please dont purchase PC controlled test equipment

[oz2cpu]

Headline : please don't purchase PC controlled test equipment
I mean of course ONLY controllable by pc

It is really driving me nuts, as a collector of funny old test equipment,
every time some stuff is controlled only by a pc, and a bit old,
it is impossible to get software, drivers, support
and operating systems changed 4 versions since...

this means such stuff (pc only control equipment)
come with a very short expected lifetime,
some of it is even just as expensive as real standalone equipment with remote capabilities,
this means in many years when the remote features are no more supported
you still have a working standalone unit, and not trash.

my latest equipment score is a good example :
https://youtu.be/93_Zg-mKWks

[TomKatt]

I wish there was a good answer to that dilemma, because it goes beyond just test equipment.  In our shop we've had to replace CNC equipment that hardware wise is still operational, but if the pc based controller fails the cost of repair is so close to an entire new machine it doesn't make sense to invest repairing it.  Recently an automated panel saw with a controller based on Microsoft's embedded "mobile" platform failed and when we located one of what appears to be the last 3 controllers left on earth the price was stupid expensive. 

I've even had to set up virtual machines trying to emulate old pc systems because the software will not run on modern 64 bit hardware.  And, as you point out, that's IF you can locate the software to begin with.

Absolutely nuts, though I have no idea what the solution is.  PC's provide flexibility and even with dedicated PLC type hardware you'll never escape the firmware / software.  The days where equipment was built with nothing but discrete hardware containing no software are gone.

[xrunner]

Headline : please don't purchase PC controlled test equipment
I mean of course ONLY controllable by pc

It is really driving me nuts, as a collector of funny old test equipment,
every time some stuff is controlled only by a pc, and a bit old,
it is impossible to get software, drivers, support
and operating systems changed 4 versions since...

 

I buy old test equipment and yea some of the software gets outdated but ... whatever you buy today will be 10 years old in ten years, 20 yrs old in 20 years, ... and so on. Who can say what system will be controlling them decades from now. I don't know what you can buy today that is future proof.

[jonpaul]

all my HP, Keysight HPIB, GPIB work fine, for decades, using the NI interface card to PCI bus PC and USB GPIB adapters.

Program with NI la view or HP Basic.

What's your problem with instruments controllers?

j

[Stray Electron]

  Uggg.  I don't recommend buying ANYTHING that is only controllable via a PC.  Over the last couple of decades I've already been through all sorts of printers, scanners, home automation controllers, etc that ran from a PC, only to have it become USELESS when MicroSloth came out with a different version of Windows and the drivers no longer work.

   It's bad enough when you're talking about a few hundred dollars worth of printers or scanners but many hundreds of dollars or thousands of dollars in TE? Nope it's NOT going to happen!

[TomKatt]

I buy old test equipment and yea some of the software gets outdated but ... whatever you buy today will be 10 years old in ten years, 20 yrs old in 20 years, ... and so on. Who can say what system will be controlling them decades from now. I don't know what you can buy today that is future proof.

Storage media formats from 20 years ago are already difficult to deal with.  Even some digital formats are problematic because of software incompatibilities.  Things will only get worse as the advance in technology rate increases...

Maybe there is some justification to not building things as well as they used to...  Considering you might not be able to operate the gear for some reason if it's still functioning.

My 1940's Waterman 'Pocket Scope" cro still works 70+ years on.  Will my Siglent SDS1104X-E still be running in 70 years?  I have a suspicion it will not   

[m k]

One answer is backup, the hardware one.

There will be time when computer part is dirt cheap.
Like a minimal PC case with necessary stuff etc.

CNC and other professional things are possibly more difficult, there can be minimal propriety parts that are not available practically ever and even that the missing part can be pretty easily substituted with something regular the process must be known.
My example here is a dongle that was actually an ISA card, possibly just some extra ROM.

Generally we are now familiar with too good hardware and what it does to manufacturers but it is changing, at least in EU, for how long I'd say that not very long, manufacturers must still live.
My prediction is that leasing will come down to hilarious levels.

[coromonadalix]

  why  not    loll  sound more a rant than nothing ??


With good drivers    you can even control one with an excel sheet having macros   loll

i did some basics for one of my meters  loll

I use agilent "old dmm softwares"  loll   hit a small wall with visa drivers and now  with R&S  ones works perfectly on Win11

[2N3055]

There are many examples where people changed a PC motherboard on and old scope with new different one to repair and/or upgrade... They could do that because scope was in fact a windows PC.
Picoscope supports 15-20 years old hardware with software that is still active...

