Keysight DSO4024 vs Tektronix MDO3 - 200 MHz bandwidth vs Teledyne lecroy 3024Z

[whiskeyjack]

My organization is getting into LED driver development business and we lack certain measurement and test equipment. One such equipment is a decent oscilloscope. Currently, we are considering Keysight DSO4024A, Tektronix MDO3 - 200 MHz bandwidth and Teledyne lecroy wavesurfer 3024Z. These are all 4 channel 200 MHz oscilloscopes. Keysight and Tek are giving some software enabled power measurement features which might come in handy. These features are not free.

I am doing some online research to form an opinion which one to go for. I have seen the demo of Tek MDO3 but still waiting for the other two vendors to arrange a demo. Price wise, these are comparable. Keysight being the cheapest and rest two approximately 20 % higher than keysight when looking at scope + probe bundle price. I am also assuming that Tek and Teledyne will give similar discounts as what keysight is offering.

Can you guys help me decide which one would be a better pick from technical and usage perspective? Since these tools are used on a frequent basis, I would like to have a tool which is very intuitive and pleasant to use. At the same time, I don't want to compromise R&D quality by buying a tool which is convenient to use but gives wrong measurement.

During my research, I came across a video from Tek -

This video compares keysight and tek and mentions the fact that keysight uses displayed data for math and measurement functions. Should I be concerned about this?

[KE5FX]

I predict this thread will remain calm and free of biased personal opinions, hurt feelings, and moderator intervention. 

That being said, don't buy the Tek without evaluating the Keysight in person.   One of the two will drive like a dream.  The other one may make you want to throw it off the top level of the company's parking garage, a la Bob Widlar.

I don't know much about LeCroy, but at least one frequent poster can speak authoritatively on them.

This video compares keysight and tek and mentions the fact that keysight uses displayed data for math and measurement functions. Should I be concerned about this?

Here's the "opinion" part: you don't want your measurements to come from data that doesn't correspond to what's currently on the screen.  That's a fairly fundamental aspect of what makes an oscilloscope an oscilloscope.  By their nature, it's easy to get bad data from an oscilloscope, and measurements that reflect the entire acquisition record make it really easy for that to happen.

That doesn't mean that the data being measured should come from the actual screenspace output buffer, though.  While measurement data should come from the part of the acquisition record that corresponds to what's on the screen, it ideally shouldn't have undergone downsampling, quantization, and clipping to conform to the LCD's resolution.  Keysight's architecture is good in this respect, since a lot of their measurements are performed on an ASIC that has direct access to acquisition memory, rather than on a CPU that doesn't.

The downside is that their ASIC imposes record size limits of its own.  So if huge records are important to you, that will be something you have to consider.  (Are they?  Are you sure?)

[RoadRunner]

From where is see I would not touch Tektronix with a stick. MDO3000 is unuseably slow. Still have lots of bugs there is no active devlepment. I had gone through almost a year of sending and receiving emails.

I would not buy any Tek scope in future as well because if you look at there new scopes they do not seems to make any progress internally on orther hand cosmetics they look nice, with nice big display and stuff.

Tektronix support at liest in DE I contacted was pertty unsupportive and Blame everthing on customer type, Specially when customer is an individual rather than a company.

I have not used any high end keysight scope ever  but i have used highend lecroy at work place so i would incline towards lecroy for ease of usablity. With lecroy price may be more though.

[Pinkus]

I own a DSOX4024a and worked with a MDO3 for a few days. Indeed the Tektronix is slow as hell and I would not work with it, even if I got one for free. Though, the MDO4 might be better - I have no experience with it.
I can highly recommend of getting a demo unit of the Keysight for a week. You will love it! It is so intuitive and responsive, working with it is always a sheer delight.
You might save some bucks by purchasing a DSOX3000 T instead. It is almost the same unit and software. Basically what is different are: screen size, 1 vs. 2 channel AWG, optional/integrated VGA/LAN, input of external 10 Mhz clock reference possible with 4000 series.

[nctnico]

Another forum member is not fond of the Lecroy Wavesurfer 3000 models due to unfixed bugs. Also Lecroy oscilloscopes typically don't have peak-detect mode which I consider a must-have. I'd consider the R&S RTB2004 and RTM3004 as well. I think these have a lot of pros compared to the other models. Faster to operate compared to Tektronix and more memory compared to Keysight.

