Analog versus digital oscilloscopes 2

[AaronLee]

I cringe whenever I see professional, high-end equipment running on any version of Windows, and subject to the blue screen of death. The same goes for the typical other alternatives of Linux, which I don't consider to be any more robust. And adding the word "embedded" to the front of the OS doesn't seem to make it any more stable. Why can't these vendors invest in a truly stable OS for their high-end gear?

If the system contains NAND flash to hold the program code, it's bound to fail sooner or later. NAND flash by it's nature will fail eventually, even if it's not being written to. Just reading it causes wear. Writing (which involves erasing) causes a lot more wear. Often times a BSP is cobbled together which doesn't take into account the extreme importance of a robust wear-leveling algorithm for NAND. Nor do most firmware engineers realize the importance of making their code to limit writing to NAND more then necessary and using strategies to limit erasures. Many years ago the company I worked for got burned by adopting a BSP for Windows CE which had a poorly designed NAND flash driver. Just like clockwork, the systems began failing about a year after installation, and they all failed within the next 6 months. It cost my company a ton of money to settle the claim. After that, we paid a bit more in order to use Micron eMMC. It's still bound to fail eventually, but with well designed firmware and assuming they got the built-in wear leveling right, it should be a long time in the future, well beyond the warrantee period.

These days, you really can't get around buying gear that doesn't have NAND or eMMC. Occasionally you might get something with NOR flash, but even that's going to fail eventually. It's just the nature of the beast. And if the flash fails, you need to have access to an image of the flash to burn it onto a new chip to replace the damaged one, or your gear just became a boat anchor. At least with the common stuff, there should be enough techies around that someone will have a flash image and willing to send you a copy. If it's not common gear, you'd better learn how to make a backup image yourself.

[RoGeorge]

better learn how to make a backup image yourself.

I'm afraid making your own backup from a working device is already not possible.   

For example, microcontrollers nowadays have internal flash that can be read-protected, so an attempt to read from such a protected chip will return only 0xFF.  Even the humble Arduino Uno (ATmega328) can be read-protected.  Same for FPGA's with internal flash (e.g. Intel/Altera MAX10 series).  And when the flash is external, then the data can be encrypted so a data sniff during boot won't help much either, unless the encryption can be defeated.

[AaronLee]

better learn how to make a backup image yourself.

I'm afraid making your own backup from a working device is already not possible.   

For example, microcontrollers nowadays have internal flash that can be read-protected, so an attempt to read from such a protected chip will return only 0xFF.  Even the humble Arduino Uno (ATmega328) can be read-protected.  Same for FPGA's with internal flash (e.g. Intel/Altera MAX10 series).  And when the flash is external, then the data can be encrypted so a data sniff during boot won't help much either, unless the encryption can be defeated.

Yes, if they've protected it, you might already be out of luck, unless someone's hacked it.

[TheUnnamedNewbie]

Another related question:  Before I retired, my employer purchased some expensive test equipment that ran Windows XP internally.  Did Agilent/Keysight upgrade the OS thereafter?

We have a 1 million USD 32 channel Keysight parallel-bit-error-rate-tester at work that is out of support for the simple reason that MS stopped providing security packages for XP.

[Trader]

Another related question:  Before I retired, my employer purchased some expensive test equipment that ran Windows XP internally.  Did Agilent/Keysight upgrade the OS thereafter?

We have a 1 million USD 32 channel Keysight parallel-bit-error-rate-tester at work that is out of support for the simple reason that MS stopped providing security packages for XP.

Big brands (Keysight, Tektronix, Fluke, LeCroy, R&S, NI, etc) are losing market to new low-cost players, and this will accelerate this process.

The Super-Computers and Web-Servers market is dominated by Linux, maybe in a couple of years, with the first Open-Source/Hardware DSO and a strong community developing all kinds of plugins and features, with customization, the big PC manufactures will enter the instruments segment.

