Is a Rigol MSO5000 overkill for a hobbyist?

[nctnico]

Late according to what reference frame?

The real signal's one.

Put your thinking hat on.
A question: what does scope care about that?

Scope has one reference frame: a trigger point.

Not necessarily. The only thing that is fixed is that the signals are sampled using the oscilloscope's clock source (unless using alternate triggering mode). So you can relate signals to eachother within the accuracy limitations of the clock source. Remember that many oscilloscopes have untriggered roll mode as well.

[2N3055]

Late according to what reference frame?

The real signal's one.

Put your thinking hat on.
A question: what does scope care about that?

Scope has one reference frame: a trigger point.

Not necessarily. The only thing that is fixed is that the signals are sampled using the oscilloscope's clock source (unless using alternate triggering mode). So you can relate signals to eachother within the accuracy limitations of the clock source. Remember that many oscilloscopes have untriggered roll mode as well.

I will repeat once again. Scopes time frame of reference is the trigger point. I mean, in internal math of the scope it is literally origin of coordinate system. At least on scopes I know about.
You can correlate signals in relation to internal timing resolution that is orders of magnitude more granular than sampling rate.
Roll mode is a special mode.

[Infraviolet]

The question of what is overkill is surely highly application dependent. Whether a particular scope type is overkill will depend on by how much its performance and features exceed those you need for the things you actually do, or plan to do.

[WaveyDipole]

Wow, didn't realise that many pages had been added since I last looked in!

Thank you for all the comments regarding amplifiers. I have learned a lot from the discussion. I can see how an amplifier might help to visualise the signal, but that is different to measurement which would require a knowledge of the characteristics of the amplifier as well as the scope to understand what one should expect to see. I can see that input impedance, amplifier DB gain and SNR all play a part in determining the result. Because, in addition to working with RF in tuner and IF stages, I am also looking to do some amateur radio work involving antennas and building filters etc, I am beginning to look at this noise issue as somewhat more relevant than I might have appreciated at first.

That still leaves the USB scope market underpopulated and with plenty of room to compete...

They don't seem to be "competing" very well with bench 'scopes in terms of bandwidth/number of channels per $$$ though.

They have a lot less hardware inside them (no screen, no application processor, no internal file system, no networking, no USB stick connector, no HDMI output, no knobs, much simpler case/chassis) so what's the excuse?

This was also explained numerous times: hardware BOM is minor part of development cost. Picoscope software has some features that you need to pay 5 digits to get otherwise..

Also since last time I wrote about this Pico added 3 new protocol decodes.. For free.
Actually, since I bought my PS3406D MSO few years ago it gained 6-7 protocol decodes, Deep measure etc...

Assuming money was no object, would you actually buy this PicoScope over the Rigol MSO5000? I have a genuine reason for asking. As I understand it, the PicoScope has the more refined software, yet both are 8-bit. The PicoScope can do Bode plots, although I am not sure whether it can do Fourier transforms. They both have a 16-channel LA and a sig gen, but how would the PicoScope fare with that 2mV signal example?

BTW, the comparisons were informative and interesting. Thank you to those who took the time and trouble to provide them. Of course, one might expect to see an improvements when going from an 8 bits to 10 or 12 bits, but even the 8-bit comparisons  were quite stark.

One feature I want in Oscope is hands free operation.  I have thought about doing this with some Python and MSO5000. 

By hands free I mean I want to use voice commands like "Single Shot", "Hold", "Autoset", "Zoom in", "Zoom out", "Reset Trigger", etc.  Wording and functionality might change but I think you get the idea.

There is nothing worse than probing something using microscope and trying to stop scope, autoset scope.  The worse is when scope probe slips as you try to one handed hold the probe and ground and blow up a chip.

I can just imagine the scene: Alexa zoom in to.... Siri reset trigger.... Cotrana stop scope.... Sorry Dave I cannot do that....
Only kidding of course, but I can see how that scenario with a microscope might be a pain. I wonder whether this type of thing might be useful to assist with working hands-free:
https://sensepeek.com/

[balnazzar]

Assuming money was no object, would you actually buy this PicoScope over the Rigol MSO5000? I have a genuine reason for asking. As I understand it, the PicoScope has the more refined software, yet both are 8-bit. They both have an LA and a sig gen, but how would the PicoScope fare with that 2mV signal example?

If 2N3055 can spare the time, he could test this. I'm interested too.
He has 3 or 4 Picos, AFAIK. I'd be interested mainly in the 2000 and 3000 series.

[nctnico]

Late according to what reference frame?

The real signal's one.

Put your thinking hat on.
A question: what does scope care about that?