You can have long debates about using acquisition memory versus screen memory for math. The first is always accurate but very slow. The latter is faster BUT it will typically only be accurate if you can see the waveforms on screen. If you zoom out to much where the waveform become a unform band you'll see that the calculations will fail. The Keysight oscilloscopes will indicate this. Another thing to look out for it whether an oscilloscope adheres to the memory depth setting or just captures enough data to fill the screen (Lecroy and modern Keysight scopes seem to only capture enough data to fill the screen). In the latter case you'll need to go back & forth between time/div settings in order to make long captures. There are several usage scenarios where this becomes really tedious. It is very nice to have more data on what happened before an event to see what happened leading up to it even if that data is off-screen at the time of capturing. It is not a given that an event is easy to reproduce so the more data you have the less time it will take to track down a intermittant bug.

[tautech]

Teledyne lecroy wavesurfer 3024Z.

Is a SDS3024X made by Siglent whom have a long standing arrangement with LeCroy whom have the western world marketing rights for them.
Siglent's new SDS5000X series supersedes the 3000Z series with higher sampling rates, more memory and individual vertical controls just to name the obvious. There's a thread here on the SDS5054X if you want to check it out.
Dave's done a couple of videos on the SDS5104X that he tore down and ran against similar class DSO's.

However whiskeyjack, the lowest BW offering in the 5000X series is 350 MHz and the pricing might exceed the other models you're looking at.

[Wuerstchenhund]

Don't go for the Tek. Seriously, don't. Tek is pretty much the bottom of the barrel when it comes to A-brands. We got all their models for evaluation and every one of them had serious issues. You pretty much pay for a name that meant something back when analog scopes were still a thing.

The Keysight DSOX4024 is a good everyday scope, we have a bunch of the higher BW models. As KE5FX says they are easy to use, but because of their architecture (which was designed for max waveform rates at all costs) they have some limitations, like the small memory (which in reality, will end up even smaller because of how the MegaZoom ASIC uses it).

Keysight support is top notch, which was necessary as we had several which suffered from excessive noise and the NAND failure. Also, Keysight artificially limits the probe interface of the InfiniVision Series scopes (of which the DSO-X4k is a part of) compared to the Infiniium Series (i.e. Windows based high end scopes) which makes some probes incompatible.

The LeCroy WaveSurfer 3024z is the successor of the WaveSurfer 3000 (non-z), both which are manufactured by Siglent (the software is designed by LeCroy). They are good scopes, although a bit slow at times. They're not as zippy as the Keysight, but they have more memory and offer more math and analysis tools (and larger FFTs). In addition, they share the probe interface (ProBus) with other LeCroy scopes, which gives them access to a wide range of probes. Like all newer LeCroy scopes they lack Peak Detect, and without going into another discussion of the value of PD (a feature designed to compensate for the shortcomings of early DSOs which had tiny memory) with modern deep memory DSOs, it's down to you to decide if this feature is critical for you or not (I've last used it some 20 years or so ago, same for my colleagues).

I know some member from Europe had some complaints about a WS3000 but I'm not sure if this is down to the specific unit (we have many of the higher BW variants and little complaint). Unfortunately LeCroy support in Europe has taken a nose dive since it was moved to Germany (U.S. support is still great, though) so this might have played a role in the complaints.

My biggest complaint with the WS3000z is that LeCroy has used the success of the modestly priced WS3000 to squeeze more cash out of the WS3000z. Still, they are good scopes for what you get for the money.

As others said, other potential candidates could be the R&S RTB2004, and there's a new Siglent scope (SDS2kX+) as well as the Rigol MSO5000. None of them have active probe interfaces like the Keysight and LeCroy do, and really are of a different (lower) class than those two.

The RTB2004 is a 10bit scope, however it's limited in functionality, it has some memory limitations (the 160M memory is only available in segmented mode, in all other modes you're limited to 10M/20M) and even lacks 50 ohms inputs. Options are truly expensive unless you're lucky enough to get one of the rare bundle offers.

