What is the "best" Oscilliscope that I can get for $30,000?

[nctnico]

Yup, people also got to work in Model-T Fords. Time moves on and better solutions are avaiable. Why not take advantage of that?

Buy a cheap, slow Chinese scope with lots of memory, and use that for protocol decoding.  Then get a Megazoom scope to go with it.  See which one you actually use most of the time.  It won't be a close contest.

Well, I more-or-less replaced a Keysight DSO7104A with a GW Instek GDS2204E. The latter is just much more convenient to use. Currently I have an R&S RTM3004 on my desk which I use mostly and I'm not going to replace that with a Megazoom Keysight scope as well. That would be a step back in my book. IOW: a lot depends on the actual use case.

[Fungus]

Love your videos Dave.  I would love to see you make one involving designing and building an excellent electronics lab from scratch with an unlimited budget!

That might be tough for him to do since he probably doesn't have an unlimited budget.   

He could use cardboard cutouts like that analog-whatsit guy does in his oscilloscope comparisons on youtube. 

[pcprogrammer]

Guess the overall consensuses is buy a relative cheap scope and hobby with that for the coming years. It will most likely be well enough for most of the project you have in mind.

There is no scope that will do everything you want and that will suffice for the rest of your live.

[Fungus]

As someone with a lab full of fairly expensive scopes, it's the small simple Keysight 1000X that I grab when I just want to measure something.

Yet you'd be doing many a great disservice today recommending a DSO with just 1 Mpts mem depth.

Why? Not everybody needs it.

If you look at the top sellers on Batronix the first six are all Keysights:

https://www.batronix.com/shop/oscilloscopes/DSO.html

(Andnote that the Rigol DS1052E is still outselling all the Siglent models  )

[Domitronic]

Why? Not everybody needs it.

If you look at the top sellers on Batronix the first six are all Keysights:

https://www.batronix.com/shop/oscilloscopes/DSO.html

(Andnote that the Rigol DS1052E is still outselling all the Siglent models  )

How do you know which are the bestsellers? This site doesn't look like its sorted by sales numbers. Rather manufacturer and series.

[nctnico]

Why? Not everybody needs it.

If you look at the top sellers on Batronix the first six are all Keysights:

https://www.batronix.com/shop/oscilloscopes/DSO.html

(Andnote that the Rigol DS1052E is still outselling all the Siglent models  )

How do you know which are the bestsellers? This site doesn't look like its sorted by sales numbers. Rather manufacturer and series.

Indeed. Simply sorted alphabetically.

[Fungus]

How do you know which are the bestsellers? This site doesn't look like its sorted by sales numbers. Rather manufacturer and series.

Because by default they're sorted by "position", not "manufacturer" (use the selector on the right)

Indeed. Simply sorted alphabetically.

Is that why there's Rigols above and below the Siglents in the list?

[Symax]

One question to OP: are You talking about scope only, or scope with set of probes, that allow You to fully enjoy abilities of the scope? Your scope will came with set of passive probes (if any), and those will limit useful bandwidth to about 500MHz. For higher bandwith You will need set of active probes, that in case of four channels may easilly cost 10k (4x2.5k), limiting budget for the scope itself. And what about current probes?

The scope only is quite a bit of an investment.  I figure once I get the base, I can add stuff later as time goes on similar to how I built my first computer with a box of parts back in the 80286 days.  The $30k price tag was an attempt to give a near limitless budget so that I could get the "best" scope of the times.  I fear that with the chip shortages and other world affairs that sometime in the future, chip manufacturing as it stands now will cease, causing some manufacturered goods to become unobtanium.

[Jan Audio]

If you have a supercomputer from the 80s it cost alot.
Now it is worth nothing.

Why invest to much ?
You can have each 2 years the newest 2000,- scope with latest technology.
Over 30 years you have alien scope.

[TheBay]

Best is what fits your use case, abilities and potential immediate future needs, not future needs as you cannot future proof here.
What is best for someone else is not best for you, irrespective of cost.

[nctnico]

One question to OP: are You talking about scope only, or scope with set of probes, that allow You to fully enjoy abilities of the scope? Your scope will came with set of passive probes (if any), and those will limit useful bandwidth to about 500MHz. For higher bandwith You will need set of active probes, that in case of four channels may easilly cost 10k (4x2.5k), limiting budget for the scope itself. And what about current probes?

The scope only is quite a bit of an investment.  I figure once I get the base, I can add stuff later as time goes on similar to how I built my first computer with a box of parts back in the 80286 days.  The $30k price tag was an attempt to give a near limitless budget so that I could get the "best" scope of the times. 

