basshead if it is about the girls - I'd spend it on a Guitar instead... and I agree with Goof - Sure you can, but a different kind - and likely better conversationalists at that.
My rigol DS1054Z does this. I set mem depth at 6M pts, and it will capture those many points, even if stop at 5ns per division. and it is a budget one.. Just zoom out.and you have about 20ms of data with three channels on, so that i do not loose SPI data.
I've just thought that feature as granted cause my cheap and old rigol ds1054z has some "zoom-out" features in single mode, if you set a memory depth it always fill out the memory with the current sample rate (that is a function of the zoom level) so depending on the sampling rate you have some extra data for zooming out and explore the waveform.
when I was working for a large automotive company i was so spoiled that this was the minimum i'd work on :) now i have a company of my own and this is the maximum i can afford currently :)
The "Run stop " is very similar to professional slow motion cameras that films slow motion footage in the RAM constantly overwriting it and when you push the trigger button and its set to capture before and after trigger the remaining slow mo footage is then stored to the rest of the RAM But here you stop it instantly at already saved data It is faster to write data to RAM than deleting and then writing to it, so it is a better/cheaper way of getting good results 👍
For the Siglent, this is like buying a 1000mL drink and only getting 100mL and being told you will have to come back for the remainder 900mL in increments of 100mL. I want it all and I want it now!
Dave, I think you’re missing one point on how the Siglent utilise its spare memory. Not sure about other manufacturers but with siglent if you limit the memory depth you gain longer history. So if you set low memory depth then in normal mode you will have access to the history of triggered captures each spanning a screen at current time division (but nothing before or after it as you're pointing out). The number of frames that fits the history depends on set memory depth. This is quite a useful feature. Not sure if other manufacturers offer an equivalent. That said I agree that in the single capture mode there's no history so ideally the scope could utilize the memory for capturing the signal before/after the screen frame.
someone even can blame keysight for changing memory depth "on the go" to improve display speed and wave capture rate but nothing is more annoying than the Tek MSO manual memory depth setup
The difference between run/stop and single is expected. It can be improved sometimes but there are technical reasons for it, especially on so fast trigger rate scope. In single it will usually capture whole memory buffer symmetrically on both sides. In Run/stop it will depend on trigger rate and some other factors. But the behavior of Siglent is definitively a bug, or a serious design flaw. Hopefully they fix it. It would drive me nuts.
At least on the Siglent, does this change with the "Slow" acquisition mode? Could be that they're throwing away the extra points explicitly to optimize the time it takes to display the data (higher waveform updates per second datasheet figure) in Fast mode than the Slow mode is more comparable in operation to the Keysight and Tek.
That "glitch" looks to me like just the handover between the master and the slave during the ACK bit. The master releases the buss temporarily before the slave pulls the ack low so you see a little glitch.
Going to guess the Siglent is using the various memory chips in a burst mode, and there is a limit to how many sequential data transfers they can do to each chip in turn before they have to change an address upper bit, limiting the number of bytes they can use. Probably a multiple of 2, but decimalised so that they can store as well trigger information and a sequence number to reassemble the jigsaw puzzle on the display. Agilent probably uses a larger number of faster chips for memory, so that they can interleave the blocks as they are written, leaving enough time between selecting each chip to allow the address to be updated for the next burst. Likely takes a lot more FPGA resources to have each memory chip with essentially it's own complete controller, along with a screaming fast multiplexer to do the interleaving and buffer for each one, to handle in the time before the next tranche of data is coming in. Thus the memory split, so that you essentially have duplicate blocks of this, and write the one for acquisition, while a read block of the FPGA is busy reading the data back and assembling it into a serial signal to transfer to the display controller, and then after the write buffer is full you get no sampling till the read is complete, unless you are sampling slow enough for the memory and display to keep up.
Do any of them capture so the trigger is a the center of the data? It would not be that hard to implement I think. Like if you kept dumping to a memory block and starting over at the other start, you just have to say where the trigger is and get the data from before that point.
Seems that the Siglent while having 200Mpoints, the limiting is the writing speed of memory/ sampling rate, so at 0.5ns/div limits there can only be 10 points captured in a trace. Meaning 2G samples per second, or 1/2 ns per sample
Nope, makes absolutely no difference. Even implies that in the manual, it literally just slows down the update rate with no other apparent advantage. Pretty dumb.