The Siglent SDS2kX+ is pretty new, has 200M of memory, but it appears to have a lot of capabilities for its class, offers a (software) 10bit mode and I believe has switchable 1M/50ohms inputs.

All came with a range of bugs, and I leave it to others to comment on what the current maturity state is of these scopes.

Take your pick.

In any case, I'd strongly recommend to get loaner units to try for yourself, and not buy based on reviews and opinions only.

[HighVoltage]

I got rid of all Tektronix scopes a few years ago, that is how bad they are compared to others.
And I do not think that I will ever have a Tek scope again.

Now I use Keysight 2000, 3000 and 6000 Series scopes and can highly recomend them.
Like others said, get a loaner for a week and play around with it.

[dreamcat4]

Since these tools are used on a frequent basis, I would like to have a tool which is very intuitive and pleasant to use.

without question the keysight

keysight uses displayed data for math and measurement functions. Should I be concerned about this?

most oscilloscopes do this so that limitation is not specific to only keysight. getting an honest picture requires learning these limitations behaviour across different modes etc and understanding the limits of that accuracy. if you need more performance than what the 200mhz model can provide at its limit then getting a better scope will probably cost significantly more. it you are not certain then ask so sample multiple keysight scopes.

other than keysight maybe a lecroy because its software user interface. but it would seem that there are more question marks over lecroy than keysight. who else to consider?

[tautech]

The Siglent SDS2kX+ is pretty new, has 200M of memory, but it appears to have a lot of capabilities for its class, offers a (software) 10bit mode and I believe has switchable 1M/50ohms inputs.

Of course it does as should any scope in this class.
https://www.siglenteu.com/wp-content/uploads/dlm_uploads/2019/12/SDS2000X-Plus_Datasheet_DS0102XP_E01A-1.pdf

[jake111]

Make sure to consider probes during your purchase.

I am back to school now but worked for a semiconductor company in the power electronics group.  None of the senior guys would use Agilent due to probe selection.  They all insisted on Tektronix DPO/MSO4000 series with those TDP series probes for their work.  Amazing probes and I will admit the scopes were wonderful though out of the price range for someone comparing it to that cost reduced DPO/MDO3000 series or anything Agilent.  The ability to capture at maximum sample depth (I recall 20Mpts per channel) at any horizontal setting on the Tek was extremely useful for some things (i.e. you can capture and then literally zoom to infinity and beyond with amazing clarity) and my Agilent DSO-X 3000 does not seem to support this at all which sucks.  I see in the Agilent documentation that they seem to put down the Tek behavior by claiming that this huge capture depth slows the scope down, and therefore update rate down, which it certainly does, however you have complete control over it at least on the Tek 4000 series and when I worked there the guy who always had the Eureka! moment solving a huge ongoing problem was always using a Tek if it was power related.  For high speed digital stuff, I saw lots of Agilent and they may dominate here, I really don't know but if you are doing constant current drivers for LED consider what probes you might want and take that into account.  As an example Tek TDP1000 probes look to be under $1500 used and for a probe with 42V common mode range and 100V damage threshold, yet 1GHz BW and such small parasitics, I don't know what Agilent has to compete but I sure can't find it on eBay for a good price if it's there.

[tautech]

The ability to capture at maximum sample depth (I recall 20Mpts per channel) at any horizontal setting on the Tek was extremely useful for some things (i.e. you can capture and then literally zoom to infinity and beyond with amazing clarity) and my Agilent DSO-X 3000 does not seem to support this at all which sucks.  I see in the Agilent documentation that they seem to put down the Tek behavior by claiming that this huge capture depth slows the scope down, and therefore update rate down, which it certainly does, however you have complete control over it at least on the Tek 4000 series and when I worked there the guy who always had the Eureka! moment solving a huge ongoing problem was always using a Tek if it was power related.

There's more than one way to skin a cat !

1) Pick the timebase setting that provides the largest mem depth or enough to suit your needs.
Place the scope in Zoom (split screen mode) and have ALL the scopes memory depth to work with using the Horizontal pan be it Run mode or for a capture.
2) Single trigger on an event then use History which for some scopes provides even greater memory than 'official' memory depth.
3) Segmented acquisition (Sequence) mode which divides the maximum record length into multiple segments WRT to trigger conditions.