In the end only you can determine what is best for your usage. If you are going to spend several $k then you should get the equipment on loan first so you can try what suits you best. Just make a pro / con list per model and rate what feature you find important. From there make a short list with 2 or 3 units you would like to try.

I fear that with the chip shortages and other world affairs that sometime in the future, chip manufacturing as it stands now will cease, causing some manufacturered goods to become unobtanium.

That sounds rather paranoia. If chip manufacturing stops then the world is in real trouble and the last thing you'll need is an oscilloscope. Such a scenario is highly unlikely due to the way the world economy is intertwined nowadays.

[kcbrown]

Yup, people also got to work in Model-T Fords. Time moves on and better solutions are avaiable. Why not take advantage of that?

Because there are tradeoffs.

When there aren't tradeoffs, then of course you go with what you call the "better solutions".

Look, Dave has an Instek (the GDS-1104B), he has at least a couple of Siglents, he has a Rigol DS-1054Z, he has a couple of Keysights, he has at least one R&S, at least one Tektronix, and probably others.  And yet, the Keysight 1000X is his go-to scope.  Why, if it's so incredibly inferior to the rest?

You yourself have argued how much better the Instek's UI is compared with the other "B-brand" scopes, in large part (as I recall) due to its responsiveness.  I'm arguing that compared with even the Instek, the Keysight's UI is superior, at the very least in responsiveness, and that, combined with its size, is why Dave chooses it to perform off-the-cuff measurements and other things that don't exceed its capabilities, and apparently it's not often that what he needs to do exceeds its capabilities.

You replaced your Keysight because you needed different capabilities.  But that's you.  And despite your arguments in favor of the Instek, your go-to scope is now an R&S, again likely because of its capabilities that you need and that the Instek doesn't have.  But again, that's you, and due to your specific requirements.  What you find most usable depends on what you are actually doing.  The Keysight shines for many typical oscilloscope uses, and doesn't do so well for a few.  Your use cases fall into that latter category.   The plain fact is that Keysight wouldn't be selling nearly as well as it does for the kind of price it commands if it didn't do most things well enough.  Its responsiveness and usability apparently more than make up for the rest.  It's not like the 1000X is some kind of specialty scope (it's rather the opposite, actually -- it's rather basic by today's standards), so the fact that it sells as well as it does for its price means it must be doing something very right.  That something is its usability.  Nothing else about it stands out.


This is all subject to change.  The competition is improving all the time.  But as of now, for responsiveness, the Keysight apparently wins over everything else.  Combine that with a diminutive size and it's clear why Dave selects it above all of the considerable number of other choices he has at his disposal for his typical uses.

[kcbrown]

The scope only is quite a bit of an investment.  I figure once I get the base, I can add stuff later as time goes on similar to how I built my first computer with a box of parts back in the 80286 days. 

Modern scopes aren't like that.  Their hardware is fixed out of the box, and determines the upper limit of what you can get out of any given unit.  The rest is just software.

You're calling the scope "quite a bit" of an investment, but that's only the case if you insist it is.  You can get all the capability you're likely to need for at least the next 5 years for less than $500.  And if you increase your budget to around $1500, you can do so in style. 

The $30k price tag was an attempt to give a near limitless budget so that I could get the "best" scope of the times.  I fear that with the chip shortages and other world affairs that sometime in the future, chip manufacturing as it stands now will cease, causing some manufacturered goods to become unobtanium.

This scenario is essentially impossible.  The only thing that's going to take out chip manufacturing worldwide is a global war, likely thermonuclear.  If that happens, we'll all be back in the stone ages, because it won't just be chip manufacturing that gets taken out, but most manufacturing.

Planning around such an event is not rational, because that's essentially planning everything around the worst possible case.  It's one thing to have contingencies against the worst possible case.  It's quite another thing to plan everything around it.  This is the latter.

No, as bad as things can sometimes look, the bottom line is that you need to have a little faith in your fellow man.  We each help to make things work, and that will continue to be the case.  The chip shortage has occurred due to a number of factors, but even now actions are being taken to compensate for it, e.g. Intel: https://www.cnbc.com/2021/03/23/intel-is-spending-20-billion-to-build-two-new-chip-plants-in-arizona.html

So fear not, and instead go on as if things will continue to be normal-ish, because just about everyone is working to ensure that.

[KE5FX]

Yup, people also got to work in Model-T Fords. Time moves on and better solutions are avaiable. Why not take advantage of that?