Scope has one reference frame: a trigger point.

Not necessarily. The only thing that is fixed is that the signals are sampled using the oscilloscope's clock source (unless using alternate triggering mode). So you can relate signals to eachother within the accuracy limitations of the clock source. Remember that many oscilloscopes have untriggered roll mode as well.

I will repeat once again. Scopes time frame of reference is the trigger point. I mean, in internal math of the scope it is literally origin of coordinate system. At least on scopes I know about.
You can correlate signals in relation to internal timing resolution that is orders of magnitude more granular than sampling rate.
Roll mode is a special mode.

You don't need a trigger point to correlate signals beyond the time resolution of the internal sample interval. And roll-mode isn't special. It is just letting the aqcuisition system run freely (and stop it to examine the signal in more detail). Just think about this for a day before shooting from the hip and missing the point. Even in roll mode, there is a point of origin but this is just as arbitrary as a trigger point.

[tautech]

The question of what is overkill is surely highly application dependent. Whether a particular scope type is overkill will depend on by how much its performance and features exceed those you need for the things you actually do, or plan to do.

Yes and no.
How can someone with limited DSO experience really know what they need or don't.
One day you might be doing the simplest blah blah blah scope work then the next diving into electronics that is new to you and therefore challenge your and your instruments capabilities.

The 'plan to do' you mention is heading in the right direction but doesn't tell the full story of how we grow and in some cases outgrow our scopes/instrument capabilities.
How would you address this better, well a better/bigger budget helps but that too doesn't address the an instrument can or can't do which is why the datasheet is our #1 guiding document followed by the User manual then followed by the likes of feedback from EEVblog members that actually own/use the instruments in question.

At one time Dave's vids were good guidance however so many are now outdated as new feature set and functionality get added and/or he didn't test the more advanced instrument features.
So we come back to this forum and the wealth of knowledge we collectively have.

[Fungus]

There's a forum member "tggzzz" with an applicable sig:

Glider pilot's aphorism: "there is no substitute for span".
Retort: "There is a substitute: skill+imagination. But you can buy span".

And in this case his sig means: you can try to twiddle and invent all kind of tricks, but you can simply buy a scope that works well in a first place.

Alternatively: If you've got that skill+imagination and it's only for occasional use you can have $600 worth of other equipment for free!

[Fungus]

TL;DR is that the AD2 seems to be incapable of managing signals <~20mVpp. I'd like you to tell me why.

It's very clearly stated in the manual.

It has a 5V measurement range and a 14-bit ADC. Resolving single mV is on the limits of its resolution.

The staircase wave I threw at the AD2 is two Vpp. Not mVpp. It's not a low signal level.

But what's the length of one 'step' compared to one sample interval?  Tell us how it's supposed to show horizontal steps with that ratio?

Does the Siglent perhaps have a smaller sample interval?

But look at that.. It's a staircase wave at TWO Vpp. It souldn't have problems with it, right?
Wrong, the staircase is not even distinguishable.

I don't think the voltage is the problem.

That's not acceptable, I'm considering returning this little sh*t.

But you were so happy with it a few days ago.

This is turning into a comedy. 

[egonotto]

Hello,

few samples from a PicoScope 5243A with a bandwith 200 MHz.

Best Regards
egonotto

[balnazzar]

TL;DR is that the AD2 seems to be incapable of managing signals <~20mVpp. I'd like you to tell me why.

It's very clearly stated in the manual.

It has a 5V measurement range and a 14-bit ADC. Resolving single mV is on the limits of its resolution.

The staircase wave I threw at the AD2 is two Vpp. Not mVpp. It's not a low signal level.

But what's the length of one 'step' compared to one sample interval?  Tell us how it's supposed to show horizontal steps with that ratio?

Does the Siglent perhaps have a smaller sample interval?

But look at that.. It's a staircase wave at TWO Vpp. It souldn't have problems with it, right?
Wrong, the staircase is not even distinguishable.

I don't think the voltage is the problem.

That's not acceptable, I'm considering returning this little sh*t.

But you were so happy with it a few days ago.

This is turning into a comedy. 

Avoiding to quote piecewise..:

1. Not single mV. It starts to have problems under 50 mV, and it sucks completely under 20 mV.

2. Please elaborate about this steps thingy. I'm not sure I understand.

3. What's the problem, then?

4a. I didn't test it against the right (that is, wrong) signals. I update my evaluation as soon as I acquire new information & experience.

4b. Tragedy, instrumentally speaking.

Another thing is that the software is so good that one gets lured by it, not paying enough attention to the instrument itself.