The old Tek TDS2024C I have to work with does not seem to have the feature either. But my personal Rigol DS2072A seems to do it just fine (not on auto of course) as well as it can do many other useful things not available with the old Tektronix. Who would have guessed that an average modern scope can be so much better than an average 20 years old one...
@@Kirillissimus I have a TDS 380 there is simply not enught memory to do that Kind of stuff. I think at some Point i should Invest in a newer skope.....
I want a scope with 144hz 17inch touchscreen, Sigilent's big touchscreen is great! And physical buttons above the inputs! And still 2 knobs! Aaaalll the things
Why this expensive Keysight scope has 16-32 times less memory than my first PC from 2000? How is it I have a 32Gb of RAM while most entry level scopes are limited to few megabytes? Madness!
I have an EPISODE SUGGESTION: DETECTING HIDDEN CAMS USING SPECTRUM ANALYZERS Or using any other appropriate way, it's not fun only but it's useful. I would love to see some specialized "difficult to detect" cameras getting busted.
You know THEY are monitoring everything YOU post on the internet, right? Even if you watch Dave telling you how to detect THEIR hidden cameras, THEY will KNOW you know before you even know. You know? Btw, looks don't matter if you're an ass to the ladies.
18:10 how do 10 ms at 10 GS/s equal 20Mpts? 10 x 10e-3 x 10 x 10e9 = 10e8 = 100M
Because I'm a dumb arse.
@@EEVblog not to worry :D had to pull out ye old confuser to check my own sanity :D
@@septicsounds3216 wow you are smart! now I feel dumb lol
I guess he was "Septic" about the result
Sweet kit. If I ever have a spare US$14,271 maybe I will pick one up.
Get a fancy used car for that. You can't get girls with an oscilloscope.
@@basshead. yess you can
basshead if it is about the girls - I'd spend it on a Guitar instead... and I agree with Goof - Sure you can, but a different kind - and likely better conversationalists at that.
+Goof Heesakkers Not hot chicks.
@@basshead. or a nice cnc machine, way more fun than girls!
Great video - hadn't given this much thought before and now it has to be a must have feature in a 'scope.
The Siglent does it that way because the 'history' mode is always recording history into the spare memory.
My rigol DS1054Z does this. I set mem depth at 6M pts, and it will capture those many points, even if stop at 5ns per division. and it is a budget one.. Just zoom out.and you have about 20ms of data with three channels on, so that i do not loose SPI data.
Even my ultra-budget Hantek DSO4072c with its gigantic 40K (Yes, K) samples can do this. You cant zoom out very far, but still.
Thumbnail looks like a fusion reactor. *thumbs up earned before even watching the video*
I did like that thumbnail! Shared several slight alternatives on twitter, people picked that one.
Good video. All my scopes work how you'd expect, full memory capture when setting for full memory.
Having the ds1054z I was thinking the other way round, thought that every scope can do that...
I like how you go ahead and show off all your scopes at the end :D
I'm just wondering, do you have a particular use for each or just a collection?
I've just thought that feature as granted cause my cheap and old rigol ds1054z has some "zoom-out" features in single mode, if you set a memory depth it always fill out the memory with the current sample rate (that is a function of the zoom level) so depending on the sampling rate you have some extra data for zooming out and explore the waveform.
Engineers: I wish I could measure things before they happen.
Keysight: What if there's something that you don't know, that you don't know.
7:08 Hey, what's going on there? Why is it displaying 3 "traces" at once? Only one of them can be realtime, correct?
when I was working for a large automotive company i was so spoiled that this was the minimum i'd work on :) now i have a company of my own and this is the maximum i can afford currently :)
Hey it actually works on the Rigol 1054Z (hacked to the 1104Z) when you change the memory depth from auto to 24M
Great to know, this is a big help!
Great video! I may have very much use of this. As the signals I look at may not be consistent over time and all data is valuable. Thanks!
The "Run stop " is very similar to professional slow motion cameras that films slow motion footage in the RAM constantly overwriting it and when you push the trigger button and its set to capture before and after trigger the remaining slow mo footage is then stored to the rest of the RAM
But here you stop it instantly at already saved data
It is faster to write data to RAM than deleting and then writing to it, so it is a better/cheaper way of getting good results 👍
For the Siglent, this is like buying a 1000mL drink and only getting 100mL and being told you will have to come back for the remainder 900mL in increments of 100mL.
