and attenuators i can test it against
<bvernoux> yes nice
<bvernoux> I need to buy more filters ;)
<bvernoux> I'm searching some digital attenuator too
<bvernoux> very good but they are crazy expensive
<bvernoux> and the must will be to have a nice preselector from DC to 26.5GHz ;)
<bvernoux> the one from here is amazing https://www.youtube.com/watch?v=AdVlSvVIGwM
<bvernoux> Agilent 8445B Auto Preselector
<bvernoux> to be hacked like in this TSP episode
<bvernoux> but the one I was checking was sold after TSP episode
<bvernoux> and the crazy things a very old one have the price x2 now ;)
<Degi> https://www.ebay.com/c/1281734245 doesnt seem that pricy
<bvernoux> before it was 90USD ;)
<Degi> If you wanna make your own, search for yig filter
<bvernoux> but it is a bit too big
<bvernoux> anyway it is a must the filter is amazing
<bvernoux> it was something which cost 10KUSD in old day
<bvernoux> what a shame they have destroyed SMA & connector
<Degi> I mean the connector itself on the thingie is fine but they couldve just unscrewed it
<bvernoux> yes
<Degi> Do they sell the loose wires separately?
<bvernoux> It seems the guy selling such things do not know what is it and destroy connectors which cost lot of money
<bvernoux> I have already seen that with some SA too
<azonenberg> bvernoux: no i dont think that's what it is at all
<bvernoux> destroying the front connectors
<Degi> SA? Spectrum analyzer?
<bvernoux> yes Spectrum Analyzer
<Degi> Like the cables or the SMA/N/BNC itself?
<bvernoux> yes ;)
<bvernoux> also on VNA
<azonenberg> a lot of organizations, especially in academia, need to use equipment until it breaks and then throw it out
<bvernoux> it is even worse as they are tuned for it
<Degi> the cables? Atl east thats easily fixable
<azonenberg> they cannot, per stupid policies, give it away or sell it
<azonenberg> if they need to retire it while it still works, it has to be destroyed
<Degi> Oh thats where I got a bunch of stuff from
<bvernoux> azonenberg, yes it is same for big company :(
<bvernoux> they prefer to destroy than to sell
<Degi> You can just fish working stuff from trash under physics building
<Degi> Once I saw 3000 V battery
<azonenberg> so power cords lopped off, etc is standard practice
<azonenberg> I dont think this is carelessness, i think it's deliberate destruction and the seller dumpster-dived this
<bvernoux> yes
<bvernoux> it was the same at Safran DS
<bvernoux> they have order to destroy
<azonenberg> Which is nonsensical. I can understand zeroizing it, removing hard drives, etc
<Degi> At my uni they arent allowed to sell and the equipment is carelessly dumped into the dumpster (sometimes typewriters) but often in good condition. Like this 520 MHz frequency counter had attachment cable, no scratches etc
<azonenberg> most companies that sell products to DoD etc actually have a section in the manual describing how to wipe it before servicing or disposal
<azonenberg> Degi: yeah i dont understand all of these wasteful practices
<azonenberg> like yes it makes accounting a bit simpler as far as not needing to calculate how much you sold it for
<azonenberg> but...
<Degi> Sometimes they put the stuff indoors if its nice stuff and then write a mail to some dept
<Degi> I think they arent allowed to sell cuz its bought with taxes
<azonenberg> like i said, stupidity of policy
<azonenberg> as a taxpayer, i would want to see the opposite
<bvernoux> yes but I doubt that will change :(
<azonenberg> public funds shall be spent as efficiently as possible, rather than taking more tax money to fund your facility
<azonenberg> if you have serviceable equipment you no longer need, you are legally *obligated* to sell it to the highest bidder
<bvernoux> yes it will be great policy
<azonenberg> and use the proceeds of such sale preferentially, then tax/grant funding only when surplus resale funds are exhausted
<bvernoux> and avoid at maximum to put things in trash
<azonenberg> exactly
<bvernoux> all my hw are old things ;)
<bvernoux> except my Picoscope ;)
<bvernoux> and my Soldering Station haha
<bvernoux> the only bad things about old HW is they are big, loudy, heavy and I'm afraid it will fail each day
<bvernoux> as there is no warranty at all and potentially it could be hard to repair ...