You can get old PCs cheap.

I guess it is not worse but more like different approach.

[nctnico]

There are many examples where people changed a PC motherboard on and old scope with new different one to repair and/or upgrade... They could do that because scope was in fact a windows PC.
Picoscope supports 15-20 years old hardware with software that is still active...

Picoscope is a rare exception. And changing a motherboard in an old scope is much harder than you'd think as the OS needs to support the newer hardware as well. Lot's of older scopes run Win2k or maybe Windows XP. Windows 95/98 is not out of the question either. Try to add support for USB to such old hardware. Been there, done that a few times already.

You can get old PCs cheap.

But really old PCs (like the ones with 486 processors and ISA slots that can run Windows 3.11 / Windows 95) are getting expensive. Recently I got rid of a whole bunch of such PCs and a guy drove over 200km to pick them up.

[oz2cpu]

thanks a lot for all the great comment about stuff that is ONLY controlled by PC...
glad to see i am not the only one with this dilemma.

the example video i posted is far from the first time i have seen this,
it was a "hate" that was build up over many years, and now i just had to let it out,
now i feel a little bit better.. need a hug, then i goto bed..

[zrq]

Still being proud of how I hacked HP 53305A (29 years old Win16 software）to work half-natively on my 2022 computer and modern GPIB adapters, I'd like to take the challenge of preserving headless PC controlled test equipment. This is probably another source of fun for me, especially after trash dump diving finding things like a 20 year vintage ROM programmer (-> magically found the software online, runs fine in Win2k VM with USB pass-through) or a 25 years old RS232 connected inclinometer (-> works with commands for newer models and magically found the software online later, runs natively on Win11, Windows's compatibility is amazing).

For example, if anyone have a HPAK L4411A for sale at a reasonable price, I'll be happy to get one. For me the lack of front panel is a just good space saving feature as I'm fully capable of hacking to at least get the obsolete Java virtual front panel working.

[coppercone2]

thats like buying kits

[thm_w]

my latest equipment score is a good example :
 

Kinda on you for buying that crap in the first place.
I'll assume you got it very cheap?

[luma]

When I’m purchasing a tool, the absolute last concern on my mind is the plight of the dude who buys it off me on eBay a decade later.  You know what you’re buying, and PC control is sometimes part of the challenge in bringing old hardware back into use.  If that’s not what you want to be doing, maybe just don’t buy it?

Suggesting that nobody buys a tool they find useful so that you have an easy getting it working some years later is preposterous.

[MK14]

Headline : please don't purchase PC controlled test equipment
I mean of course ONLY controllable by pc

So what about the Analog Discovery and Analog Discovery 2?

[2N3055]

There are many examples where people changed a PC motherboard on and old scope with new different one to repair and/or upgrade... They could do that because scope was in fact a windows PC.
Picoscope supports 15-20 years old hardware with software that is still active...

Picoscope is a rare exception. And changing a motherboard in an old scope is much harder than you'd think as the OS needs to support the newer hardware as well. Lot's of older scopes run Win2k or maybe Windows XP. Windows 95/98 is not out of the question either. Try to add support for USB to such old hardware. Been there, done that a few times already.

You can get old PCs cheap.

But really old PCs (like the ones with 486 processors and ISA slots that can run Windows 3.11 / Windows 95) are getting expensive. Recently I got rid of a whole bunch of such PCs and a guy drove over 200km to pick them up.

All you said is right but you missed the point. It is easier to find a replacement PC board (or PC) than really single purpose custom PCB for a single oscilloscope product. I.E. it is easier to find a replacement PC and ALL needed software than some custom PCB and special software for an instrument.
And most of the time you don't need it to be exactly the same, but close enough...

My old Pico 212-100 needs 32 bit Windows 7 and a software that is downloadable from Picotech.
It is more than 20 years old. Still works.
That means a bit older PC but those are still around. I actually have a laptop for it that was given to me for free..

But if we ignore Picoscope, truth is that you need to dedicate a PC for your T&M equipment. I have few Picos, Digital Discovery, a SignalHound SA, few USB logic analysers.. And have one PC and touchscreen monitor at my workstation for them. Which also comes in handy to open datasheets, BOM (very nice interactive BOM in Kicad for instance). I also run automation scripts on it , in addition to a Raspbery PI (for longer running stuff)...