[TheUnnamedNewbie]

I cringe whenever I see professional, high-end equipment running on any version of Windows, and subject to the blue screen of death. The same goes for the typical other alternatives of Linux, which I don't consider to be any more robust. And adding the word "embedded" to the front of the OS doesn't seem to make it any more stable. Why can't these vendors invest in a truly stable OS for their high-end gear?

While I understand the frustration about the OS'es, it kinda makes sense in some applications like high-end scopes. On our UXR scopes we also run other software, like Keysight VSA, Matlab, Python, etc. That needs something like Windows to support it. I've never had one of the Z-series or UXR scopes crash on us in the 5 years I'm working with them. For us, having a 'custom' OS would actually be less prefered over Windows - having Windows in the background makes a lot of things easier, as the same programs we run in simulations on our design computers can be run on the scope/VNA

[rsjsouza]

I cringe whenever I see professional, high-end equipment running on any version of Windows, and subject to the blue screen of death. The same goes for the typical other alternatives of Linux, which I don't consider to be any more robust. And adding the word "embedded" to the front of the OS doesn't seem to make it any more stable. Why can't these vendors invest in a truly stable OS for their high-end gear?

I fully agree with the misuse of the word "embedded" and there are many alternatives for real-time OSes (QNX, Nucleos, VxWorks, etc.), but I suspect the pressure to reduce development time is way too big. Not only windows OS has a larger application/coder base with resources everywhere around the internet, true experts on application development on these niche OSes are much more rare.

[bdunham7]

Big brands (Keysight, Tektronix, Fluke, LeCroy, R&S, NI, etc) are losing market to new low-cost players, and this will accelerate this process.

Are they?  In what market segments?

[Trader]

Big brands (Keysight, Tektronix, Fluke, LeCroy, R&S, NI, etc) are losing market to new low-cost players, and this will accelerate this process.

Are they?  In what market segments?

In all market segments where a non-big-brand device is being purchased.

https://www.marketwatch.com/press-release/digital-oscilloscope-market-growth-2021-cagr-value-business-opportunities-top-key-players-industry-share-size-major-manufacturers-and-driving-factors-with-covid-19-impact-till-2027-2021-06-19

[bdunham7]

In all market segments where a non-big-brand device is being purchased.
https://www.marketwatch.com/press-release/digital-oscilloscope-market-growth-2021-cagr-value-business-opportunities-top-key-players-industry-share-size-major-manufacturers-and-driving-factors-with-covid-19-impact-till-2027-2021-06-19

Huh?  That's what I would call a non-answer.

That link is to an abstract or advertisement, not the article.  It has little to no actual information, other than to apparently list the top 4 DSO manufacturers as Tektronix, Keysight, LeCroy, and Rohde and Schwarz.  It also incorrectly states that Tektronix is a subsidiary of Danaher--that's no longer true--leading me to wonder how good the rest of it is.  Did you get or read the actual article?  If so, what did it say? I would guess that the big brands are not selling products in the low cost market where they have no products (and never have) and aren't attempting to compete.

[Trader]

Big brands (Keysight, Tektronix, Fluke, LeCroy, R&S, NI, etc) are losing market to new low-cost players, and this will accelerate this process.

Are they?  In what market segments?

In all market segments where a non-big-brand device is being purchased.

I would guess that the big brands are not selling products in the low cost market where they have no products (and never have) and aren't attempting to compete.

A while ago, very few people considered buying a scope other from a big brand.

A few years ago, the low-cost chinese scopes (as Rigol or Siglent) are the bestsellers for the Entry Scope market (even in American Colleges and Companies).

The low-cost scope profits feed the R&D for mid-range and high-end scopes. It's clear to see a trend, soon the big brands will be surpassed in all devices (PSU, AWG, DSO, SA, DMM, etc) and segments (low <200MHz, mid <1.5GHz, high >1.5GHz).