<azonenberg> Yeah
<azonenberg> But it's so cheap you can buy two and use one for spare parts for the other :D
<bvernoux> but anyway still really better than new ones which are without any schematic ...
<Degi> Poor TDS520, outdoing the SRAM battery by 15 years and now something is broke
<azonenberg> that's what lain and monochroma did with their 5 GHz scopes (how many do you have total? two? three?)
<Degi> How many 5 GHz scopes do you have... lol
<bvernoux> but old HW are still expensive ;)
<bvernoux> without any warrantu
<bvernoux> waranty
<bvernoux> yes
<azonenberg> Looks like old tek gear all right
<Degi> TDS784 looks nice, is there any scope you could recommend before this is finished?
<Degi> Huh 33 GHz scope under 100 k
<azonenberg> i see a TDS8000 50 GHz sampling scope for $10K
<azonenberg> wonder what that cost new
<Degi> For 1.8 k
<azonenberg> oh
<azonenberg> that's the mainframe only, no capture modules
<Degi> With 1 aq module
<azonenberg> The $10k i was looking at
<Degi> Yes I see that too, that has nothing in tehs lots
<azonenberg> tek CSA8000 w/ 30 GHz sampling module for $6K, not bad
<Degi> Ahy es these with the pricy samle modules
<Degi> Hm do all mayor scope companies use custom ADCs?
<azonenberg> Depends on how major, and how high end the scope is
<azonenberg> LeCroy at least has used COTS ADCs on their lower and midrange stuff
<azonenberg> the upper end i think is all custom
<Degi> What is midrange? 1 GS/s?
<azonenberg> lol no
<Degi> 10? 20?
<azonenberg> midrange meaning like the HDO series
<azonenberg> yeah stuff along those lines
<azonenberg> the HDO4000 used TI COTS ADCs afaik
<azonenberg> i saw a teardown, i think it was the 4000
<azonenberg> Rigol uses the 8-bit version of the BLONDEL ADC
<Degi> With like 10 GS/s?
<Degi> Oof they couldnt be bothered to 12 bits...
<azonenberg> The HDO4024 is 200 MHz 2.5 Gsps
<Degi> Should I buy this? Kinda need a thing thats faster than a voltmeter lol... https://www.ebay.de/itm/Lecroy-LC534M-Oszilloskop-1GHz-2Gs-FFT-ohne-Floppy-Laufwerk-Ubernahmegarantie/293040302412
<azonenberg> 12-bit from Rigol? not happening :p
<Degi> Like its just 50 bucks more
<azonenberg> Degi: i have no idea about PC interfacing capability on pre-Windows LeCroy gear
<Degi> Hm...
<azonenberg> any LeCroy running MAUI or X-Stream should work with scopehal
<bvernoux> Degi, I think it is better to spend money on Rigol scope than this old obsolete 1GHz scope ;)
<Degi> WEll okay
<Degi> Meh maybe I
<bvernoux> MSO5072
<Degi> 'll wail in indecisiveness till blondel is finished
<bvernoux> you hack to 350MHz ;)
<azonenberg> Degi: if you have a bit more budget, more recent secondhand lecroy gear is very readily available on ebay as well
<bvernoux> it is 8GSPS
<bvernoux> unbeatable for 909USD
<bvernoux> FW is a bit buggy IIRC
<azonenberg> i see a DDA-5005A (5 GHz bw, 20 Gsps realtime, 200 Gsps equivalent time) for $1950
<azonenberg> monochroma got one for a bit less i think
<azonenberg> those run windows xp iirc and are known to work with scopehal
<bvernoux> Degi, take MSO5074 for 999USD it is better to have 4 chan
<Degi> I see a broken one for 1.2 k
<bvernoux> the only advantage of those Rigol MSO is they are fully hacked
<bvernoux> so for hobby it is perfect
<Degi> The 5005A seems rly nice but seems to come all from USA... So 2k for that vs 900 for rigol...
<bvernoux> with all free options
<Degi> Wait What I need to pay 587 USD entry fees...
<bvernoux> Degi, Main issue is if you have any issue with those old obsolete scope it will cost more than 2K ;)
<Degi> Good point, rigol would have 2 years warrantly
<bvernoux> and with windowsXP ...
<bvernoux> HDD will die ...
<Degi> HDD?