[nctnico]

Aaarghh the Siglent marketing campaign again  No, history and segmented recording are not an alternative. Jake111 is right: capturing data beyond the screen is a feature which is a major productivity boost. It is a real shame Keysight has dropped this from their more recent products. Personally I wouldn't buy a general purpose oscilloscope which can't sample beyond the screen.

[2N3055]

Aaarghh the Siglent marketing campaign again  No, history and segmented recording are not an alternative. Jake111 is right: capturing data beyond the screen is a feature which is a major productivity boost. It is a real shame Keysight has dropped this from their more recent products. Personally I wouldn't buy a general purpose oscilloscope which can't sample beyond the screen.

You are correct when saying segmented memory is not good for that job.

But Jake111 is not saying ANYTHING about capturing data outside the screen. That's your thing. Actually capturing at long time-base and using your hated zoom function to look up the data is explicitly mentioned. On that Tek it is only method to accomplish this.

He merely noted that MSO 4000 from Tek  has long memory (20/40 MSPS as opposed to 1/2/4 MSPS on Keysight) so you can capture many milliseconds worth of data at full sample rate (and loosing no data due to antialising). For that job is much better than Keysight 3000/4000/6000.

In which case new Siglents (2000X+ and 5000X) with 100/200 MSPS and new Rigols with 100/200/400 MSPS, and Picoscope (500MSPS/1GSMPS) are even better  at that job, having another order of magnitude more sample memory.

Regards,

[nctnico]

Aaarghh the Siglent marketing campaign again  No, history and segmented recording are not an alternative. Jake111 is right: capturing data beyond the screen is a feature which is a major productivity boost. It is a real shame Keysight has dropped this from their more recent products. Personally I wouldn't buy a general purpose oscilloscope which can't sample beyond the screen.

You are correct when saying segmented memory is not good for that job.

But Jake111 is not saying ANYTHING about capturing data outside the screen. That's your thing. Actually capturing at long time-base and using your hated zoom function to look up the data is explicitly mentioned. On that Tek it is only method to accomplish this.

Zoom can also mean using the time/div knob and the way Jake111 describes it the Tektronix is capturing beyond the screen:

...capture at maximum sample depth (I recall 20Mpts per channel) at any horizontal setting....

I have some hands-on experience with the Tektronix TBS2000 series and this does capture beyond the screen like a good DSO should.

[2N3055]

Aaarghh the Siglent marketing campaign again  No, history and segmented recording are not an alternative. Jake111 is right: capturing data beyond the screen is a feature which is a major productivity boost. It is a real shame Keysight has dropped this from their more recent products. Personally I wouldn't buy a general purpose oscilloscope which can't sample beyond the screen.

You are correct when saying segmented memory is not good for that job.

But Jake111 is not saying ANYTHING about capturing data outside the screen. That's your thing. Actually capturing at long time-base and using your hated zoom function to look up the data is explicitly mentioned. On that Tek it is only method to accomplish this.

Zoom can also mean using the time/div knob and the way Jake111 describes it the Tektronix is capturing beyond the screen:

...capture at maximum sample depth (I recall 20Mpts per channel) at any horizontal setting....

I have some hands-on experience with the Tektronix TBS2000 series and this does capture beyond the screen like a good DSO should.

I don't know about TBS2000. But zoom means zoom in Tek speak, and is widely used and taught in tek literature.
It is Tek recomended way of doing this, and frankly, most logical one to most people excluding you.
Other fact is that Tex did make it very nicely (inteligently), so it uses least amount of screen space..

[nctnico]

But jake111 isn't a Tektronix manual. Just a user. There is no need for a semantic discussion here. And using zoom mode to get most of the memory may be logical but it simply isn't the most productive way. You still need to do more steps to make a measurement. Just let the user drive the record length.

[2N3055]

But jake111 isn't a Tektronix manual. Just a user. There is no need for a semantic discussion here. And using zoom mode to get most of the memory may be logical but it simply isn't the most productive way. You still need to do more steps to make a measurement. Just let the user drive the record length.

I wholeheartedly disagree. Simplest way is to simply use time base to set length of time to be captured (like you would on analog scope), capture the single event, and then zoom in to area of interest to see details.