Like I said, PC based instruments are different, so your business continuity is different. For instance, I can easily buy another PC (or cheap PC parts), any monitor, and make an image of disk, a backup.
If an LCD on my Keysight scope dies, what are my options here?

It is simply different sets of problems and solutions.

[james_s]

I wish there was a good answer to that dilemma, because it goes beyond just test equipment.  In our shop we've had to replace CNC equipment that hardware wise is still operational, but if the pc based controller fails the cost of repair is so close to an entire new machine it doesn't make sense to invest repairing it.  Recently an automated panel saw with a controller based on Microsoft's embedded "mobile" platform failed and when we located one of what appears to be the last 3 controllers left on earth the price was stupid expensive. 

Did you take a look inside it? Earlier this week I fixed a controller for a ~$80k CNC machine. It turned out it was an industrial PC motherboard, the sort where the whole computer is on a card that plugs into a backplane, Pentium4 that runs Windows XP embedded. The machine company wanted $1k to look at it and estimated $10k-$12k to repair it. I fixed it in a day for under $50 by replacing a whole pile of bulging electrolytic capacitors on the motherboard.

[oz2cpu]

>>I'll assume you got it very cheap?

sure .. i get trash for free,
no one will be silly enough to pay anything for this set..
maybe 5 $ for the case ? since they can quite easy be re used

[TomKatt]

Did you take a look inside it? Earlier this week I fixed a controller for a ~$80k CNC machine. It turned out it was an industrial PC motherboard, the sort where the whole computer is on a card that plugs into a backplane, Pentium4 that runs Windows XP embedded. The machine company wanted $1k to look at it and estimated $10k-$12k to repair it. I fixed it in a day for under $50 by replacing a whole pile of bulging electrolytic capacitors on the motherboard.

Yes, we did.  While we also have equipment using industrial pc hardware like you describe, this unit was some oddball custom 386 base pc with MS Mobile Embedded in ROM on the board.  We even sent it out to a repair shop that claimed they could repair it (they have done many other repairs for us) and they couldn't resolve it either.  I'm sure from a technical pov it wouldn't be too hard to replace it with some generic pc system as the actual IO is run by PLC's in the cabinet, but apparently the manufacturer did not have any generic software that would interface with the PLC system as that software was all in the ROM.  So their solution was $40K+ to replace everything when the saw went for around $60K on the used market.

[Kleinstein]

It depends on the equipment. The 2 parts (DSO and AWG) from the start of the thread are testequipment with a rather short time value.  It was clear from the start that they loose much of there value over a relatively short time because better and cheaper parts come out. With such units that a evolving fast, there is not much value in a 3 year old unit - just like with PCs - there value was going down quite fast in the 1990s. So when you need to change the PC you don't loose the expensive instrument it one was, but something outdated with very limited value. A similar old stand alone DSO would also be obsolete and in many cases go to the waste - maybe a few years later, but not by much.  Having it PC controlled saves on the HW costs to start with  - so overlall not such a bad solution.

The PC control can be a problem for other, more longer lived instruments. The typical solution is to keep an old PC and maybe even an extra reseve one. I remember using an old original IBM PC-XT for such an old, but still OK scientific instrument in 2003. It just got a bit tricky to transfer and backup data. We somehow manged to move it to a newer (AFAIR 486) machine that still had an ISA bus. IFAIK it was than running some DOS 5.0 or similar.

Dedicated computers as part of instruments are often worse, as it gets even more tricky to get spares or documentation. If really needed one could still get a standard PC from the 1980s, as there were quite a lot of them around and compatibilty was reasonable. With a custom VME bus 68K system things get way more tricky.

[m k]

oddball custom 386 base pc with MS Mobile Embedded in ROM on the board.

Was the ROM embedded into something else?

Code emulation is the next step, for generic cases, special ones are special ones.
DOS and 16bit Windows code can be emulated today without any problems.
The generic problem is obsolete hardware that is controlled directly by the software using I/O commands.

Some motherboards without ISA slots are still ISA, there PS/2 connectors are not enough but serial and/or parallel port is pretty close.
You can also add a PCI to ISA bridge card and have your hardware supported but old software is still locked.
There you possibly have a completely supported environment but new bridge driver and running emulator may not compute.

Final answer is Linux, its DOS emulator can support those I/O commands.
There are still cli/sti style stuff but I've heard the system works.