I want it all and I want it now!
Well you can have your whole 1000mL, you just need to drink it a certain way.
@@EEVblog I saw the second video. It seems the other 900mL are in different containers. Who would have known? Great video.
Thank you very much!
Dave, I think you’re missing one point on how the Siglent utilise its spare memory. Not sure about other manufacturers but with siglent if you limit the memory depth you gain longer history. So if you set low memory depth then in normal mode you will have access to the history of triggered captures each spanning a screen at current time division (but nothing before or after it as you're pointing out). The number of frames that fits the history depends on set memory depth. This is quite a useful feature. Not sure if other manufacturers offer an equivalent. That said I agree that in the single capture mode there's no history so ideally the scope could utilize the memory for capturing the signal before/after the screen frame.
See my next video on this.
Sure the is no history insingle cature mode ? Could be useful imho
I swear it always sounds like Dave pronounces it as "field programmable gatorade".
Aussies slur their words together
I take that "zoom out" function for granted; I didn't even know that some lower end scope don't do that. Perhaps for a faster display update?
someone even can blame keysight for changing memory depth "on the go" to improve display speed and wave capture rate but nothing is more annoying than the Tek MSO manual memory depth setup
Like a SR71, was a bit over my head. But I liked it just the same. No pain no gain.
BIG 👍 UP
thank you M.r
The difference between run/stop and single is expected. It can be improved sometimes but there are technical reasons for it, especially on so fast trigger rate scope. In single it will usually capture whole memory buffer symmetrically on both sides. In Run/stop it will depend on trigger rate and some other factors.
But the behavior of Siglent is definitively a bug, or a serious design flaw. Hopefully they fix it. It would drive me nuts.
At least on the Siglent, does this change with the "Slow" acquisition mode? Could be that they're throwing away the extra points explicitly to optimize the time it takes to display the data (higher waveform updates per second datasheet figure) in Fast mode than the Slow mode is more comparable in operation to the Keysight and Tek.
My owon vds1022i usb scope can’t even zoom period. Tough.
Works also on my Rigol DS4014
That "glitch" looks to me like just the handover between the master and the slave during the ACK bit. The master releases the buss temporarily before the slave pulls the ack low so you see a little glitch.
ON the Micsig, I usually have the Zoom out ability if I manually choose 28meg,.
Going to guess the Siglent is using the various memory chips in a burst mode, and there is a limit to how many sequential data transfers they can do to each chip in turn before they have to change an address upper bit, limiting the number of bytes they can use. Probably a multiple of 2, but decimalised so that they can store as well trigger information and a sequence number to reassemble the jigsaw puzzle on the display.
Agilent probably uses a larger number of faster chips for memory, so that they can interleave the blocks as they are written, leaving enough time between selecting each chip to allow the address to be updated for the next burst. Likely takes a lot more FPGA resources to have each memory chip with essentially it's own complete controller, along with a screaming fast multiplexer to do the interleaving and buffer for each one, to handle in the time before the next tranche of data is coming in. Thus the memory split, so that you essentially have duplicate blocks of this, and write the one for acquisition, while a read block of the FPGA is busy reading the data back and assembling it into a serial signal to transfer to the display controller, and then after the write buffer is full you get no sampling till the read is complete, unless you are sampling slow enough for the memory and display to keep up.
I don't have a scope, but if i have to buy on, i will have the eevblog special version.
Do any of them capture so the trigger is a the center of the data? It would not be that hard to implement I think. Like if you kept dumping to a memory block and starting over at the other start, you just have to say where the trigger is and get the data from before that point.
Seems that the Siglent while having 200Mpoints, the limiting is the writing speed of memory/ sampling rate, so at 0.5ns/div limits there can only be 10 points captured in a trace. Meaning 2G samples per second, or 1/2 ns per sample
They seem to be using the extra memory for a continuous history mode feature, but only in Stop mode. Seems like a deliberate tradeoff.
I didn't know Dave got it's own special edition oscilloscope 👍
Special edition decal set.
