Keysight DSOX3014T vs DSOX1204G - is it really worth 3.5 times the money? Vote!

[Bussmann]

Hi,

A noob here. Thinking of getting a new scope for mostly audio use. No need for decoding capabilities. Still use a Tek 2247A that is great for waveform display but not accurate in measurements with anything lower than 100mV.

For the new scope, I will need 4 channels, responsive UI, accurate measurements and a robust build to last a few years. Tried the Rigol, Siglent and even R&S RTB but did not like them: the Rigol layout was busy and distracting, the Siglent encoders were mushy and overshoots most of the time, and the RTB had a strange noise floor. Tried the PC-based stuff too and did not like them not having dials.

So, for a purchase that will need to endure the next 10-20 years, is the 3014T worth the extra money when the 1204G provides most of what it can do for less than 1/3 of the money? Does "you get what you pay for" apply here?

Thanks.

[tautech]

Hi,

A noob here. Thinking of getting a new scope for mostly audio use. No need for decoding capabilities. Still use a Tek 2247A that is great for waveform display but not accurate in measurements with anything lower than 100mV.

For the new scope, I will need 4 channels, responsive UI, accurate measurements and a robust build to last a few years. Tried the Rigol, Siglent and even R&S RTB but did not like them: the Rigol layout was busy and distracting, the Siglent encoders were mushy and overshoots most of the time, and the RTB had a strange noise floor. Tried the PC-based stuff too and did not like them not having dials.

Welcome to the forum.

Yes, first time use of a DSO after only using CRO's can be a bit awkward as the controls take no effort to rotate.
What model Siglent did you trial ?

[Mr. Scram]

Hi,

A noob here. Thinking of getting a new scope for mostly audio use. No need for decoding capabilities. Still use a Tek 2247A that is great for waveform display but not accurate in measurements with anything lower than 100mV.

For the new scope, I will need 4 channels, responsive UI, accurate measurements and a robust build to last a few years. Tried the Rigol, Siglent and even R&S RTB but did not like them: the Rigol layout was busy and distracting, the Siglent encoders were mushy and overshoots most of the time, and the RTB had a strange noise floor. Tried the PC-based stuff too and did not like them not having dials.

So, for a purchase that will need to endure the next 10-20 years, is the 3014T worth the extra money when the 1204G provides most of what it can do for less than 1/3 of the money? Does "you get what you pay for" apply here?

Thanks.

The 3014T seems to mostly make sense for higher bandwidths and MSO functionality. The 1000X series does offer ridiculous value for money and seems to be really pushing the boundaries of a high end experience at budget segment prices.

[thm_w]

R&S RTB or Siglent (you didn't specify which model) are going to be the best bet for audio measurements.
If you let some small UI difference or "strange noise floor" (which is what, high RMS noise? was it outside of the spec?) distract you, then you are selling yourself short.

Rigol MSO5000 is a great choice for digital, so would not fit your use case. 

Spending a ton of money up front and expecting a scope to last 10-20 years is also misguided IMO. Get the 1204G if it fits your unique requirements, then in 5-10 years a replacement will be released and you can upgrade to that with the money saved.

[Bussmann]

Hi tautech,

Thanks for the welcome. I actually splurged and got the rebatched 1104 version for double the money hoping to get local support. Attached is a picture of the pubs that came with the T3D version. The scope looked identical except for the badge on the bezel and there is no Siglent logo on the screen. There was no logo period.

Hi Mr. Scram,

I asked the question as I suspect you are right. If you have a 1000X on hand, can you comment on the feel of the knobs - detent or no detent, feel like 3000T or feel cheap? It was a major reason I did not keep the Siglent.

Hi thm_w

RTB2004 and T3DSO1104, the re-badged version of the SDS1104X-E.

The "strange noise floor" is one that varied by nearly 70µV over the course of 2 hours during use, but there was no published noise specs for the RTB so it cannot be considered "out of spec". It varied between 30µV and 100µV RMS AC1M no BWL.