[balnazzar]

Hello,

few samples from a PicoScope 5243A with a bandwith 200 MHz.

Best Regards
egonotto

These flexible ADCs make A LOT of difference. Maybe you can test it with a more irregular wave?

[egonotto]

These flexible ADCs make A LOT of difference. Maybe you can test it with a more irregular wave?

Hello,

few new samples from a PicoScope 5243A with a bandwith 200 MHz.

Best Regards
egonotto

PS: At tpico_16 plus 4_20MHz.jpg, the limit was exceeded. That is no longer possible.

[jasonRF]

This is mostly repeat, as it is with a pico 5244B.  I am only posting it because I think I must have a lower noise signal than used in the other plots.  The plots below are made using a tinySA as the source, generating a nominally -50 dBm (1 mV peak) sinusoid at 8 MHz.  It is connected to the scope using the SMA cable that came with the tinySA and a cheap SMA-to-BNC adapter.  These are single captures using the most sensitive 2mV/div scale (I have zoomed the scales for the images). 

The first two are with the full 200 MHz bandwidth at 8 and 12 bits, and the second two are with the 20 MHz bandwidth limiter turned on.   

jason

[Fungus]

2. Please elaborate about this steps thingy. I'm not sure I understand.

3. What's the problem, then?

Clue #2: I believe what you're seeing is a lack of horizontal resolution, not a lack of vertical resolution.

[WaveyDipole]

I have been comparing the datasheets for the PicoScope 3406D and the Rigol MSO5000 series and they seem neck to neck in a number of ways, but there are some differences about which I am curious.

Firstly there is the sample rate which is 1Gs/s on the PicoScope with only 250Ms/s per channel vs 8Gs/s and 2Gs/s per channel. For the price point, the PicoScope seems rather poor for a scope costing over 2k?

Secondly, there seems to be no memory depth specification for the PicoScope. Is this because of its different architecture, the software running on a PC rather than on hardware inside the instrument? I.e is the memory depth bound only by the amount of PC memory available as opposed to the amount of physical memory installed in the instrument?

Thirdly, the probes. Generally the standard for input impedance is 1MΩ. On the PicoScope this appears to be 200kΩ. Wouldn't this mean that probes are not interchangeable and one would have to use only PicoScope provided probes? For example, wouldn't a "standard" 10:1 probe become something like a 50:1 probe on the PicoScope?

[2N3055]

I have been comparing the datasheets for the PicoScope 3406 and the Rigol MSO5000 series and they seem neck to neck in a number of ways, but there are some differences about which I am curious.

Firstly there is the sample rate which is 1Gs/s on the PicoScope with only 250Ms/s per channel vs 8Gs/s and 2Gs/s per channel. For the price point, the PicoScope seems rather poor for a scope costing over 2k?

Secondly, there seems to be no memory depth specification for the PicoScope. Is this because of its different architecture, the software running on a PC rather than on hardware inside the instrument? I.e is the memory depth bound only by the amount of PC memory available as opposed to the amount of physical memory installed in the instrument?

Thirdly, the probes. Generally the standard for input impedance is 1MΩ. On the PicoScope this appears to be 200kΩ. Wouldn't this mean that probes are not interchangeable and one would have to use only PicoScope provided probes? For example, wouldn't a "standard" 10:1 probe become something like a 50:1 probe on the PicoScope?

Hem, what datasheet are you looking at?

https://www.picotech.com/oscilloscope/3000/picoscope-3000-oscilloscope-specifications

First, current model is 3406D.
Just a quick excerpt:

Analog Input:  impedance   1 MΩ ±1%, in parallel with 14 pF ±1 pF. 
Digital inputs: Input impedance 200 kΩ ±2% ∥ 8 pF ±2 pF
Buffer memory: 512 MS

[WaveyDipole]

The very same datasheet, but I omitted the 'D' in my post. I have also made a mistake regarding the input impedance data. There are two separate tables, one listing analog specifications and the other the digital specifications. I obviously took that from the wrong table.

I wasn't sure whether "buffer memory" corresponded to "Max. Memory Depth" or not. One is specified in MS, the other in Mpts and I am not sure whether they correlate with each other. Assuming that a "point" is the same as a "sample" and MS stands for mega samples, then its '500' '512' vs '200 (optional)' then this would just be a different way of stating the same statistic. But is my assumption correct?

The only thing that would stand out as different then would be the sample rate with the PicoScope capturing 100,000 waveforms/s against the MSO5000's 450,000. I would appreciate a little help in understanding what 'Maximum sampling rate' vs 'Maximum effective sampling rate' in the attached snapshot means please?
 