<Degi> Huh
<bvernoux> I really think it is very risky to buy such old scope
<bvernoux> and they have no option
<Degi> Lol like the TDS520 with 5 year battery backed SRAM
<bvernoux> yes the battery die and the things is not usable anymore as it loose calibration ...
<Degi> Hm so 70 MHz MSO5072 can be hacked to 350 MHz?
<bvernoux> yes and even more in theory ;)
<bvernoux> it is 8GSPS ADC
<bvernoux> 200Mpoints also
<bvernoux> all is SW
<Degi> And 4 channels too?
<bvernoux> take the 4 chan version as I think 2 chan have only 2 chan ;)
<bvernoux> so 999USD ;)
<Degi> Where do I buy that
<bvernoux> for 4chan
<bvernoux> it depends where you live
* azonenberg didn't know 4chan sold oscilloscopes
<bvernoux> from EU rigolna
<Degi> Whas up with all these 2 channel scopes
* Degi stacks 10 of blondel
<bvernoux> 2chan is too limited ;)
<Degi> 80 channel isnt :P
<bvernoux> especially when 4chan cost just 80USD more
<azonenberg> Lol
<bvernoux> anyway I have an old RIGOL DS1102E with 2 Chan ;)
<bvernoux> my 1st scope
<bvernoux> and it still work nicely
<bvernoux> for basic stuff ;)
<azonenberg> My DS1102D failed a year or so ago
<azonenberg> would turn on then randomly shut off, probably PSU issue
<azonenberg> i gave it to somebody whitequark hooked me up with in seattle, no idea if she ever got it working or not
<bvernoux> 100MHz 1GSA/s with 2chan is clearly limited but it depends what you are doing ...
<azonenberg> and yes that's what i'm looking forward to about BLONDEL. Having lots of slow channels i can use in combination with my higher speed scopes
<bvernoux> it is not clear if the 2Chan have 4Chan by default disabled by SW ;)
<Degi> So the 5000 series is all the same hardware?
<bvernoux> I have just doubt on the 2chan version
<bvernoux> I know for other stuff all is same
<bvernoux> memory, BW, options
<bvernoux> all is SW
<bvernoux> the hack is trivial ;)
<bvernoux> just change a file to say all option ;)
<Degi> Hm is there a tutorial or do I need to scan that thread?
<bvernoux> ha yes 2chan have 4chan in fact but comes with 2 probes ;)
<bvernoux> so not good ;)
<bvernoux> as you shall buy the 2 other probes after ...
<Degi> Ok if the probes are any good ill probably just get the 4 chan model
<Degi> Hm is that kinda the best scope you can get for under 1 k?
<bvernoux> I think yes
<Degi> Why are the cheap scopes only 4 channels anywasy..
<bvernoux> it is fun there a new 5000
<bvernoux> do not buy it ;)
<bvernoux> 999USD
<bvernoux> without any option/upgrade ...
<bvernoux> and 4GSA/s
<Degi> Huh it has less GS/s and BW?
<bvernoux> yes
<bvernoux> I do not understand strategy of Rigol ...
<Degi> Sell to companies I guess
<bvernoux> hehe they provide free decoders
<Degi> Does the 999 $ version come with logic probes and BNC probes?
<bvernoux> I doubt there is logic probes
<bvernoux> but IIRC someone have hacked them ;)
<Degi> And I can just stream that scope to scopehal?
<bvernoux> good question
<bvernoux> Digital Chan are optional
<Degi> I mean the analog chans
<bvernoux> 399USD Digital Prob kit
<bvernoux> Logic Probe for MSO5000 Series Oscilloscopes. Includes cable, leads and grabber clips
<Degi> The 999 $ HV probe kit looks nice but you can probably diy that cheaper and better
<bvernoux> I just think it is unbeatable with the spec and 10inch screen ;)
<bvernoux> evne if there is some quirks in FW
<bvernoux> I suspect they are solved now
<bvernoux> there ethernet alsop
<bvernoux> which is nice
<bvernoux> and HDMI ;)
<bvernoux> for 999USD ;)
<bvernoux> I will probably buy it too
<bvernoux> but I need something >1GHz BW ;)
<Degi> Heh what arey ou planning to measure?