Ergonomics is a big deal to me so imprecise encoders and sloppy switches do not pass the test.

Thanks to everyone for your comments.

[Mr. Scram]

Dave comments on the knobs and indents being the same as on the 2000X and 3000X series around the 1:40 mark.

https://youtu.be/9KcOQsVxtoU

[Bussmann]

Thank you Mr. Scram! Nice.

Dave comments on the knobs and indents being the same as on the 2000X and 3000X series around the 1:40 mark.

https://youtu.be/9KcOQsVxtoU

[tautech]

Hi tautech,

Thanks for the welcome. I actually splurged and got the rebatched 1104 version for double the money hoping to get local support.

You already have great support in Ohio:
https://www.siglentamerica.com/

It's unknown how up to date LeCroy are with SDS1*04X-E firmware as there's been several versions that have addressed self cal accuracies and added lots of additional functionality since the X-E's were released.

On the encoders, I've just reached across to my pre-release beta unit and I notice they are no different to any I do pre-despatch checks on......detents are quite well enough defined however use style does have an affect on 'felt' feedback. The modern compact DSO can be best used for some operations with three fingers atop of the unit with thumb and forefinger rotating the controls and that better hand support promotes more precise adjustments where the larger and heavier scopes often don't require this operation style.
Encoder knob size can have a bearing on this too but all in all it's related to previous CRO use where knobs take some effort to rotate which is not the case for DSO's.

EEVblog (Dave) has done several videos on 'noise on DSO's' and a couple more recent ones on this baseline noise 'real' measurement thank you might want to watch:

[Mr. Scram]

In case it had gone unnoticed I'll quote Dave.

Be aware that Tautech is a Siglent distributor, so he has a vested interest in talking up Siglent and does so at every opportunity on the forum, he's (in)famous for it.

[Bussmann]

HI tautech,

Thank you so much for the pointers!

The touch thing is kinda personal and I know what works for me may not be good for anyone else. Perhaps the most important to me to know that an 1/2/5 adjustment means 3 clicks that i can feel without looking at the scope. I was not able to feel this on the T3/Siglent. I could on the RTB, and of course, on the Tek, you know you clicked it.

There are now a few videos to watch so thank you. I am glad I signed up here!

Hi Mr. Scram,

I am a noob so I just soak it up. I thank everyone for helping!!

Hi tautech,

Thanks for the welcome. I actually splurged and got the rebatched 1104 version for double the money hoping to get local support.

You already have great support in Ohio:
https://www.siglentamerica.com/

It's unknown how up to date LeCroy are with SDS1*04X-E firmware as there's been several versions that have addressed self cal accuracies and added lots of additional functionality since the X-E's were released.

On the encoders, I've just reached across to my pre-release beta unit and I notice they are no different to any I do pre-despatch checks on......detents are quite well enough defined however use style does have an affect on 'felt' feedback. The modern compact DSO can be best used for some operations with three fingers atop of the unit with thumb and forefinger rotating the controls and that better hand support promotes more precise adjustments where the larger and heavier scopes often don't require this operation style.
Encoder knob size can have a bearing on this too but all in all it's related to previous CRO use where knobs take some effort to rotate which is not the case for DSO's.

EEVblog (Dave) has done several videos on 'noise on DSO's' and a couple more recent ones on this baseline noise 'real' measurement thank you might want to watch:

[Helix70]

Hi,

I use the MSOX3014T on a daily basis. We also have a 1000x series. To answer the title of the thread, Yes. It is really worth it - *If you use it*.

The bandwidth is largely irrelevant. I only have, and need, 100Mhz for my use. But 1M waveforms per second means it has the horsepower to process all the waveforms, so you don't miss the trigger. There are many videos that give more details of the features.

The touch screen is very useful and intuitive, is was more probe options, can be updated more if thats your thing, does far better with decoding serial, much better UI in general.