IMHO, the FIRST manufacturer to release the Hardware Design and the Code in a "GPLv3" license, and support the community to improve it, will "Dominate" (conquer the passion) from the Open-Source Communities and Crowdsourcing Market.

AFAIK, there is no market for Developing APPs for Instruments and today All bench instruments have firmware and options, think about how much customization could be done for every single instrument (to show more information, to do more tasks, new features and statistics graphics, etc), this would be a good reason to replace the current device to a "Smart"(PSU, AWG, DSO, SA, DMM) like the SmartPhones replaced a lot of devices (including the Phone).

The manufacturer could be concerned just to Make the Hardware and selling services, like an App Certification (similar to Google Play or App Store).

An inspiration maybe could be Tesla that opened all the patents and is free to use any improvements made to its technology by another party.

https://www.vennershipley.co.uk/insights-events/does-teslas-open-source-patent-philosophy-mean-they-are-free-to-use/

https://electrek.co/2015/11/10/a-number-of-companies-are-now-using-teslas-open-source-patents-and-it-has-some-interesting-implications/

IMHO, Keysight could do that with the "Keysight Smart Bench Essential instruments".

[bdunham7]

A while ago, very few people considered buying a scope other from a big brand.

Well that one is wrong.  There have been low-cost second and third tier oscilloscopes for many decades--from manufacturers like BK Precision, GW Instek, Hitachi, Kikusiui and a variety of house brands.  'Big brand' scopes like the ones that are a few hundred bucks on eBay now were just as unaffordable to the average Joe then as a decent Tek scope is today.

A few years ago, the low-cost chinese scopes (as Rigol or Siglent) are the bestsellers for the Entry Scope market (even in American Colleges and Companies).

Perhaps your behind-the-paywall article would have that info, but I don't know.  And schools have been buying lower-cost scopes for years--BK Precision, for example--so unless you have numbers, I wouldn't assume anything.

The low-cost scope profits feed the R&D for mid-range and high-end scopes. It's clear to see a trend, soon the big brands will be surpassed in all devices (PSU, AWG, DSO, SA, DMM, etc) and segments (low <200MHz, mid <1.5GHz, high >1.5GHz).

They do seem to be nipping at the heels of the big dogs, but IMO they still aren't really competitive with anything but the lowest tiers of the A-brand offerings.  And unless they whip their support and service into the proper condition for dealing with the types of clients the A-brands cater to, they aren't going to surpass anyone anytime soon.

IMHO, the FIRST manufacturer to release the Hardware Design and the Code in a "GPLv3" license, and support the community to improve it, will "Dominate" (conquer the passion) from the Open-Source Communities and Crowdsourcing Market.

Yeah, sure.  Open source has been around quite a while, so show me a market segment anywhere where open source 'dominates', especially anything that is hardware based.  The main issue is that there's no money in it, so everyone who wants to make money and is able to do something else does something else. 

[Trader]

WRONG - You are confusing low-cost with low-quality, when comparing "BK Precision, GW Instek, Hitachi, Kikusiui" VS chinese devices during that Same period, just a few years ago the chinese devices achieved a minimum entry-level quality.

WRONG - your perception is that chinese instruments aren't dominating in the entry-level market?  ok, just google "best oscilloscopes", LOL.

They do seem to be nipping at the heels of the big dogs, but IMO they still aren't really competitive with anything but the lowest tiers of the A-brand offerings.

WRONG - they are Very Competitive in Making Hundreds of Millions, and reinvesting on R&D. Try think out the box...  A Ferrari is much better than a Toyota (IMHO), but Toyota make Much, Much More Money than Ferrari. I don't think Toyota wants to make supercars, but I'm sure Rigol and Siglent want to do top high-end instruments, and soon they will start "nipping the BALLS of the big dogs".

show me a market segment anywhere where open source 'dominates'

Android, Red Hat, All Super-Computers OS, Majority of WebServers, movie effects, and many open-source applications in every field, I'm sure your Browser is Open-Source.