[2N3055]

The very same datasheet, but I omitted the 'D' in my post. I have also made a mistake regarding the input impedance data. There are two separate tables, one listing analog specifications and the other the digital specifications. I obviously took that from the wrong table.

I wasn't sure whether "buffer memory" corresponded to "Max. Memory Depth" or not. One is specified in MS, the other in Mpts and I am not sure whether they correlate with each other. Assuming that a "point" is the same as a "sample" and MS stands for mega samples, then its '500' vs '200 (optional)' then this would just be a different way of stating the same statistic. But is my assumption correct?

The only thing that would stand out as different then would be the sample rate with the PicoScope capturing 100,000 waveforms/s against the MSO5000's 450,000. I would appreciate a little help in understanding what 'Maximum sampling rate' vs 'Maximum effective sampling rate' in the attached snapshot means please?

Scope's sampling memory is specified in number of samples not bytes. For instance 12bit Siglent scopes use 2 bytes per sample... So number of sample points you can fit in internal memory. Samples and points are equivalent terms in this context. Those 500 MPoints are shared between all channels.

'Maximum effective sampling rate'  is name for ETS sampling Picoscope supports. On repetitive signal you can achieve equivalent of 25GS/s sampling rate. Signal has to be stable and very self correlated for that to work well.
Realtime sample rate is 1GS/s total for all channels.

[Jay_Diddy_B]

WaveyDipole,

The way to read this specification is:

1) They have a single ADC that converts at a rate of 1 GS/s that is shared between the analog channels. They could have four ADC that are sampling at 250MS/s.

2) They have the option to interleave the sampling or multiplex the ADC input depending on the arrangement of the ADCs.

3) The raw ADC performance determines the ability to capture a single trigger event. The waveform is reconstructed from the ADC samples. Often a sinc function is used.

4) If the input signal is repetitive, for example a continuous sinewave, multiple triggering events, shifted in time by a fraction of the sampling interval can be used to obtain more points. This is why the Maximum Effective sampling rate only applies to repetitive signals.

Jay_Diddy_B

[balnazzar]

As a matter of fact, the 2K and 3K Picos seem indeed to be overpriced with respect to many bench scopes with similar specs. I don't know why. Maybe it's that they are the de facto usb monopolysts, or that the software is much better than the software coming with benchtops.
OTOH the analog discovery pro (4ch) is just 250 MSa/s and costs 1600 eur vat included.

Other models do have peculiarities that make them expensive. E.g. the flexible ADCs on the 5K series. All those bits are quite effective, as demonstrated by the tests performed by egonotto and jason.
There are others with separately grounded channels.

[tautech]

OTOH the analog discovery pro (4ch) is just 250 MSa/s and costs 1600 eur vat included.

A SDS2104X Plus, 10" touch display 2 GSa/a and 200Mpts is a good bit cheaper than that.
And you can put on a big screen via its webserver.

[JeremyC]

Hello,

The Analog Discovery has only 2 sensitivity levels. Everything else is done using digital zoom. If I remember correctly, only the settings 500 mV/div and 5 V/div have the best resolution

Best Regards
egonotto

That's not a good thing. Even the Owon vds1022i does much better at a few microvolts, no matter being 4X cheaper.

500 mV are *a lot*.

Keep in mind that with 14bit resolution you will get 30.5uV resolution in the 500mV range.
The AD2 has about 5mVp2p and 0.63mVrms noise in 2mV/div range (checked using the fly wire, +->gnd and ->gnd) and it's probably affecting your trace.

[jasonRF]

Hello,

The Analog Discovery has only 2 sensitivity levels. Everything else is done using digital zoom. If I remember correctly, only the settings 500 mV/div and 5 V/div have the best resolution. 

Best Regards
egonotto

That's not a good thing. Even the Owon vds1022i does much better at a few microvolts, no matter being 4X cheaper.

500 mV are *a lot*.

Keep in mind that with 14bit resolution you will get 30.5uV resolution in the 500mV range.
(Attachment Link)

You missed a factor of 10.  The spec says 320 uV resolution.  A back of the envelope gives 5V/16384 = 305 uV. 

[balnazzar]

OTOH the analog discovery pro (4ch) is just 250 MSa/s and costs 1600 eur vat included.

A SDS2104X Plus, 10" touch display 2 GSa/a and 200Mpts is a good bit cheaper than that.
And you can put on a big screen via its webserver.

The ADP is 14 bit but... 4 channels at 250 MSa/s in total? No thanks. Of course the Siggy is better in terms of price for specs and features. Even a Pico 2408 is better. Or the HDO1000.