Are 'premium' scope brands still justified?

[tautech]

Although I have to say, after a little thought, I would have to cancel the 2000Xplus again, a clear improvement on the 800X HD would in fact only be the 2000X HD.
There is nothing in between, not even the 1000X HD.
Main reason:
Guess what...

1. Low noise
2. Full memory management

[Martin72]

1. ERES not only as a mathfunction
2. Low noise
3. Full memory management
 

[Electro Fan]

Although I have to say, after a little thought, I would have to cancel the 2000Xplus again, a clear improvement on the 800X HD would in fact only be the 2000X HD.
There is nothing in between, not even the 1000X HD.
Main reason:
Guess what...

1. Low noise
2. Full memory management

From the spec sheets:
2000X HD Jitter CH1~CH4  <10 ps rms (typical), Ext <200 ps rms
1000X HD Jitter CH1~CH4 <100 ps, Ext <200 ps rms
  800X HD Jitter CH1~CH4 <100 ps, Ext na

https://siglentna.com/wp-content/uploads/dlm_uploads/2023/03/SDS2000X-HD_Datasheet_EN01A.pdf
https://siglentna.com/wp-content/uploads/dlm_uploads/2024/03/SDS1000X-HD_Datasheet_EN01F.pdf
https://siglentna.com/wp-content/uploads/dlm_uploads/2024/03/SDS800X-HD_Datasheet_EN01C.pdf

[Performa01]

From the spec sheets:
2000X HD Jitter CH1~CH4  <10 ps rms (typical), Ext <200 ps rms
1000X HD Jitter CH1~CH4 <100 ps, Ext <200 ps rms
  800X HD Jitter CH1~CH4 <100 ps, Ext na

https://siglentna.com/wp-content/uploads/dlm_uploads/2023/03/SDS2000X-HD_Datasheet_EN01A.pdf
https://siglentna.com/wp-content/uploads/dlm_uploads/2024/03/SDS1000X-HD_Datasheet_EN01F.pdf
https://siglentna.com/wp-content/uploads/dlm_uploads/2024/03/SDS800X-HD_Datasheet_EN01C.pdf

As explained several times already, the low-end device jitter specs are just some house numbers that can never be violated. By contrast, e.g. the SDS2000X HD provides true specifications for that property - and very conservative ones, as the actual jitter is still some five times better.

In actual fact, we can derive the trigger jitter from the peak detect specification, since this indicates the sample rate in the trigger path. As this is 2 ns on the SDS800X HD vs. 1 ns on the SDS2000X HD, we can expect the trigger jitter to be twice as high. Consequently, a proper specification would read <20 ps RMS (typical).

Actually, my measurement was a bit higher (almost 5 ps), but that's because the lower bandwidth of the 800 series forced me to use a test signal with lower slew rate.

Here is my measurement for the SDS824X HD in Reply #2, Section "Trigger Jitter":
https://www.eevblog.com/forum/testgear/sds800x-hd-review-demonstration-thread/msg5293747/#msg5293747

[Keysight DanielBogdanoff]

Or have the 'big boys' been caught napping, and are now about to find you can't charge £10k+ for obsolete hardware any more?

We're definitely not napping, and naturally we keep an eye on this part of the scopes market. There's a lot of technology "trickle down" that happens with our chip architectures (both processors and ADC/front ends). So, it's common for us to push the envelope with a super high end scope (like a UXR) and then leverage those learnings into the next chipsets.

On the chip front, you'll also see the "big boys" doing their own ICs but what's described in this thread as "B brands" are typically using off-the-shelf components and are stuck at that level of performance.

Ultimately I wouldn't call them "B brands" but personally think of them more as "value" brands. They will always be the cheapest for a given bandwidth, even if "A brands" dropped prices the "B brands" would undercut. But, if you buy one that's what you get. I could get an Ender 3 3d printer for $100, but there's a reason I paid 10x for my Bambu.