[David Hess]

In the past test instruments were divided into separate markets of service instruments and laboratory instruments.  The low tier manufacturers provided service instruments and the high tier ones laboratory instruments.  Some companies like Tektronix tried to provide both in separate product lines.  (1) The various newer companies like Rigol and Siglent may have started providing low end instruments only really suited for service work but now bridge that divide with their high end instruments.

With repairability getting worse, the service category does not have as much meaning as it used to these days and low end test instruments are just that, but I think what has changed is that depreciation of technology has made them cheaper to manufacture so there is a larger market exclusive of the diminishing service market to support them.

(1) Examples of Tektronix "service" product lines include the T900 series, 22xx series, and 3-bay 7000 mainframes.  In that last case, notice that only the 7B53A timebase intended for the 3-bay mainframes included optional TV sync triggering.

[Trader]

Another important effect: who starts using a specific brand, tends to buy (other devices/models) and also recommends to others the Same brand.

FMPOV an entry-level market is very important even for lab-grade manufacturers.

[bdunham7]

WRONG - your perception is that chinese instruments aren't dominating in the entry-level market?  ok, just google "best oscilloscopes", LOL.

I'm not even going to try to respond to the rest of your incoherent ramblings, but the whole point of what I said is that there has been an 'entry level' market for oscilloscopes and other test equipment for a very long time and none of the players in that market have gone on to become an A-brand leader.  The A-brands don't really compete in this area and they never have.  Where did I say that Chinese instruments weren't dominating the current entry-level market?  I'm well aware that they are and I don't think that gives them a clear path to join HPAK/Tek/R&S/LeCroy anytime soon.  "Making hundreds of millions"--where do you get your facts?

[TheUnnamedNewbie]

WRONG - your perception is that chinese instruments aren't dominating in the entry-level market?  ok, just google "best oscilloscopes", LOL.

I'm not even going to try to respond to the rest of your incoherent ramblings, but the whole point of what I said is that there has been an 'entry level' market for oscilloscopes and other test equipment for a very long time and none of the players in that market have gone on to become an A-brand leader.  The A-brands don't really compete in this area and they never have.  Where did I say that Chinese instruments weren't dominating the current entry-level market?  I'm well aware that they are and I don't think that gives them a clear path to join HPAK/Tek/R&S/LeCroy anytime soon.  "Making hundreds of millions"--where do you get your facts?

I'm with you on this one. Our lab keeps buying some of the Rigol and siglent entry level equipment now and then, and every time we get one, and then decided that it's just inferior to the ones by the A brands and nobody uses it. None of the big labs I've gone to used anything but the big names (Keysight, R&S, Tek, Yokogawa, Anritsu...).

It's all about service. The fact that we regularly get asked by R&S to provide questions we have which they will then have their engineers put into a presentation and come explain us. The fact that Keysight sent a applications engineer when we struggeled with their equalization software package, free of charge, within a week.

[james_s]

That's certainly no surprise. The entry level gear is a very good value, it offers a lot of bang for the buck and it is affordable to the hobbyist or pro-sumer who is probably not going to spend thousands or tens of thousands of dollars on high end gear. It is obviously not as good as the expensive stuff and it doesn't try to be, the thing it focuses on is being affordable. Getting a reputation for making decent affordable entry level gear is not beneficial to entering the high end market. Bugatti, Ferarri, Porsche, Mercedes and others don't try to complete on the low end by making basic economy subcompact cars. Companies like Kia make a pretty good product in terms of a reliable and affordable transportation appliance but they are not likely to start building high end supercars, and if they did I would bet they would come up with a separate brand for that line.

[nfmax]

That's certainly no surprise. The entry level gear is a very good value, it offers a lot of bang for the buck and it is affordable to the hobbyist or pro-sumer who is probably not going to spend thousands or tens of thousands of dollars on high end gear. It is obviously not as good as the expensive stuff and it doesn't try to be, the thing it focuses on is being affordable.