Buy a cheap, slow Chinese scope with lots of memory, and use that for protocol decoding.  Then get a Megazoom scope to go with it.  See which one you actually use most of the time.  It won't be a close contest.

Well, I more-or-less replaced a Keysight DSO7104A with a GW Instek GDS2204E. The latter is just much more convenient to use. Currently I have an R&S RTM3004 on my desk which I use mostly and I'm not going to replace that with a Megazoom Keysight scope as well. That would be a step back in my book. IOW: a lot depends on the actual use case.

Hmm, that's actually really interesting.  Not being familiar with the GDS2204E, I looked it up, and it seems to fall far short of the specs of the DSO7104A in some key areas.  Memory depth is a wash (10M versus 8M for the Keysight.)  Seems like a great deal for the price, but that's all that jumps out at me.  So you must really like it to make a statement like that.  What are some of the things you find more convenient about the Instek (or, conversely, intolerable about the Keysight)?

I'd hope the RTM3004 would be a good scope, being a first-tier R&S-branded product that's 10 years or so newer than the Keysight UI and its underlying chipset.  When it comes to scopes, Keysight seems to be resting on their laurels like Tektronix did for a long time, and I wouldn't be surprised if they lose their leadership position in mid-range DSOs as a result.  But not to a 200 MHz 1 GS/s Instek that sells new for $1200.  What am I missing?  Maybe I'm letting some old prejudices do the thinking for me here.

[tautech]

As someone with a lab full of fairly expensive scopes, it's the small simple Keysight 1000X that I grab when I just want to measure something.

Yet you'd be doing many a great disservice today recommending a DSO with just 1 Mpts mem depth.

Indeed. The OP stated protocol decoding and for that you really want deep memory.

The reality is KS 1kX is a 2 GSa/s DSO and in providing 2 GSa/s the capture length is just half that of a 1 GSa/s DSO with the same mem depth. Activate another channel and that's halved again !
The power of a DSO is severely hobbled with such limiting mem depth for capture length.

[KE5FX]

The power of a DSO is severely hobbled with such limiting mem depth for capture length.

The rule of diminishing returns applies.  My TDS 3034 only had a 10K record length (and only then when operated in the right 'mode', another pet peeve.)  But it was a big step up from the 1K-point 2430A it replaced, because the 10K data record helped me get some things done that were much harder to accomplish with 1K points. 

Upgrading to 1M points in the MSO 6034A felt like winning the lottery in comparison.  And yes, I've found that limiting as well... I think it's happened two or three times in the last 10 years.  I got over it. 

I'm sure if I had a 100M record, I'd hit that limit as well, but at that point I'd probably be better off using some other instrument to capture and work with such large amounts of data.  That problem becomes one of streaming data acquisition for offline processing, which in turn becomes a matter of writing some custom software for decoding whatever data I'm dealing with, or perhaps recording complex data from an SDR.  I think the newer Tek scopes can actually do that sort of thing, which brings us back to the $30,000-and-up topic.

[IAmBack]

One question to OP: are You talking about scope only, or scope with set of probes, that allow You to fully enjoy abilities of the scope? Your scope will came with set of passive probes (if any), and those will limit useful bandwidth to about 500MHz. For higher bandwith You will need set of active probes, that in case of four channels may easilly cost 10k (4x2.5k), limiting budget for the scope itself. And what about current probes?

The scope only is quite a bit of an investment.  I figure once I get the base, I can add stuff later as time goes on similar to how I built my first computer with a box of parts back in the 80286 days.  The $30k price tag was an attempt to give a near limitless budget so that I could get the "best" scope of the times.  I fear that with the chip shortages and other world affairs that sometime in the future, chip manufacturing as it stands now will cease, causing some manufacturered goods to become unobtanium.

So, you are looking for doomsday scope
Maybe you should start with power generator with engine capable to run on alcohol, and instruments to produce this fuel?
In case of shortage of chips you'll need few more tools than a scope itself. Maybe several sets, hidden in different places...
And don't forget to prepare your own pcb manufacturing line together with all needed materials. Don't forget about components needed to build/repair stuff...
If a great "shortage of ewerything" is going to happen, your 30k scope won't be very useful. BTW, I've survived almost decade of such world, where lack of ewerything is something normal - thanks to political system (beware of the left side). I don't think, that advanced tools (as of these times) could help us, but basic and simple equipment and materials were wery helpful.

[tautech]

The power of a DSO is severely hobbled with such limiting mem depth for capture length.