<bvernoux> RF stuff ;)
<bvernoux> I need more a working SA
<bvernoux> than a very fast scope today
<bvernoux> but a fast scope is always nice especially something with 20GSPS or more ;) but so expensive
<bvernoux> MSO8000 are nice but also expensive
<bvernoux> 8KUSD min ;)
<bvernoux> at that price I prefer to buy other brand than Rigol
<Degi> Maybe a LC534M
<bvernoux> I like the Tektronix MDO4104C-SA6
<bvernoux> especially the fact it has a 6GHZ SA linked/synchronized to other scope chan
<bvernoux> but refurbished version cost is 14KUSD ;)
<Degi> I mean this scope looks good and costs about half of the rigol but it probably has no warranty and no way to easily interface with PC and low sample depth... https://www.ebay.de/itm/Lecroy-LC534M-Oszilloskop-1GHz-2Gs-FFT-ohne-Floppy-Laufwerk-Ubernahmegarantie/293040302412
<bvernoux> yes too old
<bvernoux> with unreadable screen
<Degi> Do rigols have some factory calibration?
<Degi> Heh unreadable?
<bvernoux> and 2GSPS
<bvernoux> it is clearly not 1GHZ ;)
<Degi> Why not
<bvernoux> because 2GSPS is maximum 200MHz
<Degi> Hm why would that be?
<Degi> Like the broken scope here has 500 MHz at 500 MS/s...
<bvernoux> as a good point is to divide by 10
<bvernoux> even if niquist say /2
<electronic_eel> Degi: if you want to buy Rigol in Germany, go to batronix
<bvernoux> it is better to have 10points ;)
<Degi> 2.5 should be enough with a brick wall filter
<Degi> thx
<Degi> Huh they have fancy pcb holders
<bvernoux> yes or 250MHz but not more ;)
<Degi> I mean the 500 MS/s scope actually seems to have 500 MHz, now its broke...
<bvernoux> or doing sampling trick but it is not a scope
<bvernoux> no
<bvernoux> 500MSPS is max 50MHz
<bvernoux> some say it is ok to 100MHz
<Degi> It can do ET sampling to 100 GS/s
<Degi> But the actual ADC is 500
<bvernoux> yes for ET sampling
<Degi> Tbh in some cases it makes sense to have more bandwidth than sample rate, for example if you want to know the signal value accurately at a certain position
<bvernoux> but it is not scope mode ;)
<Degi> rigolna seems to be a bit cheaper...
<bvernoux> check VAT price too
<Degi> Well then its a raw adc idk
<bvernoux> IIRC it is not included
<Degi> Oh
<Degi> Hm rigolna wants you to contact them for a quote if youre from not USA
<electronic_eel> Degi: if you have some time and want to lower the price a bit, use the lowest-competitor-offer of batronix: search around a bit for any lower price offer than batronix in europe. then send them a link/screenshot of the offer and they will give you the same price
<electronic_eel> I got my SA like 200 EUR lower because of that, there was a sellout offer at some dealer in sweden, I used that offer
<Degi> In EU?
<electronic_eel> and bought at the lower price at batronix
<Degi> Its currently 962 € with VAT and 809 without
<bvernoux> I doubt there is such offer for the MSO5000
<bvernoux> but it is a good point
<Degi> Everybody else is at 962 € too even conrad
<Degi> Wait thats the 2 ch variant...
<bvernoux> I want the RSA3000 for that price ;)
<electronic_eel> na, forget the price search engines, you got to invest a bit of time to search for the smaller sellers
<bvernoux> upgradable to 7.5GHz by SW ;)
<bvernoux> Degi, the good things with the 70MHz version is probes are 350MHz ;)
<bvernoux> as they have only 1 type of probes for all variant ;)
<Degi> lol
<bvernoux> so all the HW are identical ;)
<bvernoux> I see batronix => € 1,069.81 incl. 19% VAT
<electronic_eel> bvernoux: they might have just one type of probe, but this one type is shitty
<bvernoux> electronic_eel, it works up to 350MHz anyway it is a passive probe
<bvernoux> why do you say they are shitty ?
<bvernoux> compared to other brand I will say of course ;)
<Degi> Hm 5074 is 1069 € with VAT on batronix, batterfly has it for 1068 €, arrow for 1018 but I think thats without VAT, eleshop.eu has it for 1050 € with protocol analysis bundle
<electronic_eel> it works somehow, but attaching proper ground clips is not that easy. nearly no accessories. also high capacitance
<Degi> Is protocol decoder bundle included in the hack?
