4ch 100MHz 1GS/s per channel cheap-o-scope

[MrWolf]

Hello,

usually I mess around with low frequency and have been spoiled by having excessive resolution of waveforms. But specific task has emerged that will require 4 channels, at least 100MHz and most importantly at least 1GS/s on all channels at the same time.
Currently I have 25MHz PicoScopes and no real complaints about them. They do take abuse and software is feature-packed (FFT!). Initially I wanted to buy PicoScope 2408B or Rigol DS1054Z  (and hack it), until I realized that instead of being cheap-o-scopes these are actually fake-o-scopes because they are only 250MS/s on 4ch and 100MHz - which is basically a sad joke, not scope.
After quite a bit of search I found that only (?) cheap-o-scope doing full 1GS/s on 4 channels is Tek TBS1104, but it hopelessly outdated otherwise (low memory etc).
Proper PicoScopes start at €2,795.00 with 6402C, but this is out of budget currently (about €1500 max).
I do like (toy-o-scope  ) GW INSTEK GDS-2204E (has FFT!) but its only doing 500MS/s with all ch in use.
Currently almost out of ideas. Maybe buy two of something 2ch with trigger in-outs and  sync but this will probably go out of budget.
Used stuff do not like, only maybe refurbished with new knobs. Do not like knobs that have been rubbed by someone else
Maybe TBS1104 low mem is not an issue at full tilt since there is only 10 samples per wfm anyway?
Almost forgot - there must be option do display non-interpolated raw waveform. 10 samples per wfm is not much and when interpolated all the useful info will be lost for good.

[rf-loop]

SDS2104X all 4 channels on 1GSa/s for all channel. 2 channels on, 2GSa/ch
Also enough memory for most use,
Powerful fast segmented memory acquisition.
Always backround working Waveform history buffer.
In your budget due to factory offer.
Also it keep raw data independent of interpolation and display settings.

[TopLoser]

Keysight DSOX2004 hacked up to 200MHz does 1GS/s on all 4 channels and is way under your budget.

[Ice-Tea]

Or you could, you know, get an older scope from a reputed brand with the specs you need...

*whistles innocently*

[MrWolf]

Thats good news about SDS2104X, I actually looked at it but important info in datasheet was only in fine print (missed that) so I decided that it is probably only 500MS/s and went searching further. Since this is selling point they should make it more prominent in datasheet I think...
Also I think people should stop buying substandard fake-o-scopes and just push them out of the market. Dave makes some good points in this old video:

Sadly seems that many newbies do not realize this and go for products based on analog front end rating only (almost did fall for that).

[JPortici]

but they are cheap! and can get the job done.
And they are 100 MHz scopes.
You don't use a 100 MHz tool for a 100 MHz signal. 5-10 times the bandwidth and the adequate rise time if you want to have a good approximation of a square wave.

[tggzzz]

But specific task has emerged that will require 4 channels, at least 100MHz and most importantly at least 1GS/s on all channels at the same time.

Why is 1GS/s important to you, given the 100MHz spec? Often it is better to concentrate on the risetime.

You don't mention whether the signals are repetitive or single-shot, nor whether it is necessary to export waveforms to other devices.

[rf-loop]

Thats good news about SDS2104X, I actually looked at it but important info in datasheet was only in fine print (missed that) so I decided that it is probably only 500MS/s and went searching further. Since this is selling point they should make it more prominent in datasheet I think...

No. Absolutely not. It have 1GSa/s for every single channel simultaneously when all 4 channels are on.

There is 2 channel group.
I
CH1 + CH2 they have one 2x1GSa  Analog devices ADC. If one of these channel is in use ADC work in single 2GSa/s mode.  140M memory.
II
CH2 + CH3 they have one 2x1GSa  Analog devices ADC. If one of these channel is in use ADC work in single 2GSa/s mode. 140M memory. 140 for one channel and 70 per channel if both are in use.

If there is in group I or II both channel on, then other group can use only 1GSa/s even if in this group only one channel is in use.
(principle is that channels in use at same time, can use only same speed)

If select group I  one channel in use and group II  one channel... both can use 2GSa/s and both have max 140M acquistion memory (+ more for waveform history buffer, what principle is explained more in SDS1000X thread. Same principle in SDS2000X but 2kX have lot of more processing power and memory)

Also about many information in web. You need really be careful because most of informations are Obsolete. And if you look some videos, even from Dave, be careful and keep in mind there is many kind of "user errors".   Also many times there have been perhaps even many FW updates after these some videos.