The rule of diminishing returns applies.  My TDS 3034 only had a 10K record length (and only then when operated in the right 'mode', another pet peeve.)  But it was a big step up from the 1K-point 2430A it replaced, because the 10K data record helped me get some things done that were much harder to accomplish with 1K points. 

Upgrading to 1M points in the MSO 6034A felt like winning the lottery in comparison.  And yes, I've found that limiting as well... I think it's happened two or three times in the last 10 years.  I got over it. 

I'm sure if I had a 100M record, I'd hit that limit as well, but at that point I'd probably be better off using some other instrument to capture and work with such large amounts of data.  That problem becomes one of streaming data acquisition for offline processing, which in turn becomes a matter of writing some custom software for decoding whatever data I'm dealing with, or perhaps recording complex data from an SDR.  I think the newer Tek scopes can actually do that sort of thing, which brings us back to the $30,000-and-up topic.

We need remember a DSO can be the whole acquisition system or just part of it.
Analysis can be performed within the scope or offline and capture depth/length is very important.
Comparatively cheap scopes offering 100+ Mpts are common today yet you act like this is something special when it is the direction the industry is heading and makes perfect sense where the modern DSO is far more than a tool for just looking at waveforms.

[KE5FX]

Comparatively cheap scopes offering 100+ Mpts are common today yet you act like this is something special when it is the direction the industry is heading and makes perfect sense where the modern DSO is far more than a tool for just looking at waveforms.

Point being, I hardly ever need that, and when I do, I need something besides an oscilloscope to acquire and process the data.  Obsession over record length is an objectively terrible way for most people to choose a scope.

[tautech]

Comparatively cheap scopes offering 100+ Mpts are common today yet you act like this is something special when it is the direction the industry is heading and makes perfect sense where the modern DSO is far more than a tool for just looking at waveforms.

Point being, I hardly ever need that, and when I do, I need something besides an oscilloscope to acquire and process the data.  Obsession over record length is an objectively terrible way for most people to choose a scope.

Interesting when you said getting a scope with larger mem depth was like winning the lottery so why you would not include a more capable capture length as a high priority specification and just one part of choosing a DSO ? 

[nctnico]

Yup, people also got to work in Model-T Fords. Time moves on and better solutions are avaiable. Why not take advantage of that?

Buy a cheap, slow Chinese scope with lots of memory, and use that for protocol decoding.  Then get a Megazoom scope to go with it.  See which one you actually use most of the time.  It won't be a close contest.

Well, I more-or-less replaced a Keysight DSO7104A with a GW Instek GDS2204E. The latter is just much more convenient to use. Currently I have an R&S RTM3004 on my desk which I use mostly and I'm not going to replace that with a Megazoom Keysight scope as well. That would be a step back in my book. IOW: a lot depends on the actual use case.

Hmm, that's actually really interesting.  Not being familiar with the GDS2204E, I looked it up, and it seems to fall far short of the specs of the DSO7104A in some key areas.  Memory depth is a wash (10M versus 8M for the Keysight.)  Seems like a great deal for the price, but that's all that jumps out at me.  So you must really like it to make a statement like that.  What are some of the things you find more convenient about the Instek (or, conversely, intolerable about the Keysight)?

I'd hope the RTM3004 would be a good scope, being a first-tier R&S-branded product that's 10 years or so newer than the Keysight UI and its underlying chipset.  When it comes to scopes, Keysight seems to be resting on their laurels like Tektronix did for a long time, and I wouldn't be surprised if they lose their leadership position in mid-range DSOs as a result.  But not to a 200 MHz 1 GS/s Instek that sells new for $1200.  What am I missing?  Maybe I'm letting some old prejudices do the thinking for me here.

Where the Instek wins is the menu structure. Like the Tektronix TDS500/600/700 series it has a vertical and horizontal row of buttons next to the screen. The horizontal buttons select the menu and the vertical buttons shows the sub-menu with the current settings. This gives the user a very clear overview of what settings are where. Also the GW Instek has a seperate select button. What drove me mad on the DSO7104A is the select button rotating to the next item while it is pressed. Which then required to go into the menu again and adjust the setting. On top of that the GW instek also has clever things like a default 1:10 button in the probe setup menu. Also don't be fooled by Keysight's memory depth specs; in most use cases it is only 1/4th of what it says on the badge. So it is 10Mpts (GDS2204E) versus 2Mpts (DSO7104A). A specific issue of the DSO7104A is that it has a noisy front-end (AFAIK modern Keysight scopes should be much better in that respect) so I usually had the high-res mode on as a noise filter which comes with it's own set of problems when displaying a signal. Another problem is lack of processing power in the DSO7104A. 128kpts FFT on the DSO7104A is basically useless due to the excruciatingly slow update rate where the GDS2204E has a reasonable update rate even with 1Mpts FFT on. The same goes for math where the GDS2204E uses actual data instead of decimated data. The latter is also why I'm keeping the GDS2204E around: filtering and math work only well on the actual data and not on decimated data. This is also a point where the RTM3004 falls short.