<bvernoux> it is 3 years warranty not 2 ;)
<bvernoux> hehe
<bvernoux> it is first time I see that
<electronic_eel> I know some guy, I think he is called azonenberg or similar, he currently develops better probes ;)
<bvernoux> electronic_eel, ha yes of course ;)
<Degi> 1049 CHF which is 996 EUR on toppreise.ch but not sure if vat included
<bvernoux> electronic_eel, but for scope probe they are correct anyway of course they are not ultra low capacitance it is just standard probes
<bvernoux> and Rigol does not have 50Ohm input by default ;)
<bvernoux> not on their scope
<Degi> Do they have selectable 50 ohm?
<electronic_eel> the MSO5000 does not have 50 ohms?
<bvernoux> I doubt even the MSO8000 have such feature
<electronic_eel> I think they have, just the lower ones don't
<electronic_eel> my MSO4000 has 50 ohms
<bvernoux> to be checked
<bvernoux> ha ok
<electronic_eel> wouldn't have bought it without
<Degi> Should I buy this instead? https://eleshop.eu/rigol-mso5074.html
<bvernoux> Degi, yes and ask freebundle ;)
<electronic_eel> I disliked the standard rigol probes so much that I bought the RP5600A probes
<bvernoux> as anyway all reseller shall have the offer with free bundle all options free
<bvernoux> for serial decoders ...
<electronic_eel> but a full set gets pricy
<bvernoux> ha it is included too
<Degi> Geez 400 bucks for a probe
<electronic_eel> yes, now you know why I'm interested in azonenbergs passive probes
<Degi> Indeed
<Degi> Is that an active probe?
<electronic_eel> no, passive 600MHz
<Degi> lol
<electronic_eel> now, go look what passive tek probes in that category cost
<Degi> 64 bucks for 200 MHz
<Degi> Hm any upside to buying from batronix than eleshop?
<bvernoux> amazing
<bvernoux> Siglent have now a VNA up to 7.5GHz
<bvernoux> a bit expensive >8K
<bvernoux> last time i checked their VNA was limited to 3GHz
<electronic_eel> Degi: I don't know eleshop, but bought several times from batronix. batronix was fast and nice to work with
<bvernoux> the SVA10xxX have amazing spec too
<Degi> Like on eleshop they have protocol decoder option and 3 years warraants
<electronic_eel> also batronix is in germany, so lower shipping costs if you have to return something or similar
<electronic_eel> Degi: if you see some better offer somewhere, batronix will match that
<Degi> Even things like extra features?
<electronic_eel> just send them the link or screenshot
<electronic_eel> I think so
<Degi> How do I do that? Click request offer?
<Degi> Nvm they have options bundle too
<electronic_eel> send them an email, add your postal address, and they will send you a full offer
<bvernoux> does anyone have bought a Siglent SVA1032X ?
<Degi> Nah I noticed that batronix actually has the better offer
<bvernoux> It will be interesting to know if it is same HW vs the SVA1075X
<electronic_eel> hmm, didn't they introduce the 3ghz version first and the 7.5ghz version came later?
<bvernoux> yes
<electronic_eel> that would indicate that the hardware is different
<bvernoux> ha ok Dynamic Range is a bit low
<electronic_eel> otherwise they would have started with the 7.5
<bvernoux> between 60dB to 90dB ;)
<bvernoux> so less than my HP 8753D ;)
<bvernoux> electronic_eel, I suspect they have designed first the 7.5Ghz then the do different version and all is SW ;)
<bvernoux> it is more economic for the production cost ...
<bvernoux> it is the same for other vendor like Agilent ;)
<electronic_eel> from a marketing standpoint this is not how you do it. you release with a big bang and the top model to get all the pr
<bvernoux> SA E440xB
<bvernoux> they have same hw on lot of part
<bvernoux> electronic_eel, yes it is why I ask as maybe the 7.5Ghz is different
<bvernoux> woo sweep points are very limited
<bvernoux> 101 to 751 !!
<Degi> huh?