This. And indeed, especially compared with old 'service' type CRT oscilloscopes, affordable, entry level DSOs offer spectacularly good value for money. It's a whole different world now!

[bdunham7]

In the past test instruments were divided into separate markets of service instruments and laboratory instruments.  The low tier manufacturers provided service instruments and the high tier ones laboratory instruments.  Some companies like Tektronix tried to provide both in separate product lines.  (1) The various newer companies like Rigol and Siglent may have started providing low end instruments only really suited for service work but now bridge that divide with their high end instruments.

Service and entry-level aren't the same thing, IMO, although perhaps they overlap at the lower end.  Entry level and educational instruments have always occupied a level below 'service', or at least the better tiers of service.  Something like a Tek 2235 was clearly a service-oriented scope, but certainly not entry-level priced.  (something like $2K in the '80s) There were low-end 10 and 20 MHz models that were the market equivalent of a Rigol DS1054Z.  It might appear that the new entry-level makers have encroached on the service market, but that's only because the stuff that is getting fixed is often pretty old so relatively low performance scopes suffice.  Instruments intended to service modern high-speed electronics do exist, but that level of equipment is still well beyond the hobbyist price point.

I don't know the current state of the marketplace, but from my perspective the TOTL Rigol and Siglent technology is only competitive on a 'bang-for-buck' basis or with decades old used instruments.  And assembling an instrument with COTS components doesn't make you technologically advanced.  I'm not knocking their products, but they are nowhere near cutting-edge technology except in the cost-optimization arena.  And while that is great for the bang-for-buck enthusiast who gets a boatload of features and performance for a relative pittance, I think the idea that they will soon surpass the A-brands is silly.  They've only recently come up to the level of the two-decades old TDS5104, if that.

[RoGeorge]

...
With the current trend, I see everything will be closed and locked-down and encrypted, first to a certain brand and environment, and very soon (as in a few decades away from now) we will see everything (everything as in objects, services, etc.), even your faithful DMM or your faithful handgun, everything will require a personally encrypted key and a valid personally authorization, or else that object/service won't be available to you.

This will happen whether we like it or not, simply because the actual level of technology allows to have that.

... 2 days later ...
https://hackaday.com/2021/08/02/home-depot-is-selling-power-tools-that-require-activation-in-store/   

Oh well, I didn't mean it that fast, though! 

[magic]

So far limited to 3rd world countries struggling with 3rd wold problems; still no pay per use

 

[RoGeorge]

still no pay per use

Literally pay per screw!   

[james_s]

...
With the current trend, I see everything will be closed and locked-down and encrypted, first to a certain brand and environment, and very soon (as in a few decades away from now) we will see everything (everything as in objects, services, etc.), even your faithful DMM or your faithful handgun, everything will require a personally encrypted key and a valid personally authorization, or else that object/service won't be available to you.

This will happen whether we like it or not, simply because the actual level of technology allows to have that.

... 2 days later ...
https://hackaday.com/2021/08/02/home-depot-is-selling-power-tools-that-require-activation-in-store/   

Oh well, I didn't mean it that fast, though! 

It's perfectly reasonable given the massive losses from shoplifting and from what I gather losses have increased sharply. This is not some kind of rental thing, the tool is activated at the time of purchase and then it's yours forever. As long as you don't buy stolen tools it should not be a problem. I'm also fairly sure a determined person could bypass the protection but the sort of people who make a living selling stolen power tools might not want to bother with that.

I can't help thinking it would be more effective to simply keep the stock of valuable items locked up and use the business model where you grab a tag off the shelf, take it to the checkout counter and pay, then they give you the item you purchased. They could streamline it and use an automated conveyor to deliver to the checkout stand. Or just embed anti-theft tags inside the tools at the factory, I've seen things that had that. 

[TimFox]

When CDs were new and easy to shoplift, it was common to bring the “jewel case” to the checkout where they inserted the disc from stock behind the counter, whereafter it was your property.  Same thing.