Basically, the way i feel about my MSO3014T is you would need to pry it from my cold dead hands. But it isn't for the shelf. It wants and needs to be used to be justified.

[Bussmann]

Hi there,

Thank you for your feedback. This is what I wanted to know. I also read the manual of the 1000X in the meantime, and found that the built-in measurements is a subset of the others. Make sense for a bargain but I'd better to sure before I leap and get something short of a dollar!

Thanks again.

Hi,

I use the MSOX3014T on a daily basis. We also have a 1000x series. To answer the title of the thread, Yes. It is really worth it - *If you use it*.

The bandwidth is largely irrelevant. I only have, and need, 100Mhz for my use. But 1M waveforms per second means it has the horsepower to process all the waveforms, so you don't miss the trigger. There are many videos that give more details of the features.

The touch screen is very useful and intuitive, is was more probe options, can be updated more if thats your thing, does far better with decoding serial, much better UI in general.

Basically, the way i feel about my MSO3014T is you would need to pry it from my cold dead hands. But it isn't for the shelf. It wants and needs to be used to be justified.

[nctnico]

Make sure to try the R&S RTM3000 series as well. IMHO the 10 bit ADCs do offer an advantage over 8 bit for analog signals especially when you want to zoom in on details.

[Bussmann]

Will do. Thanks for the suggestion.

Make sure to try the R&S RTM3000 series as well. IMHO the 10 bit ADCs do offer an advantage over 8 bit for analog signals especially when you want to zoom in on details.

[nez]

Hi!  My own noob questions out of curiosity:)

It seem's that 10 bit resolution would be quite useful as nctnico mentioned.

But how about waveforms per second?  If one is working in the audio frequency range (we'll say only audio, for sake of discussion), what's the approximate ceiling on useful waveforms/sec, and what would be overkill?

My other question is on averaging to clear the noise.  Is the audio-centric workflow straightforward to the point where you'd want to use averaging to handle noise (i.e. the signal shape you analyze is repeated with no cycle-to-cycle variation)?  Or, is it actually more complex than that, and averaging would not really be used?

[Bussmann]

Well, the math would say 10-bit ADCs should have an advantage, but practically for audio, even 10-bit is not ideal for frequency domain. I have a 24 192 sound card for that.

And I think you nailed in on the head: what analog scopes take for granted easily showing XY waveforms most DSO simply struggle at. Only DSOs with very fast waveform update rate like Keysights can do it easily. Hence the original question - is 1204G fast enough to do a good XY or is the 1,000,000 rate on the DSOX3014T really necessary? No answer yet so far.

Averaging is the other thing because noise at low frequencies are particularly challenging - both for display and triggering. So far, not sure there is a way out of spending the big bucks for the T.

Hi!  My own noob questions out of curiosity:)

It seem's that 10 bit resolution would be quite useful as nctnico mentioned.

But how about waveforms per second?  If one is working in the audio frequency range (we'll say only audio, for sake of discussion), what's the approximate ceiling on useful waveforms/sec, and what would be overkill?

My other question is on averaging to clear the noise.  Is the audio-centric workflow straightforward to the point where you'd want to use averaging to handle noise (i.e. the signal shape you analyze is repeated with no cycle-to-cycle variation)?  Or, is it actually more complex than that, and averaging would not really be used?

[nctnico]

Hi!  My own noob questions out of curiosity:)

It seem's that 10 bit resolution would be quite useful as nctnico mentioned.

But how about waveforms per second?  If one is working in the audio frequency range (we'll say only audio, for sake of discussion), what's the approximate ceiling on useful waveforms/sec, and what would be overkill?

My other question is on averaging to clear the noise.  Is the audio-centric workflow straightforward to the point where you'd want to use averaging to handle noise (i.e. the signal shape you analyze is repeated with no cycle-to-cycle variation)?  Or, is it actually more complex than that, and averaging would not really be used?