There is also what I think of as "market trickle up" - for example what options/capabilities are included as "standard." Back at release of this generation of Agilent/Keysight scopes you had to pay for options that are now standard, mostly due to market pressure of including certain decodes (for example) for free.

Personally, as someone who uses/has used basically all the scopes, using some brand's scopes annoys me...

[2N3055]

Personally, as someone who uses/has used basically all the scopes, using some brand's scopes annoys me...

Yep, I will sign that sentence !!

Good to hear from you!!

[AndyC_772]

Hi Daniel, thanks for stopping by, it's interesting to hear a manufacturer's viewpoint - and as you can hopefully tell, I'm still a big fan of my MSO-X3000A scopes. Until fairly recently, wouldn't have considered trading them for anything else. I've also recommended them to others who have gone on to purchase their own too, and as far as I know, they're very happy too.

However... I'm not sure I'd agree with your comment that it's the 'B' brands who are "stuck at that level of performance".

I've no doubt that when the current MegaZoom ASIC was developed, it was a state-of-the-art, fit for purpose solution to the problem of how to build a solid, responsive and reliable everyday bench scope. But that was really rather a long time ago now, and it's entirely reasonable to suggest that what may have needed an ASIC back then might be equally well done in an FPGA today.

Moreover, an ASIC is inherenltly a snapshot in time, lots of NRE and no real ability to incrementally improve it year on year, whereas with off-the-shelf components there's always the option to move to this year's hot new device, and provided it's still programmed in Verilog, VHDL, C, or whatever, that migration is much less costly.

I'd expect scopes made that way to get meaningfully more capable year on year, while those built on ASICs don't - and that's exactly what seems to have happened. If the 3000G series really is, as I've heard it described, "just the 3000T with APPBUNDL included", that's just a lick of paint and a tweak to the pricing structure - not really progress as such.

That pricing does warrant highlighting in the context of today's market. Whether the cost is in the options or the base price of the scope, 3000G is still priced way higher than, say, the new Siglent 3000X-HD for equivalent bandwidth. £5200 (call it £6k with the digital cable plus a few options for a typical configured unit) vs £17899 for 1GHz puts the Keysight at about 3x the price, for technology that must be 10 yrs old by now.

That, surely, warrants justification, and I think I'm right to ask if a 3000G is still a good choice for a small business.

As it happens, the (unopened) SDS804X-HD that's currently sitting on my desk was also about 1/3 the cost of the various Keysight 1000 series scopes we have dotted around the lab. (Coincidence?)

I don't doubt there's cool new stuff in the pipeline, but as a mere end-user I've not seen it yet. Something like a '3500G' series, based on the 3000G platform but updated with a 12 bit front end, 100 M points / channel, and priced at a level which isn't totally blind to what's been coming out of China the last few years? Sign me up.

[nctnico]

Still, one of the rather unique features of the Keysight oscilloscopes is that these can decode across segments. When recording a bunch of segments, the decode table shows data from all segments (with timestamps ofcourse!) in one so you can hop through the various segments with ease. On other oscilloscopes I've seen/used so far you'll need to go through each segment and look at the decoded data for that specific segment; there is typically not an easy way to aggregate all decoded data.

[2N3055]

Still, one of the rather unique features of the Keysight oscilloscopes is that these can decode across segments. When recording a bunch of segments, the decode table shows data from all segments in one so you can hop through the various segments. On other oscilloscopes I've seen/used so far you'll need to go through each segment and look at the decoded data for that specific segment; there is typically not an easy way to aggregate all decoded data.

That is a good point.
But if you need to do that a lot, Picoscope does that and to level no embedded scope does.

And for the price of MSOX 3104G you can buy several Siglent and Picoscope scopes(or whatever your favourite brand is) at the same time.

[Martin72]

Something like a '3500G' series, based on the 3000G platform but updated with a 12 bit front end, 100 M points / channel, and priced at a level which isn't totally blind to what's been coming out of China the last few years?