[rsjsouza]

Another alternative is the DS4014E, which goes up to 2GSPS per channel and is around €1.6k.

If I am not mistaken, the GW Instek GDS2104A has 1GSPS per channel at 100MHz. At €1.3k it fits in your budget.

[MrWolf]

but they are cheap! and can get the job done.
And they are 100 MHz scopes.
You don't use a 100 MHz tool for a 100 MHz signal. 5-10 times the bandwidth and the adequate rise time if you want to have a good approximation of a square wave.

No plans to use it for digital signal tinkering. Analog sine-waves in 50...100MHz range with phase shifts involved. Basically want to learn about transmission line stuff in freq area where lines are getting reasonably short for experimenting. If I can get general picture at 4ch 1GS and pinpoint more exactly at 2ch 2GS will be more-less on top of things. With stable signals can even enable ETS and get hires. However I do realize that this is absolute minumum tech for this... But you do not spend big bucks in area where you are novice. Upgrade to GHz gear after year or two learning curve...

[MrWolf]

You don't mention whether the signals are repetitive or single-shot, nor whether it is necessary to export waveforms to other devices.

CSV export would not hurt. Signals analog, modulated, repetitive eg transmission line suff. Replys got me thinking maybe should go for 200MHz, 1GS. In some cases seems to fit in budget.

[Ice-Tea]

Or, you know, 1GHz, 4GSa/S

*Whistles some more*

[Fungus]

I wanted to buy PicoScope 2408B or Rigol DS1054Z  (and hack it), until I realized that instead of being cheap-o-scopes these are actually fake-o-scopes because they are only 250MS/s on 4ch and 100MHz

Later on you'll find out that 250MS/s is the correct sample rate for sampling 100MHz signals and then you'll realize that your post is embarrassing.

at least 100MHz and most importantly at least 1GS/s on all channels at the same time.

That specification makes no sense.

[tggzzz]

but they are cheap! and can get the job done.
And they are 100 MHz scopes.
You don't use a 100 MHz tool for a 100 MHz signal. 5-10 times the bandwidth and the adequate rise time if you want to have a good approximation of a square wave.

No plans to use it for digital signal tinkering. Analog sine-waves in 50...100MHz range with phase shifts involved. Basically want to learn about transmission line stuff in freq area where lines are getting reasonably short for experimenting. If I can get general picture at 4ch 1GS and pinpoint more exactly at 2ch 2GS will be more-less on top of things. With stable signals can even enable ETS and get hires. However I do realize that this is absolute minumum tech for this... But you do not spend big bucks in area where you are novice. Upgrade to GHz gear after year or two learning curve...

In that case you really want a TDR. A Tektronix 1502 will enable you to resolve small impedance changes 2cm apart. That's a system risetime of  140ps, roughly equivalent to 2.5GHz - and you get a 50ps step generator included.

See https://entertaininghacks.wordpress.com/2016/08/30/a-traditional-tdr-cable-tester-with-2cm-resolution/ for pretty pictures.

Its primary use is finding small and large problems in cables and connectors and PCBs, or cable length, or dielectric constant. Let me know offline how much you would pay for a working example, because I have a spare one for sale I'll see if anyone wants it at a hamfest on Sunday

N.B. my spare is the "normal" variant, calibrated in feet and inches, sigh!

[EEVblog]

Will you ever use the full 100MHz bandwidth on all 4 channel simultaneously?

[MrWolf]

That specification makes no sense.

Has it occurred to you that I have actually tested my scenarios at lower frequencies with exactly matching analog freq / sample ratio? Findings conform exactly to that old EEVBlog video. "Fatal data loss" occurs at about 10 samples per wave for basic sine wave. So in short you are saying that owner of this blog is talking BS in that video. Way to go.

Will you ever use the full 100MHz bandwidth on all 4 channel simultaneously?

Barely enough for starters. Monitoring for both voltage and current tends to eat up channels fast...