[kcbrown]

Where the Instek wins is the menu structure. Like the Tektronix TDS500/600/700 series it has a vertical and horizontal row of buttons next to the screen. The horizontal buttons select the menu and the vertical buttons shows the sub-menu with the current settings. This gives the user a very clear overview of what settings are where.

I agree.  Once you get used to that, it makes submenu navigation fast and obvious.

Also the GW Instek has a seperate select button. What drove me mad on the DSO7104A is the select button rotating to the next item while it is pressed. Which then required to go into the menu again and adjust the setting.

Yeah, that's a major usability win for the Instek.  I've never understood why manufacturers use knobs without detents for the select knob/button.  I'm particularly surprised that Keysight would make such a mistake.

That said, I think Instek still made a mistake by not using a detent-based knob for the select knob.  Selecting items from a list is annoying with it because it's easy to overshoot the item you want, so you have to backtrack.  If it had detents it would be a lot easier to be precise with it.

[Sighound36]

Is this still going on really?

My goto scope is 6000B Lecroy, if I didn't require the accuracy and the toolbox, plenty of the suggestions in the previous 5 pages would more than fit the bill without question easily and the cost of the probes would be a darn site cheaper!

Howabout a good everyday scope with decent tool box and a quality power analyser, DMM and Function Genny, load etc and you will have change and have a far more capable set up the just putting all your eggs in one basket.

In the last few years I ahve looked at a good number of mid range scopes up to £100K with tools and probes, some better than others, some far more flexible and accurate some bloody noisy!

If you wish to travel this route you will need exeperiance of using these more powerful and complex devices have indosyncraticness which can be frustrating.

But genuinelly a dailydriver will be a simple to use, quick boot up times and intutive plus you have to wish to use it regularly rather than Oh do I have to switch this thing on?

Take your time, have plenty of long term demo's at home and set up the sorts of testing you would be performing with these types of scopes all of the big four will wish to make a sale and sould be very happy to supply a  demo unit, I have no issues here on the UK othher that breakdowns

[KE5FX]

Interesting when you said getting a scope with larger mem depth was like winning the lottery so why you would not include a more capable capture length as a high priority specification and just one part of choosing a DSO ? 

$100M would be nice, but not if I have to play Squid Game with my oscilloscope to win it.  I'll take $10M and a Megazoom ASIC, thanks.

10K points isn't enough for some of my use cases, but 1M to 10M is fine.

Yeah, that's a major usability win for the Instek.  I've never understood why manufacturers use knobs without detents for the select knob/button.  I'm particularly surprised that Keysight would make such a mistake.

Apparently that was one difference with the DSO/MSO6000 series versus the 7000s.  I've only used the former.  It uses a rotary knob with no detent for menu selection, but there's no 'push to accept selection' feature.  Instead, the function is selected immediately while scrolling through the options.  No detent is necessary, since it's easy enough to back up if you overshoot.

The drawback is that there's no way to close the menu; you have to wait a few seconds for it to close itself, assuming you don't need to open another menu right away.  I assume that's why they added the ability to push the knob.  I can see how doing that without adding a detent was a mistake. 

In the absence of a good touchscreen, the obvious right way to do it would be a capacitive sensor in the knob that closes the menu as soon as you take your hand away.

[tautech]

Interesting when you said getting a scope with larger mem depth was like winning the lottery so why you would not include a more capable capture length as a high priority specification and just one part of choosing a DSO ? 

$100M would be nice, but not if I have to play Squid Game with my oscilloscope to win it.  I'll take $10M and a Megazoom ASIC, thanks.

10K points isn't enough for some of my use cases, but 1M to 10M is fine.

You have that arse about face when for Megazoom you need cough up the $ $ $.
The #1 recommended DSO in this thread uses 2 ADC's each with 200 Mpts of memory support !
However if the thought of such memory depth scares you, you can wind it back to factory default 20 Mpts or lower if that would make you more comfortable.