Waveforms/s is a way overhyped number for marketing purposes. Remember that a waveform needs to be captured first before display so it is a function of samplerate versus time/div; therefore the highest waveforms/s are achieved at a very specific time/div setting. Anything over 1000 waveforms/s is basically overkill. Now some people will say that with low waveforms/s you won't catch glitches but if you are hunting for glitches a much better way is to use mask testing, longer memory (show many cycles) or a trigger. That way you don't need to stare at the screen without blinking.

Back to audio use: you can use averaging but then the signal needs to be repetitive indeed. A more useful mode is high-res which is even better when starting from 10 bits instead of 8.

[Mr. Scram]

Some critics have argued the waveform update rate isn't a relevant metric, but the numerous manufacturers emphasizing the improved waveform update rates of their new product ranges suggest the market may feel differently. That being said it does appear a high waveform update rate is useful for a specific subset of problems and I'm not sure audio related work is part of that subset. Are incidental glitches a problem in the audio world, other than in digital buses? Don't expect the XY mode of any DSO to be as good as those on CROs, Keysight or any other brand. DSOs outperform CROs in pretty much every metric but XY mode appears to be an exception.

[Someone]

Some critics have argued the waveform update rate isn't a relevant metric, but the numerous manufacturers emphasizing the improved waveform update rates of their new product ranges suggest the market may feel differently. That being said it does appear a high waveform update rate is useful for a specific subset of problems and I'm not sure audio related work is part of that subset. Are incidental glitches a problem in the audio world, other than in digital buses?

Its a little more nuanced than just banner specifications and biggest number is always best. Its important to take in all the characteristics of the scope, for example the waveform rate of a Tek DPO4000:

The solid line for each memory depth is where the full memory fits on the screen and there is no aliasing. Dotted lines off to the right the visible window is a short segment of the set memory depth, while off to the left is where aliasing occurs. Most audio work would be down around the 1E-2 to 1E-4 seconds per division region where the update rate is dominated by the acquisition and not the blind/turnaround time... except the longer memory depths which are required to avoid aliasing on that scope.

[EEVblog]

Hi,

A noob here. Thinking of getting a new scope for mostly audio use. No need for decoding capabilities. Still use a Tek 2247A that is great for waveform display but not accurate in measurements with anything lower than 100mV.

For the new scope, I will need 4 channels, responsive UI, accurate measurements and a robust build to last a few years. Tried the Rigol, Siglent and even R&S RTB but did not like them: the Rigol layout was busy and distracting, the Siglent encoders were mushy and overshoots most of the time, and the RTB had a strange noise floor. Tried the PC-based stuff too and did not like them not having dials.

So, for a purchase that will need to endure the next 10-20 years, is the 3014T worth the extra money when the 1204G provides most of what it can do for less than 1/3 of the money? Does "you get what you pay for" apply here?

For your needs the 1204G is plenty, but it's not the best for low level signals. A scope with a true 500uV range might be what you are after for audio work.
It uses the exact same Megazoom IV ASIC and the 3000 and 4000 series does, so apart from the front end bandwith the hardware is essentially the same (minus touch screen).
What they do is limit in software the memory and features available in that ASIC.
All the manufacturers limit products like this, it's about market segmentation.

[EEVblog]

EEVblog (Dave) has done several videos on 'noise on DSO's' and a couple more recent ones on this baseline noise 'real' measurement thank you might want to watch:

[Bussmann]

Everyone has valid points and this noob is no closer to a decision. So I am asking for votes to see if the numbers tell a clear picture.

Hi,

A noob here. Thinking of getting a new scope for mostly audio use. No need for decoding capabilities. Still use a Tek 2247A that is great for waveform display but not accurate in measurements with anything lower than 100mV.