If you have dominated the market for decades together with 2...3 others, I can well imagine that you can react rather sluggishly to the cheaper competition from the East.
Or want to.
This wasn't necessary in the early years either, firstly because the quality of the devices didn't even come close to their own products, and secondly because they had a customer base.
But in the meantime, something has grown up that could cause problems for the major suppliers.
The future will show whether they will ignore this and continue to rely on their customer base, or whether they will actually be willing to counter it.
And we're not talking about ambitious hobbyists who want to invest more than €500.
For example, if I can get a Siglent 2000XHD/3000XHD today that is just as good as a scope that costs €10000....15000, then I don't have to think twice - as a company.
I had equipped our department with scopes that together cost as much as ONE Waverunner.
And we won't be the only ones who have done this or are thinking of doing so.
Can you afford to ignore the cheaper competition in the long term?

[Antonio90]

Something like a '3500G' series, based on the 3000G platform but updated with a 12 bit front end, 100 M points / channel, and priced at a level which isn't totally blind to what's been coming out of China the last few years?

Can you afford to ignore the cheaper competition in the long term?

That's a very interesting point. And raises the question as to why the price difference really.
As a company, buying for more than needed is shooting yourself in the foot, but relying fully on foreign production because it is cheaper is setting yourself up to be shot in the head later.

This would be a very good argument, I think, if it wasn't for the A brands moving production to the east too.

[nctnico]

Still, one of the rather unique features of the Keysight oscilloscopes is that these can decode across segments. When recording a bunch of segments, the decode table shows data from all segments in one so you can hop through the various segments. On other oscilloscopes I've seen/used so far you'll need to go through each segment and look at the decoded data for that specific segment; there is typically not an easy way to aggregate all decoded data.

That is a good point.
But if you need to do that a lot, Picoscope does that and to level no embedded scope does.

AFAIK Picoscope has no protocol triggering and when you need to do really long term decodes, then triggering is your first filtering stage to get rid of excess data. Think about debugging a rare (once in an hour or longer) problem with a specific device / access pattern on a I2C or SPI bus.

[nctnico]

Something like a '3500G' series, based on the 3000G platform but updated with a 12 bit front end, 100 M points / channel, and priced at a level which isn't totally blind to what's been coming out of China the last few years?

Can you afford to ignore the cheaper competition in the long term?

That's a very interesting point. And raises the question as to why the price difference really.

You pay for pedigree (as in having decades of knowledge on how to build particular test equipment), proper product/firmware regression testing and long term support. When I ask Keysight a really tough question about the internals of a piece of equipment I bought 10 years ago (in order to assess whether the quality of measurement results can be impacted by a potential failure) I get a good, detailed answer.

The price of test equipment is not in the hardware, it is in the engineering and support. All major test equipment manufacturers are producing in low wage countries.

[2N3055]

Still, one of the rather unique features of the Keysight oscilloscopes is that these can decode across segments. When recording a bunch of segments, the decode table shows data from all segments in one so you can hop through the various segments. On other oscilloscopes I've seen/used so far you'll need to go through each segment and look at the decoded data for that specific segment; there is typically not an easy way to aggregate all decoded data.

That is a good point.
But if you need to do that a lot, Picoscope does that and to level no embedded scope does.

AFAIK Picoscope has no protocol triggering and when you need to do really long term decodes, then triggering is your first filtering stage to get rid of excess data. Think about debugging a rare (once in an hour or longer) problem with a specific device / access pattern on a I2C or SPI bus.

You are correct. So doesn't Saelae.
But Picoscope can capture 10s of thousands of packets and filter them afterwards.

But one wrong packet in 2 hours is perfectly manageable without decoding across segments.
20-30 packets a day are handled with history and manual run through.

Segments + trigger mask are used when you have fast incoming packets, where you want to ignore millions and capture thousands.