I’m still using an old bk precision analog scope. I’m looking to upgrade to a more modern digital scope. Do you know any modern good upgrades?
Did you try it in roll mode?
have you set the acquisition mode of the lecroy to slow instead of fast? maybe that does it?
Nope, makes absolutely no difference. Even implies that in the manual, it literally just slows down the update rate with no other apparent advantage. Pretty dumb.
Audio samplers have been able to do that for ages. I guess you need about 1000x the bandwidth though :)
Tektronix oscilloscopes are missing in this video! I'd like to can see them so as to compare.
Err, watch it again...
@@EEVblog Sorry, you are right! 😅 I apologise to you. I had forgotten my glasses. 🤓 Hahaha.
What about the Keysight DSOx1204G, does it zoom out?
Migsig STO (C/E) same same when your not in auto as I recall and take advantage of 30 or 70meg..
@EEVBlog Curious if you heard from Siglent?
what about siglent Siglent SDS1202X-E ??? ( it is more realistic choise for hobiest)
Wait a minute.. 2016 Special Dave Edition? What's the deal there...?
Year i think my old Tektronix can't do anything linke that ...
The old Tek TDS2024C I have to work with does not seem to have the feature either. But my personal Rigol DS2072A seems to do it just fine (not on auto of course) as well as it can do many other useful things not available with the old Tektronix. Who would have guessed that an average modern scope can be so much better than an average 20 years old one...
@@Kirillissimus I have a TDS 380 there is simply not enught memory to do that Kind of stuff. I think at some Point i should Invest in a newer skope.....
Vertical time base? ;)
It's when your bench is so cramped that you have to use the scope lying on it's side.
I want a scope with 144hz 17inch touchscreen, Sigilent's big touchscreen is great! And physical buttons above the inputs! And still 2 knobs! Aaaalll the things
What abt the Rigol 5000 series?
Same chipset as the 7000, so presumably same result. I don't have one to check.
Best keysight ad I've seen
Why this expensive Keysight scope has 16-32 times less memory than my first PC from 2000? How is it I have a 32Gb of RAM while most entry level scopes are limited to few megabytes?
Madness!
Singlet uses all its memory in storing traces in history
I have an EPISODE SUGGESTION:
DETECTING HIDDEN CAMS USING SPECTRUM ANALYZERS
Or using any other appropriate way, it's not fun only but it's useful. I would love to see some specialized "difficult to detect" cameras getting busted.
You know THEY are monitoring everything YOU post on the internet, right? Even if you watch Dave telling you how to detect THEIR hidden cameras, THEY will KNOW you know before you even know. You know? Btw, looks don't matter if you're an ass to the ladies.
My old school scope is 20 megs crt scope :(
I've got a sweet multi-meter, BM235... it has Min. Max as well as average feature.
Ping pong seems to be Round Robin with 2 buckets...
Enhance!
that's a 14 grand scope, holey moley!
The different fonts, text sizes and lines on the Rigol hurt my eyes every time I see it. Terrible!
I think someone counted once and it was like 8 different font or something.
My analog scopes get all of the points.
Hi, I don't use an oscilloscope every day, am I not winning?
Clearly not!
@@EEVblog I guess I better buy my first scope, and then.... success.
Can your oscilloscope can do this???
damn, my newly acquired hantek isn't so great now.
Hantek doesn't do it either?
@@EEVblog Not sure about the higher end models. Mine is low low end.
if you told me that not all scopes behave this way, I WOULD TAKE THE RED PILL AND SEE HOW DEEP THE RABBIT HOLE GOES
👍
I got a CS-4025, why do I watch this lmao
I don't even have one lol
Made me think of this Spaceballs clip: th-cam.com/video/5drjr9PmTMA/w-d-xo.html
why would one want to compare 3k $ with 15k $ oscilloscope?
It's not, it's about comparing architectures from different manufacturuers. The lower end ones usually have the same arhictecture.
Who cares. I have UNI-T UPO 2072 CS and it is the best scope. Cos I say it is. There.
Holy shit!! Thats a $14k scope!!!!!!!! It does NOT look like a $14,000 scope.....not sure what one does look like....well, a Rohde & Schwarz I guess.
and that scope costs $14k
US$14,271 and memory size of a 1990s computer ;-D
Only Keysight can do that