<bvernoux> I always use 1601 ;)
<bvernoux> yes on the VNA ;)
<electronic_eel> I think they started with the 3ghz. they probably planned to do the 7.5, but couldn't get it to work from the start. then they released the 3ghz to get to market fast
<bvernoux> I'm reading spec of SVA10xxX
<Degi> RAM is pricy
<bvernoux> Degi, it does not consume lot of RAM for 1601 pts ;)
<electronic_eel> later they improved their hw/sw so they now reach 7.5
<bvernoux> it is something like 30KB ;)
<bvernoux> in worst case
<bvernoux> electronic_eel, yes I suspect that too
<monochroma> i replaced the hard drives in both of my DDA-5005s with SSDs, you can still download the software from lecroy and such
<bvernoux> electronic_eel, anyway the dynamic range is not amazing
<Degi> Hmm I should ground my shelf when I get this scope...
<bvernoux> yes buy anti ESD stuff
<bvernoux> it is always a good things ;)
<Degi> Its kinda EMP deluxe when I play around with HV stuff on the ground floor..
<Degi> I should get an ESD mat... Any recommendations? And any recommendations for a 19 V isolated PSU? My laptop PSU has some big voltage swing on the output relative to ground
<bvernoux> electronic_eel, their SVA1075X is clearly not as good a PicoVNA
<bvernoux> and more expensive
<bvernoux> will be interesting to see a teardown of PicoVNA
<Degi> Hm I kinda wanna design a bunch of probes...
<azonenberg> Degi: if you wanna try your hand at a DC-100 MHz active probe, we need one for BLONDEL
<Degi> With 50 ohm output? Specific input impedance or "as high as possible"?
<azonenberg> "as high as possible" input, 50 ohm single ended output
<azonenberg> for initial rev focus on getting the probe design itself to work and just run off a bench supply
<azonenberg> we can make the fancy usb power version later
<Degi> We can stick a betavoltaic into it
<Degi> Is the output swing important? Should it be +- 5 V or is +- 1.25 V also ok?
<azonenberg> I'd target lower, we dont want to saturate the input
<azonenberg> +/- 1 give or take a bit is probably about right, we can run the frontend at ~unity gain then
<Degi> The OPA859 seems kinda nice if +- 1 V input range is OK lol. Though I guess some passives should come before that
<Degi> Nvm OPA656 looks better
<Degi> Is there some input range we wanna have? Is +- 2.5 V good?
<Degi> Or more like +- 1 kV?
<azonenberg> I'd target significantly wider. We will probably want an attenuator on the input to lower loading. Say +/- 10V at least operating range, with damage threshold of at least +/- 20?
<Degi> Hmm k. Any suggestions for R?
<Degi> 10 GOhm?
<azonenberg> 1M or even 100K input impedance is probably fine
<electronic_eel> +- 10V operating range? that sounds borderline unusable to me. you can't even use it on a regular 12v supply then
<electronic_eel> more like +- 50v to cover most common low voltage electronics stuff, including 48v supplies
<azonenberg> ok that works too
bvernoux has quit [Quit: Leaving]
<Degi> Hmm
<Degi> We can have 20 0V
<azonenberg> If we are going to have operating ranges above touch-safe levels then we need to start thinking about proper insulation etc
<azonenberg> rather than some 3d printed plastic
<Degi> Well we can advertise it as 30 V
<azonenberg> I have no problem with designing a board capable of wider range, like you said we can always claim lower specs
<electronic_eel> here is some documentation about a design with a OPA659, up to 500mhz: https://www.bartelsos.de/dk7jb.php/fet-tastkoepfe
<electronic_eel> unfortunately in German though
<Degi> Can we use a resistor as a capacitor?
<azonenberg> if you characterize it properly? sure
<azonenberg> what, do you need a 30 fF cap? :p
<Degi> Yes exactly 30 fF but that is parallel plate assumption, in reality more like 60
<azonenberg> why such a $$$ resistor?
<Degi> 0.1 % 50 PPM/°C 0.1 PPM/V
<Degi> I mean whats the price range for the probe? 15€? 50€?
<azonenberg> i'd like sub 100 USD if possible
<azonenberg> so a 7.55 @ qty 1 resistor is a lot
<Degi> Well we only need 1
<azonenberg> (the ones i use in the 2 GHz passive probe are only around 2.3 USD each)
<Degi> Well and the low side resistor but I bet I can find a cheaper one
<azonenberg> Look around a bit more. But by all means draw up a schematic and we'll think about it
<electronic_eel> oh, one thing came to my mind: AC coupling
<Degi> oh
<electronic_eel> if you want to keep the capabilities of a conventional scope, we need ac coupling
<azonenberg> Yeah
<Degi> Wanna have that in the scope? Otherwise external SMA dongle...