[tszaboo]

Something like a '3500G' series, based on the 3000G platform but updated with a 12 bit front end, 100 M points / channel, and priced at a level which isn't totally blind to what's been coming out of China the last few years?

If you have dominated the market for decades together with 2...3 others, I can well imagine that you can react rather sluggishly to the cheaper competition from the East.
Or want to.

Honestly, why would they care about the low end market? They can sell an active probe with more profit on it than a DSOX1000 or 2000 scope, and people will buy those in pairs. Some high end several GHz ones will turn more profit than a dozen scope.

Something like a '3500G' series, based on the 3000G platform but updated with a 12 bit front end, 100 M points / channel, and priced at a level which isn't totally blind to what's been coming out of China the last few years?

Can you afford to ignore the cheaper competition in the long term?

That's a very interesting point. And raises the question as to why the price difference really.
As a company, buying for more than needed is shooting yourself in the foot, but relying fully on foreign production because it is cheaper is setting yourself up to be shot in the head later.

This would be a very good argument, I think, if it wasn't for the A brands moving production to the east too.

You compare the list price of a product to a price which is made with negotiations in mind.
If you need to buy an expensive scope you are either using taxpayer's money, and pay the list price because nobody cares. Or you are trying to make profit, and negotiate the price down from the list price.

[nctnico]

Have you had any luck trying to get the price down for a business purchase? I have not. In my experience list prices are pretty firm. Ofcourse you might get a good deal during a promotion, on a demo / older unit but other than that, list price it is.

[madires]

Keysight 33210A arb: encoder doesn't work properly after five years and again they are some custom thing made by I think Bourns. Replacement encoder price: astronomical (I don't remember exactly but it was around €180 for two lousy plastic encoder.)

And some people think optical encoders are expensive. This begs the question: Why aren't optical encoders chosen for the most frequently used knobs, at least for professional T&M devices?

[tszaboo]

Have you had any luck trying to get the price down for a business purchase? I have not. In my experience list prices are pretty firm. Ofcourse you might get a good deal during a promotion, on a demo / older unit but other than that, list price it is.

Somehow a a "refurbished" unit turns up always. Not for the cheap stuff usually.

[Martin72]

Honestly, why would they care about the low end market? They can sell an active probe with more profit on it than a DSOX1000 or 2000 scope, and people will buy those in pairs. Some high end several GHz ones will turn more profit than a dozen scope.

If they could only rely on the sale of these "rocket science scopes" in the 6-digit range, they wouldn't have the others.
And that's what it's all about, their price/performance ratio.
"Back then" you had no choice but to buy an expensive scope, no matter what you wanted to do with it, no matter how complex the task.
Now things are different and the question is whether you want to leave this "low-end" market to the others in the hope that enough expensive products can continue to be sold or not.
I don't know their economics to say they don't care about this market.
Teledyne Lecroy, on the other hand, has made it very easy for themselves with the sparsely masked Siglent products in their range.
But then I might as well buy the original.
The so-called "low end" market is nowadays dominated by brands such as Rigol and Siglent, which is already "gone".
But at least Siglent didn't stop there and with the 3000X HD and 6000A/Pro at the latest, they have models that can be placed in the mid-range market in terms of features and performance.
And yet both are still significantly cheaper than other brands in the sector and that would "worry" me as the top dog.
In addition, Siglent has made a start and secured a place in the upper house with the 7000A.
3Ghz, PC platform and for just under 20000€....

[pdenisowski]

The price of test equipment is not in the hardware, it is in the engineering and support.

Generally speaking, this is very true.  The higher you go (in frequency, performance, etc.) the great the impact of component cost. 

You also have to factor in manufacturing / assembly cost.  I worked with one particular product (not a scope) that a competitor of ours tried to copy.  After producing a few prototypes, they gave up when they realized that the real trick was the manufacturing process, not the parts themselves

All major test equipment manufacturers are producing in low wage countries.

