<azonenberg>
Quick update, i have a bunch of refactoring and work on file loading that is going to land at some point but i don't want to push incomplete/potentially breaking stuff
<azonenberg>
there's a few things i need to work out around things like IDs for instruments
<azonenberg>
and the fact that when you have multiple inheritance the pointer to the Oscilloscope view of an object and the Multimeter view of the object may not be the same memory address
<azonenberg>
(which is a problem when you have a table mapping raw pointers to numberic IDs)
<azonenberg>
I have power supply serialization working in a write-only fashion, i.e. you can save scopesessions that will eventually be loadable
<azonenberg>
but there is currently no loading support
<azonenberg>
When load support does come, it will be offline initially
<azonenberg>
i want to look at a broader refactoring of how some of the non-scope instruments are handled from a gui perspective
<azonenberg>
e.g right now the power supply dialog "owns" the connection to the PSU, and the session disconnects from it when the dialog is closed
<azonenberg>
but this is not at all obvious to the user
<azonenberg>
i want to make it more like the scope model where you have global settings (somewhere) and then per channel settings you can get by double clicking nodes in the filter graph or something. going to have to think about that
<azonenberg>
basically, i want to allow you to be connected to the psu without keeping the dialog open and in view all the time
<d1b2>
<kumarabhishek> Hi, I was discussing this in the Supercon chat this morning (with @aleksorsist) and thought I might seek some opinion here as well.
<d1b2>
<kumarabhishek> If I have a little bit of budget to buy a mid-range scope (one step above the 1054Z but still not exorbitantly priced) - I was considering the Siglent SDS2000X+ series and the DHO1xxx series from Rigol but I also saw the MSO5000 series of Rigol which is very compelling at the price point it is right now.
<d1b2>
<kumarabhishek> The EEVblog forums seem to suggest that the MSO5x series doesn't have the best frontend performance and there also seems to be some consensus wrt firmware issues. I was pretty biased towards the Siglent and would have perhaps bought it, but I'm still thinking how big of a deal the software issues on Rigol scopes or the frontend noise is and would like to know if someone has first hand hardware experience with either of these models.
<d1b2>
<kumarabhishek> As it stands today, the Rigol is still a pretty compelling purchase, but I want to be sure that there are no hiccups.
<d1b2>
<johnsel> What is your use-case? What kind of performance do you need? I have a Rigol MSO8k here which is pretty nice, but it is somewhat noisy and using it remotely over SCPI the performance is pretty bad.
<d1b2>
<johnsel> It all depends on what specifically you are looking to achieve if something is suitable or not.
<d1b2>
<kumarabhishek> I need a scope that has reasonably good high speed performance on a budget. But you're right, I do need to be more specific
<d1b2>
<kumarabhishek> 'reasonably good' is a pretty subjective term
<d1b2>
<johnsel> So is high-speed. Our big boss azonenberg might call 16GHz high speed whereas we pleb might call 2GHz high speed
<d1b2>
<kumarabhishek> 200+ MHz of bandwidth would be nice, yes
<d1b2>
<kumarabhishek> You're right
<d1b2>
<kumarabhishek> I don't want to spend more than 2K $, but get the maximum value out of it
<d1b2>
<johnsel> Noise becomes important for measuring power supplies etc, less so to diagnose i2c or spi
<d1b2>
<kumarabhishek> Yep, don't need scope for those things though. I'd say I'd much rather look at say, high speed differential signals starting with maybe HDMI
<d1b2>
<kumarabhishek> Or if I'm doing high speed digital i.e. > 100-200 MHz, make sure my signal integrity is intact
<d1b2>
<johnsel> Modern HDMI signals are pretty fast, much faster than 200MHz
<d1b2>
<johnsel> That's more along the lines of a few GHz
<d1b2>
<kumarabhishek> Yeah, but the Pico can do lower resolution HDMI still while overclocked
<d1b2>
<johnsel> Okay well if you want to make eye diagrams etc you'd want to use ngscopeclient for that, none of the scopes can do this on the instrument itself as far as I am aware. So you might want to look at the quality of the drivers
<d1b2>