<Degi> Or in the probe?
<electronic_eel> but doing it on the 50 ohms side of the scope means you have a high pass filter with very high freq
<electronic_eel> or you need a huuuughe cap
<Degi> You mean a big cap would be needed @50?
<azonenberg> My plan is for BLONDEL to be DC coupled only, and use a SMA DC block if you need one
<electronic_eel> the passive probe is designed for higher freq stuff, so I don't see much of an issue there
<azonenberg> how low freq do normal sma dc blocks go to?
<electronic_eel> but the active probe for lower freq stuff - that is where the missing ac coupling could become a problem
<electronic_eel> so could we put the ac coupling into the probe?
<azonenberg> I see DC blocks down to 50 kHz available commodity
<electronic_eel> yes, but isn't it more like 20 or 50 Hz you want?
<azonenberg> Do you know what the bandwidth of normal scope AC coupling is?
<azonenberg> serious q, i can't remember the last time i've used it
<Degi> Hmm
<Degi> I think my old scope has 40 pF 1 MOhm
<electronic_eel> hmm, I use it when I want to measure very low signals and have to turn up the gain to the max
<Degi> Can we have a switch in the probe
<electronic_eel> it could probably also be done with dc coupling if the scope would automatically follow with the offset or similar
<Degi> Hmh nah too complicated
<Degi> We could add a second low freq ADC lol
<azonenberg> i have a filter in scopehal that does AC coupling in postprocessing
<azonenberg> right now it's across a whole capture but i guess you could do it with higher bandwidth too
<azonenberg> i.e. it subtracts the mean of all samples from each sample
<electronic_eel> it would have to be coupled with the offset stage, otherwise you quickly go out of range with max gain
<Degi> So this circuit I posted should provide inverted output, 10.00 MOhm input at DC, idk about capacitance errors tbh...
<Degi> Its possible to make it noninverting too tbh
<Degi> But then input VCM varies idk if thats bad
<azonenberg> inversion is fine because we'll be able to ID the probe via the usb interface
<azonenberg> and then invert in post
<Degi> It should provide 32 kOhm impedance at 100 MHz
<Degi> At Vout = +- 1 V Vin = +- 100 V
<Degi> If we want Vin +- 10 V then we'd need a capacitor parallel to the 10M
<Degi> Maybe a TVS from input to ground
<azonenberg> what about variable gain on the active probe? just a thought
<azonenberg> rather than having ultra low voltages on the output at lower ranges
<azonenberg> we can easily integrate it with gain settings on the scope itself
<electronic_eel> maybe switchable?
<azonenberg> lecroy has integration with the tetris frontend to maui so gain/offset controls actually change the probe config
<electronic_eel> not free variable with a trimmer or varicap, but two or three values you can switch between
<azonenberg> yeah thats what i was thinking, a few decades or so should be enough
<electronic_eel> if it is controlled from the mcu on the probe, that could be done with a button (for non BLONDEL-users) or via the scope gui
juli963 has quit [Quit: Nettalk6 - www.ntalk.de]
<azonenberg> yeah it will be controlled by the probe mcu
<azonenberg> if we have a button there, we'll need synchronization between the mcus bidirectionally
<azonenberg> so the probe can tell the scope you pushed the button and push the config to the gui
<electronic_eel> yes, why not? we plan to have usb there, don't we?
<azonenberg> yeah i just mean it will slightly complicate firmware
<azonenberg> not a reason to not do it
<azonenberg> just a design consideration
<electronic_eel> I think planning the probes to be of use without BLONDEL on a conventional scope will increase sales significantly
<azonenberg> yeah agreed
<azonenberg> i think inexpensive high quality probes and accessories for existing scopes will sell well
<electronic_eel> exactly
<azonenberg> And will be something we can use to fund R&D of the scopes, which *we* really want to be doing
<azonenberg> but we have to pay for that somehow :p
<Degi> Hm for universal, we could have a probe power & control hub
<azonenberg> yeah we'd need to definitely have a standalone host to provide power etc
<Degi> Variable gain: Is it OK to add after input stage or should input stage components be variable?
<azonenberg> Ideally input stage variable, but i'm not sure. Havent put much thought into the active probe
<electronic_eel> variable gain on the input stage is better. maybe mux in different values?