For the new scope, I will need 4 channels, responsive UI, accurate measurements and a robust build to last a few years. Tried the Rigol, Siglent and even R&S RTB but did not like them: the Rigol layout was busy and distracting, the Siglent encoders were mushy and overshoots most of the time, and the RTB had a strange noise floor. Tried the PC-based stuff too and did not like them not having dials.

So, for a purchase that will need to endure the next 10-20 years, is the 3014T worth the extra money when the 1204G provides most of what it can do for less than 1/3 of the money? Does "you get what you pay for" apply here?

For your needs the 1204G is plenty, but it's not the best for low level signals. A scope with a true 500uV range might be what you are after for audio work.
It uses the exact same Megazoom IV ASIC and the 3000 and 4000 series does, so apart from the front end bandwith the hardware is essentially the same (minus touch screen).
What they do is limit in software the memory and features available in that ASIC.
All the manufacturers limit products like this, it's about market segmentation.

[Mr. Scram]

How do you intend to use the XY mode?

[EEVblog]

Everyone has valid points and this noob is no closer to a decision. So I am asking for votes to see if the numbers tell a clear picture.

Unfortunately the numbers won't tell you. Small details matter, a lot.
Your need for an X-Y mode is fairly unique. For that why not stick with an analog scope?
I would focus on the low noise audio requirement first, for which the Keysight isn't well suited due to it's software magnified front end.
But it's just one of many (often conflicting) requirements.
For starters, what is your budget? and if you are considering the 3000T it's obviously rather high.
When you have a large budget then other things come into play, like maybe more than one scope for different requirements.
e.g. a general cheap-ish 4CH scope for general daily use + a specialised scope for analog audio performance (could even be a high resolution USB scope) might be a better solution and than s single expensive scope that doesn't do all that you want.

[Bussmann]

@Mr. Scram - XY for channel separation waveform display

@Dave

Here are observations of various scope behavior before I posted my questions:

1. I played with a friend's 14-bit National Instrument PXI-5122 scope and it was dreadfully slow, even though it was supposed to be sampling at 100MS/s.

2. While we know 8-bit and even 10bit scope noise floor is likely inadequate for audio, even the NI 14-bit scope's noise floor was no better than 105dBV. And that's not anywhere near any 24 192 soundcard at -140 to -150dBV.

3. I already have a 24 192 soundcard running ARTA and it worked pretty well for frequency domain work but for time domain, its <400kS/s sampling means very jagged waveforms on DSO, especially when zoomed in a bit.

4. I also already have a Tek 2247a analog scope - it's a hybrid scope - ADCs for measurements and readouts, analog tube for waveform display, but it has no storage, and when zoomed in to see rise time on higher frequencies, the waveform is barely visible.

The quest to find a scope that can do near-analog waveform displays, give better than 0.5% accuracy measurements (the Tek 2247A is like 15% to my eyes) and can zoom in - live or history mode - to see and measure without squinting. The 24 192 soundcard can do all the frequency domain stuff like FFTs, THD and sweeps but they cannot do good waveforms. Therefore my preference is really to see a waveform better than to measure it better.

Budget is under $5k - the DSOX3014T would be a real stretch as it's a lot more than the DSOX3014A and 3.5x that of the DSOX1204G. The RTB seems to give up responsiveness when it's taxed.

Everyone has valid points and this noob is no closer to a decision. So I am asking for votes to see if the numbers tell a clear picture.

Unfortunately the numbers won't tell you. Small details matter, a lot.
Your need for an X-Y mode is fairly unique. For that why not stick with an analog scope?
I would focus on the low noise audio requirement first, for which the Keysight isn't well suited due to it's software magnified front end.
But it's just one of many (often conflicting) requirements.
For starters, what is your budget? and if you are considering the 3000T it's obviously rather high.
When you have a large budget then other things come into play, like maybe more than one scope for different requirements.
e.g. a general cheap-ish 4CH scope for general daily use + a specialised scope for analog audio performance (could even be a high resolution USB scope) might be a better solution and than s single expensive scope that doesn't do all that you want.