Not so long ago I bumped into INT 34h and the likes of Turbo C, it's a DOS era x87 emulation system.
There supporting compiler generates a code that will be replaced in case of actual x87 being present, means that first round is slow but interrupt modify its presence away and then from that on the actual x87 command is executed.

Modifying I/O commands is trickier since they are native, short and not meant to be anything else, so replacement is not an option.
But nowadays exceptions are generated, there program execution is completely rerouted so problem solved.
Until it's not, that exception must be handled and that handler is not very usually working, if even implemented.
M$ case is clear, old software is not fully supported.

Are GDI printers still around?

[artag]

Still being proud of how I hacked HP 53305A (29 years old Win16 software）to work half-natively on my 2022 computer and modern GPIB adapters, I'd like to take the challenge of preserving headless PC controlled test equipment.

That sounds useful. I've run it on an old laptop but I don't really want to preserve it for just that. I'd really rather run it in wine etc. as I don't even want to run recent windows, but I'm sure documenting what you did would be of interest and not just to mel.

[zrq]

Still being proud of how I hacked HP 53305A (29 years old Win16 software）to work half-natively on my 2022 computer and modern GPIB adapters, I'd like to take the challenge of preserving headless PC controlled test equipment.

That sounds useful. I've run it on an old laptop but I don't really want to preserve it for just that. I'd really rather run it in wine etc. as I don't even want to run recent windows, but I'm sure documenting what you did would be of interest and not just to mel.

I have a post for that https://www.eevblog.com/forum/metrology/using-hp-53305a-phase-analyzer-software-on-modern-windows/msg4466365/#msg4466365 . Please use the DLL burried in the source code zip and ignore the attachment in the main post, I recall I fixed some bugs.

Wine is more complicated...You may be lucky and make another layer of shim then it works, https://stackoverflow.com/questions/39574355/is-it-possible-to-call-native-linux-api-from-a-windows-application-running-in-wi , or it's completely impossible, I haven't look deep into it so I don't know.

Wait.. didn't you already replied to that thread half year ago?? 

[james_s]

Yes, we did.  While we also have equipment using industrial pc hardware like you describe, this unit was some oddball custom 386 base pc with MS Mobile Embedded in ROM on the board.  We even sent it out to a repair shop that claimed they could repair it (they have done many other repairs for us) and they couldn't resolve it either.  I'm sure from a technical pov it wouldn't be too hard to replace it with some generic pc system as the actual IO is run by PLC's in the cabinet, but apparently the manufacturer did not have any generic software that would interface with the PLC system as that software was all in the ROM.  So their solution was $40K+ to replace everything when the saw went for around $60K on the used market.

Well that's annoying. It illustrates the issue with equipment that should last many decades relying on computer technology that becomes obsolete after just a few years.

[2N3055]

Yes, we did.  While we also have equipment using industrial pc hardware like you describe, this unit was some oddball custom 386 base pc with MS Mobile Embedded in ROM on the board.  We even sent it out to a repair shop that claimed they could repair it (they have done many other repairs for us) and they couldn't resolve it either.  I'm sure from a technical pov it wouldn't be too hard to replace it with some generic pc system as the actual IO is run by PLC's in the cabinet, but apparently the manufacturer did not have any generic software that would interface with the PLC system as that software was all in the ROM.  So their solution was $40K+ to replace everything when the saw went for around $60K on the used market.

Well that's annoying. It illustrates the issue with equipment that should last many decades relying on computer technology that becomes obsolete after just a few years.

First of all, what manufacturer is claiming a CNC machine will last 25 years? Or a scope?.
They are very good if they provide any support more than 7 years after it was stopped being produced.

This is a good example how it would be better if they made a machine run on off the shelf DELL PC for which you could find replacement. But they used custom MB, so only course of action is that manufacturer should stock hundreds of spare motherboards for repair of machines in supported period..  Which they don't..

As I said, people can like or not PC based T&M equipment based on preferences or use case, but reparability is actually slightly better than for fully custom hardware.

[precaud]

What exactly is the issue?
I still have almost every computer I have ever bought, starting in the early 80's with HP 9845B and Z-80 CP/M-based Eagle IIE's.
They ALL still work fine.
Most have never had any maintenance or repair, other than cleaning (it's dusty here) or replacing disk drives.
Statistically, that makes them more reliable than the standalone equipment being suggested by the OP.

[m k]

Wrong part have had too many working hours.

[james_s]

First of all, what manufacturer is claiming a CNC machine will last 25 years? Or a scope?.
They are very good if they provide any support more than 7 years after it was stopped being produced.