I usually want to capture some known time interval (exact or approximate), which DIRECTLY corelates with time base. Period

Doing it your way, I need to calculate in my head what length at which sample rate I need and at which time/div.

My way is much simpler. Except for screen real estate, it is simpler and more logical to do. And Tek has really good implementation on MDO4000 series, for instance, with minimum screen wasted, and even a special shuttle buttons for really nice navigation trough zoom.

And also, you're confusing capability with comfort. On a 200 MSPS Siglent in zoom mode, you CAN capture 40ms at full 5GS/s.
On very expensive MSOX-3104T you can capture 800us worth of data at full 5GS/s, zoom or no zoom mode... You CANNOT capture 40ms however you use it...

[nctnico]

But jake111 isn't a Tektronix manual. Just a user. There is no need for a semantic discussion here. And using zoom mode to get most of the memory may be logical but it simply isn't the most productive way. You still need to do more steps to make a measurement. Just let the user drive the record length.

I wholeheartedly disagree. Simplest way is to simply use time base to set length of time to be captured (like you would on analog scope), capture the single event, and then zoom in to area of interest to see details.

I usually want to capture some known time interval (exact or approximate), which DIRECTLY corelates with time base. Period

Doing it your way, I need to calculate in my head what length at which sample rate I need and at which time/div.

You misunderstand. It is the other way around. What I do every now and then is look at the part of an I2C or SPI message to catch a specific piece of information. In some cases getting a measurement does take some time to wait so doing many captures gets expensive quickly. With the memory depth set to full (10MPts, 20Mpts, 80Mpts, whatever) this means I automatically capture a lot of information outside the screen. If I want to see a more of the signal the only thing I need to do is turn the time/div knob to zoom out. It doesn't get any simpler. No need to think about what precise setting to use. Forgot to put the trigger point back to the center? No problem, the info is there. Forgot to set the time/div exactly right? No problem, the info is there. You can work very lazy this way and not care so much about the exact settings. Just capture lots of data and analyse afterwards. It gets especially rewarding if something unexpected happens. Often the data leading up to the random event is also there.

[jake111]

Sorry I triggered you guys (HA GET IT, IT'S A SCOPE JOKE!) but I really think the capture depth is very important and this is why the senior guys I worked with used them.  That and the probes, Tek's active probes seem to be second to none.  Segmented memory and all that is great when you are looking at a repetitive waveform and want to observe some trend over time but when you have one shot to observe a failure, for instance in a case where you get catastrophic IC failure under a specific condition and want to capture everything in great detail for analysis later in that one shot, the Agilent (at least my DSO-X 3000) seems kinda useless for that.  I assume this was a cost reduction feature and this is why the Tek 4000 series still cost 10 grand more than 10 years later while the Agilent can be had for two thousand.   If you don't need this deep capture depth because all you look at are repetitive waveforms and never care to capture something in one shot in great detail then the cheaper scope is probably better for you unless you need good active probes.  Don't discount the value of the probes.  Those guys had Tek active probes that were approaching a decade old, i.e. the Tek TDP and TCP series and those probes still passed cal every year and to be able to observe a medium voltage switching power waveform at those voltage levels without damage and also while exposing the DUT to such miniscule parasitics from the probe front end was pretty damn cool.

Anyway I have an Agilent DSO-X because I don't have the money for something better, for the price I paid I am not complaining at all.  I have used the Rigol MSO5000 series and though those seem to pack a lot of claimed functionality in a small price, I would take the Agilent over it for the small price increase any day.  You get what you pay for.

And in all of this I never mentioned the LeCroy because I've never seen one before other than their SPARQ VNA and their USB 3.0 test/cert setup which was more of a holy grail than a scope...  I know nothing about these...