<johnsel> As far as I am aware the siglent driver is much more stable than the rigol one, and I don't think we've had many people using the newer rigols with ngscopeclient at all
<d1b2>
<kumarabhishek> Yep. I was told the Siglent scopes are well supported in ngscopeclient
<d1b2>
<johnsel> azonenberg might be able to chime in on that but afaik not
<d1b2>
<kumarabhishek> Got it
<d1b2>
<johnsel> New rigols that is
<d1b2>
<kumarabhishek> I'd actually like to be able to observe practical things like PCIe or HDMI or Gigabit ethernet signals but not sure what's the minimum I'd need to spend in order to actually get a scope capable of doing this and I'm sure it would not be cheap
<d1b2>
<kumarabhishek> The gap seems to be too large I guess
<d1b2>
<kumarabhishek> On one side, you have cheap 1GSa/s 4-channel 100MHz bandwidth scopes and on the other side you have the fancy $10K scopes that can debug the high speed protocols but not sure if something's there in the middle
<d1b2>
<johnsel> PCIe starts at a few GHz and goes to many GHz quickly, HDMI also goes to 10GBps nowadays which you would want 16GHz to do SI measurements on, gigabit ethernet is the most doable of them as it is only 1000/4=125gbps per channel
<d1b2>
<johnsel> It's almost certainly an expensive undertaking to look at anything modern. Second hand market for Lecroys is pretty much the only reasonable way to get a few GHz measurement capability
<d1b2>
<johnsel> It may be better to look at your traces with a VNA if you are interested in SI
<azonenberg>
"fancy $10K scopes"
<azonenberg>
the scopes you use for doing modern stuff like higher rate PCIe and 100G Ethernet etc are well into the six figures list price
<azonenberg>
my 16 GHz scope was $33K on ebay and that was the deal of a lifetime
<azonenberg>
$10K might get you low single digit GHz if you shop around and are OK with buying several generations behind current
<azonenberg>
even siglent's 2 GHz scope i dont think is that cheap
<azonenberg>
ok actually it is, its $9990 list price
<azonenberg>
but that gets you 2 GHz at like 5 Gsps (so barely nyquist sampling rate... they claim 10 Gsps equivalent but that's with some interpolation or something)
<tnt>
I wish sampling scope were still current... I don't need real time capabilities, if I could do a reasonable eye pattern for those modern standards, I'd be happy.
<azonenberg>
and super slow scpi performance
<azonenberg>
Fast sampling scopes still exist
<azonenberg>
I actually was looking into the possibility of building one with a 7 seires or ultrascale+ GTX
<azonenberg>
they have some special features i think would be really useful for that
<azonenberg>
@kumarabhishek the siglent driver is reasonably stable since a lot of folks use it, but they're all slow
<tnt>
azonenberg: well AFAICT the ones you can buy today are meant for standard where real-time isn't a realistic option (i.e. way above what I need and thus my budget).
<azonenberg>
expect mid to low single digit waveforms per second, i think the 6000 sereis hits maybe 3-5 WFM/s at shallow memory and the lower end are less
<azonenberg>
@tnt well my plan in the not too distant future is to build a BERT + sampling scope using an xcau25p and see how far i can push it
<tnt>
(or they are for a specific stuff ... like we had a HD-SDI field debug thingie that was doing eye pattern but it was specialized for hd-sdi).
<azonenberg>
it's project #2 in my pipeline ahead f the trigger crossbar
<azonenberg>
ahead of*
<tnt>
that'd be nice for sure.
<azonenberg>
as a minimum i know for a fact i can get BER eyes
<d1b2>
<kumarabhishek> Wow
<azonenberg>
but i think its very likely i can get full eyes within a limited vertical range
<azonenberg>
Wrt Ethernet, you can probably get away with looking at 100baseTX with a scope that has at least 250 Msps and ~100 MHz BW if all you want to do is decode. my 350 MHz 2 Gsps lecroy did it with plenty of margin
<azonenberg>
gig is more problematic to do anything with not just because faster but because it's bidirectional signaling on the same wires (10/100 are one pair in and one out, gig uses all four pairs in both directions *simultaneously* vs taking turns or anything)
<azonenberg>
you need special test fixtures with directional couplers to split the two waveforms apart
<azonenberg>
i have a prototype of one but its not fine tuned to the point i'd recommend anyone else build one yet
<azonenberg>
gig over fiber is doable with a 1-2 GHz scope easily enough