<Degi> If input stage 1:1 required, the input C will be several pF
<Degi> Like 2 or so
<electronic_eel> hmm 1:1, not sure if we need that
<Degi> Also each probe would have a varicap for compensation for production variations
<electronic_eel> if you really need very low voltage low freq stuff you are much better off with a special high gain probe
<Degi> Hm k
<Degi> We can do 10:1 and 100:1?
<azonenberg> That sounds reasonable
<azonenberg> Just ensure that if you set 10:1 it won't be damaged by out-of-range inputs
<electronic_eel> maybe something in between? lik 10:1 and 50:1?
<azonenberg> clipping is ok, dying is not
<Degi> I think Ill usea second OP amp for that then
<electronic_eel> you mean a 2nd stage or switching the one for the first stage?
<Degi> A second stage tbh
<electronic_eel> hmm, if you attenuate 100:1 it won't be easy to get low signals back out of the noise
<Degi> hmh not sure
<Degi> Like how much noise do we have here vs rest of analog chain?
<electronic_eel> this is just a few 100 mhz, there should be muxes that work for that freq range
<Degi> Meh I can just put a relay there
<electronic_eel> relay always works
<electronic_eel> just takes some space
<Degi> At 10:1 we'd need I think 500 fF input capacitance
<azonenberg> So here's the question, what bw do we want to target with this?
<Degi> Not sure if a mux is suitable because the low side capacitance would be 5 pF
<Degi> 100 MHz?
<azonenberg> more might be good if we can do it without much effort
<Degi> Idk we can try
<electronic_eel> would be nice if it work for DUDDELL too
<Degi> Thats 350 MHz?
<electronic_eel> the scope is planned for 250, but the probe should go a bit higher
<azonenberg> Yeah. I want something for ZENNECK too that can do lower speeds as well, although i guess it's probably OK to have a low bw active probe and then the high bw passive probe
<electronic_eel> I think targeting something like the rigol crowd, with the 350mhz MSO5000ers, would make sense from a business perspective
<azonenberg> Makes sense
<azonenberg> But it has to be demonstrably better than oem passive probes
<azonenberg> (and work with whatever inputs those scopes have)
<electronic_eel> yes, of course. I think they have 50 ohms
<electronic_eel> urgh. just checked the datasheet: the rigol MSO5000 just have 1Meg, no 50 ohms option
<Degi> Hm oof
<Degi> Should I still buy it?
<Degi> (The active probe should work for both)
<Degi> I do have T's and a 50 ohm terminator for BNC after all...
<electronic_eel> you'd need an external terminator. they degrade the signal
<azonenberg> inline BNC term would work
<azonenberg> but yeah
<azonenberg> also VNA is here woo
<electronic_eel> next level up, the mso7000 series, has 50 ohms
<azonenberg> the cal cert is printed on A4 paper and doesn't fit in my hole punch so getting it into my cal binder is annoying :p
<sorear> metric instrument, metric cal cert, checks out
<azonenberg> lol
<azonenberg> i swear i've punched A4 before and it fit
<azonenberg> but this doesnt
<azonenberg> maybe it was a different punch
<Degi> Get a drill
<Degi> Anyways what do you think of this schematic? https://i.imgur.com/5gUP9NY.png
<Degi> I can upload the kicad folder if you wanna
<miek> better send the punch out for cal too :p
<azonenberg> lol
<electronic_eel> lol
<Degi> lol
<electronic_eel> Degi: I suggest drawing the schematics in ltspice and then use it to have a look at the performance...
<electronic_eel> if the design is finished switch to kicad
<Degi> https://www.schaeffer-ag.de/en/ This looks interesting
<electronic_eel> they are a well established front panel manufacturer
<electronic_eel> convenient, but a bit pricey
<electronic_eel> ok, I'm going to bed now
<Degi> Good night
<electronic_eel> good night
<Degi> Hmh the circuit seems to be good to 34 MHz with 220 MHz OP amp
<Degi> I like how you can add negatie capacitances in ltspice
<azonenberg> Ok so i need a little time to set up the VNA
<Degi> K looks like op amp choice matters
<Degi> I guess input impedance and exact ratio isnt that important?
<Degi> According to LTSpice not using the OP amp in the loop is slightly better
<Degi> Somehow the op amps in ltspice have a peak at higher frequenies at gain 1
<Degi> Does this work as a unity gain buffer https://www.mouser.de/datasheet/2/609/626810f-1279531.pdf