[MrWolf]

Or, you know, 1GHz, 4GSa/S

Advice is correct but as ex-CCCP citizen Im used to making fine carpentry with axe

[JPortici]

oh get that sweet lecroy of his already

[MrWolf]

oh get that sweet lecroy of his already

This kind of stuff will come after I had my way with something fresh and probably chinese. Also not very portable... forgot to mention portability req because all the low-end stuff is pretty portable nowdays...
...ok who Im fooling - Its too big to hide from my wife in small apartment   

[JPortici]

the lecroy can be disguised as a vintage tv 
or a futuristic microwave 
get creative

anyway measuring sine wave phase shift.. is a scope the correct tool? and i think you'd need a lot higher bandwidth than the signal as i believe the input antialiasing filter is not gaussian.. sure is not on old DSOs or cheap modern DSOs, with lower samplerate-bandwidth ratio.

Is the input channel bode plot even provided on higher end scopes?

[Ice-Tea]

oh get that sweet lecroy of his already

This kind of stuff will come after I had my way with something fresh and probably chinese. Also not very portable... forgot to mention portability req because all the low-end stuff is pretty portable nowdays...
...ok who Im fooling - Its too big to hide from my wife in small apartment   

But... It has a carying handle! And you can use it as a boat anchor!

I see your point, though

[Fungus]

That specification makes no sense.

Has it occurred to you that I have actually tested my scenarios at lower frequencies with exactly matching analog freq / sample ratio? Findings conform exactly to that old EEVBlog video. "Fatal data loss" occurs at about 10 samples per wave for basic sine wave. So in short you are saying that owner of this blog is talking BS in that video. Way to go.

Which video? What 'scope?

[rsjsouza]

That specification makes no sense.

Has it occurred to you that I have actually tested my scenarios at lower frequencies with exactly matching analog freq / sample ratio? Findings conform exactly to that old EEVBlog video. "Fatal data loss" occurs at about 10 samples per wave for basic sine wave. So in short you are saying that owner of this blog is talking BS in that video. Way to go.

Which video? What 'scope?

Fungus, the video is linked above.

Regarding analog bandwidth versus digital sampling rate, there are many discussions in this forum (here, here and here for example) with varying degrees of experience and opinion, with the offending pissing contest and mud flinging...
 

[nctnico]

Used stuff do not like, only maybe refurbished with new knobs. Do not like knobs that have been rubbed by someone else

Just clean the knobs and they'll look as new. Easy job for the amount of money you save. Perhaps your best paid job ever! If you don't need protocol decoding then there are many choices out there on the second hand market.

Almost forgot - there must be option do display non-interpolated raw waveform. 10 samples per wfm is not much and when interpolated all the useful info will be lost for good.

Sorry but no information gets lost due to interpolation. Interpolation is there to make the signal interpretable with our eyes. Just read about the Nyquist theory and you'll learn that 10 times oversampling is overkill and you won't get any extra information. Just more money gone from your wallet.

Perhaps you could elaborate on your other requirements. You mention FFT but are they other things which are important like connectivity, math, measurements, filtering, etc?

[Someone]

That specification makes no sense.

Has it occurred to you that I have actually tested my scenarios at lower frequencies with exactly matching analog freq / sample ratio? Findings conform exactly to that old EEVBlog video. "Fatal data loss" occurs at about 10 samples per wave for basic sine wave. So in short you are saying that owner of this blog is talking BS in that video. Way to go.

Which video? What 'scope?

Fungus, the video is linked above.

Regarding analog bandwidth versus digital sampling rate, there are many discussions in this forum (here, here and here for example) with varying degrees of experience and opinion, with the offending pissing contest and mud flinging...
 

You forgot this epic thread too: https://www.eevblog.com/forum/testgear/will-keysight-upgrade-the-2000-3000t-x-series-oscilloscopes-within-a-few-months/?all

Almost forgot - there must be option do display non-interpolated raw waveform. 10 samples per wfm is not much and when interpolated all the useful info will be lost for good.

Sorry but no information gets lost due to interpolation. Interpolation is there to make the signal interpretable with our eyes. Just read about the Nyquist theory and you'll learn that 10 times oversampling is overkill and you won't get any extra information. Just more money gone from your wallet.

I would like to live in your magical world where measurements have no noise in them and the front end antialias filters are perfect, the rest of us will happily continue buying scopes with sample rates 5-10x higher than required by nyquist (or more!).