[jake111]

Sorry I triggered you guys (HA GET IT, IT'S A SCOPE JOKE!) but I really think the capture depth is very important and this is why the senior guys I worked with used them.  That and the probes, Tek's active probes seem to be second to none.  Segmented memory and all that is great when you are looking at a repetitive waveform and want to observe some trend over time but when you have one shot to observe a failure, for instance in a case where you get catastrophic IC failure under a specific condition and want to capture everything in great detail for analysis later in that one shot, the Agilent (at least my DSO-X 3000) seems kinda useless for that.  I assume this was a cost reduction feature and this is why the Tek 4000 series still cost 10 grand more than 10 years later while the Agilent can be had for two thousand.   If you don't need this deep capture depth because all you look at are repetitive waveforms and never care to capture something in one shot in great detail then the cheaper scope is probably better for you unless you need good active probes.  Don't discount the value of the probes.  Those guys had Tek active probes that were approaching a decade old, i.e. the Tek TDP and TCP series and those probes still passed cal every year and to be able to observe a medium voltage switching power waveform at those voltage levels without damage and also while exposing the DUT to such miniscule parasitics from the probe front end was pretty damn cool.

Anyway I have an Agilent DSO-X because I don't have the money for something better, for the price I paid I am not complaining at all.  I have used the Rigol MSO5000 series and though those seem to pack a lot of claimed functionality in a small price, I would take the Agilent over it for the small price increase any day.  You get what you pay for.

And in all of this I never mentioned the LeCroy because I've never seen one before other than their SPARQ VNA and their USB 3.0 test/cert setup which was more of a holy grail than a scope...  I know nothing about these...

Thanks this is kind of what I was trying to point out.  If you are looking at repetitive waveforms with periodically repeating "glitches", sure you can get away with the poor capture depth of the Agilent.  However if you are looking at a "once in a blue moon" failure that you spent 30 minutes carefully setting up triggering for, or a catastrophic failure that causes damage when it occurs, you only get one shot, and I think these two use cases encompass a lot of professional use.  The repetitive stuff is for more run of the mill problems.  The statement that you can just change your horizontal settings, or set up this bandaid segmented memory to capture something with more clarity just gives me the though "Oh, well that sounds nice, doesn't it... But in the real world..."... This waveform update rate thing is another annoyance - Yes Agilent that's great, you can update more waveforms per second and say "hey, that repeating glitch happened" - But what if it only happens once? Can I zoom in on that glitch with great clarity without reacquiring?  And for lots of professional use, that really, really sucks.

To each their own but if you are trying to bash Tek's deep capture depth and say that Agilent's segmented memory is an acceptable replacement you're either regretting your purchase or haven't had to do any troubleshooting of something that is a rare or single failure.

[2N3055]

But jake111 isn't a Tektronix manual. Just a user. There is no need for a semantic discussion here. And using zoom mode to get most of the memory may be logical but it simply isn't the most productive way. You still need to do more steps to make a measurement. Just let the user drive the record length.

I wholeheartedly disagree. Simplest way is to simply use time base to set length of time to be captured (like you would on analog scope), capture the single event, and then zoom in to area of interest to see details.

I usually want to capture some known time interval (exact or approximate), which DIRECTLY corelates with time base. Period

Doing it your way, I need to calculate in my head what length at which sample rate I need and at which time/div.

You misunderstand. It is the other way around. What I do every now and then is look at the part of an I2C or SPI message to catch a specific piece of information. In some cases getting a measurement does take some time to wait. With the memory depth set to full (10MPts, 20Mpts, 80Mpts, whatever) this means I automatically capture a lot of information outside the screen. If I want to see a more of the signal the only thing I need to do is turn the time/div knob to zoom out. It doesn't get any simpler. No need to think about what precise setting to use. Forgot to put the trigger point back to the center? No problem, the info is there. Forgot to set the time/div exactly right? No problem, the info is there. You can work very lazy this way and not care so much about the exact settings. Just capture lots of data and analyse afterwards. It gets especially rewarding if something unexpected happens. Often the data leading up to the random event is also there.

Well my way is exactly the same, except i turn time base in other direction .... 

Don't get me wrong, I respect you have way that works for you. I have a different way that works for me... They both do the job.

But you are actually making important point. On my Keysight, I can capture at slow timebase, and then I can also change timebase and position, and traverse whole captured buffer (your way). Without using zoom. On little Rigol DS1074Z same thing.

Can you do it on Siglent? Or new Rigols? Or does change of timebase erases the existing buffer? Can you change timebase to something shorter and traverse magnified portion on screen (as a sort of full screen zoom) ?