Yes, but not exclusively.  We (R&S) still do quite a bit of manufacturing in Germany - touring our facilities is one of my favorite side-trips when I'm in Germany.

https://www.rohde-schwarz.com/us/career/professionals/manufacturing_253344.html

[pdenisowski]

Somehow a a "refurbished" unit turns up always. Not for the cheap stuff usually.

We (R&S) have a demo pool of equipment that we use for, well, demonstrations, and that also loans units to account managers, AEs, customers, etc. on a short- or long-term basis. All of these "demo" units are always for sale.  If you purchase one, it goes back to our service center for a refurb / cal and then is sent to the customer with the full factory warranty. This applies to both expensive and "value" instruments. 

And when I say all of our demo units are for sale, I really do mean all of them.  When I got one of our new NPA power analyzers pre-release, it already had a demo sales sticker on it

[hfleming]

All I can say, is that most of the cheap new stuff on the market blows the lid off from the brand names from not so long ago… spectrum analyzers, oscilloscopes and function generators… when I first used them in the 1980’s, the spectrum analyzer that I used at work (Tektronix 292) costed a lot more than my house. The tektronics oscilloscopes and function generators were not so much cheaper. At few years later we got the first DSO, that didn’t even have a vector display, just a bunch of dots that you had to join mentally in your mind (Nicolette XX). For the next bunch of years I used HP, Tektronix, R&s stuff that costed more than a house. 20 years ago the brand name DSO’s were still crap, always going back to my Tektronics 485 or later 2285 analogue scopes. Lately equipment like the cheap siglent DSO’s, function generators, Spectrum analyzers blows all the pro equipment out of the world that I used not so long ago in the lab. Onliest thing that remains very too pricey for a home lab is a decent RF-generator…. Would love to have But that is just my own opinion. Would love to have modern Keysight, R&S or Tektronix stuff in my home-lab, but can’t afford them, but even the budget stuff I have now still blows my mind compared to the very expensive stuff I used in a lab not so many years ago, IMHO.

[2N3055]

The price of test equipment is not in the hardware, it is in the engineering and support. All major test equipment manufacturers are producing in low wage countries.

This statement is interesting.

So you are saying that my 15000 USD Keysight is so expensive because of the support?
What support do you mean?

Warranty is same as others: return to repair/replace.

Do you mean call them for some support ?
Well you DON'T get that...
In warranty you get some few tickets on general level.

Their (by the way excellent to that matter) support you have to PAY for.

So I paid for support twice: once for overpriced hardware and then if I actually want support I have to pay for it...

That is your truth. What you are saying is simply not true anymore. It used to be.. But not anymore.

It pretty much a racket.

Not to mention that what they called support ticket was me reporting bugs.
Out of two, one was never fixed and never will be.
They should have paid ME for that...
They kept reminding me that these were two complementary "support requests" and that from that point on they will charge me by request..
Or I could get this affordable Support Contract.....

No, big brands have knowledge and if you have to use them for that, you can have it, but it will cost you  additional money.

So you are wrong.
You don't get any support included in very overpriced hardware price.
Only normal warranty like with any other manufacturer.

[nctnico]

I'm not wrong, your experience is just different. I have asked Keysight and Tektronix for support in the past and didn't need to have warranty or a limited number of tickets. I got my technical questions answered to a satisfactory level. It is simple as that. But I did make sure to be really clear, without adding any excess information, in describing the exact problem and what my exact question was.

[2N3055]

I'm not wrong, your experience is just different. I have asked Keysight and Tektronix for support in the past and didn't need to have warranty or a limited number of tickets. I got my technical questions answered to a satisfactory level. It is simple as that. But I did make sure to be really clear, without adding any excess information, in describing the exact problem and what my exact question was.

So you are implying it was my fault... Nice...

I don't know about Tek. But how recent is your Keysight experience?
Did you miss all the hubbub in last year or so about how they changed support structure and whole market approach...