[2N3055]

Sorry I triggered you guys (HA GET IT, IT'S A SCOPE JOKE!) but I really think the capture depth is very important and this is why the senior guys I worked with used them.  That and the probes, Tek's active probes seem to be second to none.  Segmented memory and all that is great when you are looking at a repetitive waveform and want to observe some trend over time but when you have one shot to observe a failure, for instance in a case where you get catastrophic IC failure under a specific condition and want to capture everything in great detail for analysis later in that one shot, the Agilent (at least my DSO-X 3000) seems kinda useless for that.  I assume this was a cost reduction feature and this is why the Tek 4000 series still cost 10 grand more than 10 years later while the Agilent can be had for two thousand.   If you don't need this deep capture depth because all you look at are repetitive waveforms and never care to capture something in one shot in great detail then the cheaper scope is probably better for you unless you need good active probes.  Don't discount the value of the probes.  Those guys had Tek active probes that were approaching a decade old, i.e. the Tek TDP and TCP series and those probes still passed cal every year and to be able to observe a medium voltage switching power waveform at those voltage levels without damage and also while exposing the DUT to such miniscule parasitics from the probe front end was pretty damn cool.

Anyway I have an Agilent DSO-X because I don't have the money for something better, for the price I paid I am not complaining at all.  I have used the Rigol MSO5000 series and though those seem to pack a lot of claimed functionality in a small price, I would take the Agilent over it for the small price increase any day.  You get what you pay for.

And in all of this I never mentioned the LeCroy because I've never seen one before other than their SPARQ VNA and their USB 3.0 test/cert setup which was more of a holy grail than a scope...  I know nothing about these...

Thanks this is kind of what I was trying to point out.  If you are looking at repetitive waveforms with periodically repeating "glitches", sure you can get away with the poor capture depth of the Agilent.  However if you are looking at a "once in a blue moon" failure that you spent 30 minutes carefully setting up triggering for, or a catastrophic failure that causes damage when it occurs, you only get one shot, and I think these two use cases encompass a lot of professional use.  The repetitive stuff is for more run of the mill problems.  The statement that you can just change your horizontal settings, or set up this bandaid segmented memory to capture something with more clarity just gives me the though "Oh, well that sounds nice, doesn't it... But in the real world..."... This waveform update rate thing is another annoyance - Yes Agilent that's great, you can update more waveforms per second and say "hey, that repeating glitch happened" - But what if it only happens once? Can I zoom in on that glitch with great clarity without reacquiring?  And for lots of professional use, that really, really sucks.

To each their own but if you are trying to bash Tek's deep capture depth and say that Agilent's segmented memory is an acceptable replacement you're either regretting your purchase or haven't had to do any troubleshooting of something that is a rare or single failure.

It's okay that is happening all the time here. 

Seriously, there is no replacement for deep acquisition buffer when you need it. Period.
I actually have a combination of Keysight MSOX-3000T and a Picoscope with 500 Msamples buffer, because Keysight sucks at longer captures.
So, tool for the job...

But Tektronix is not only one with more memory than Keysight. There is number of R&S, LeCroy, Siglent, Rigol, Picoscope(USB scopes) etc..
And most of them are working better than Tek at most things.

But.... If you have bunch of active probes for Tek ecosystem, then Tek might be the choice...

Tool for the job.

[jake111]

I just have to say, I wouldn't consider anything that erases the captured memory when you change the timebase to be an oscilloscope... I would consider that to be a poor chinese copy, you have to understand these 9 year old kids are working 16 hour days to get this product to you and they can't do everything, give them a break...   

[nctnico]

But Tektronix is not only one with more memory than Keysight. There is number of R&S, LeCroy, Siglent, Rigol, Picoscope(USB scopes) etc..
And most of them are working better than Tek at most things.

But.... If you have bunch of active probes for Tek ecosystem, then Tek might be the choice...

I have to agree. Going by the experience of others on this forum the Tek MDO3000 / MDO4000 scopes are not that good when it comes to actually using the deep memory for math and decoding. Makes me sad though because Tektronix has made so much nice equipment in the past. I really liked the elegant designs (the electronics) of the Tek 2230 and TDS500/TDS700 series oscilloscopes.