I don't know what the manufacture claims, but these machines cost hundreds of thousands of dollars new, the latest one is my friend's shop was just over $1M, they HAVE to last a long time. On average I think they have retired/sold/scrapped machines when they are around 20 years old, you run it until the ways, ballscrews, spindle, etc are worn out and the tolerance drops below what is acceptable, then sometimes you shift it to jobs that require less tolerance or sell it to somebody else doing lower tolerance jobs and it keeps running. A 10 year old machine can easily be worth $50k-$100k, it is well worth keeping them going. They still have a few much, much older machines in use but mostly manual stuff. It's annoying if you have to scrap an expensive machine that is not worn out because the controller has failed.

[nctnico]

Yes, we did.  While we also have equipment using industrial pc hardware like you describe, this unit was some oddball custom 386 base pc with MS Mobile Embedded in ROM on the board.  We even sent it out to a repair shop that claimed they could repair it (they have done many other repairs for us) and they couldn't resolve it either.  I'm sure from a technical pov it wouldn't be too hard to replace it with some generic pc system as the actual IO is run by PLC's in the cabinet, but apparently the manufacturer did not have any generic software that would interface with the PLC system as that software was all in the ROM.  So their solution was $40K+ to replace everything when the saw went for around $60K on the used market.

Well that's annoying. It illustrates the issue with equipment that should last many decades relying on computer technology that becomes obsolete after just a few years.

First of all, what manufacturer is claiming a CNC machine will last 25 years? Or a scope?.
They are very good if they provide any support more than 7 years after it was stopped being produced.

AFAIK pick & place machines for electronics assembly see very long service lifes. From what I've read & heard it is not out of the ordinary to have machines serviced/refurbished that are 20+ years old.

[graybeard]

Headline : please don't purchase PC controlled test equipment
I mean of course ONLY controllable by pc

It is really driving me nuts, as a collector of funny old test equipment,
every time some stuff is controlled only by a pc, and a bit old,
it is impossible to get software, drivers, support
and operating systems changed 4 versions since...

this means such stuff (pc only control equipment)
come with a very short expected lifetime,
some of it is even just as expensive as real standalone equipment with remote capabilities,
this means in many years when the remote features are no more supported
you still have a working standalone unit, and not trash.

my latest equipment score is a good example :
https://youtu.be/93_Zg-mKWks

I agree.  I have always refused to buy equipment that will not work in a stand-alone mode, both for my home lab, and at work.
If I want remote capability the equipment must have either a IEEE488 or serial port.   For home eithernet connections work great.  At work I am not allowed to connect anything to a network or USB port that is not supplied by IT, so only IEEE488 and serial are acceptable.

[nigelwright7557]

Maybe the answer is to keep the old pc with its version of Windows ?

I sell software and you would be gobsmacked how many people ask if it will work with Windows XP.

[tautech]

If I want remote capability the equipment must have either a IEEE488 or serial port.   For home eithernet connections work great.  At work I am not allowed to connect anything to a network or USB port that is not supplied by IT, so only IEEE488 and serial are acceptable.

Really, in this day and age your work IT are so behind the times ? 

Me thinks you need drop them a link to SDS6000L and watch them go into a spin. 

[m k]

Device doesn't matter.
With USB or RJ45 nobody can really say what happens behind the scenes.
Other issue is what IT knows but that's their headache and no salary breaks for others.

For machine years,
maybe somebody should ask Heidelberg how long their GTO 46 or "Windmill" will last.

I think Tiegel is incredible machine.
You just can't believe it before you see it operating.

[Wolfgang]

Disagree.

The problem is not that a piece of equipment is only controllable by a PC, the problem occurs when this control is not done in a standard way like SCPI and special, soon outdated software is used. Then I agree, never buy this.

On the other hand, a lot of stuff used in automation will never be operated by hand once the programming is done. This means you dont need a screen, a keyboard, switches, ... all that stuff. What I do is using standard lab equipment (DMMs, generators, scopes, whatnot) during the design phase, and then I make special appliances that are SCPI controlled to do the automated stuff. With an Arduino its possible to make quite smart stuff that costs only a fraction of all the fancy gear with GUIs. That works great and its fun.

regards
  Wolfgang

[alm]

The problem is not that a piece of equipment is only controllable by a PC, the problem occurs when this control is not done in a standard way like SCPI and special, soon outdated software is used. Then I agree, never buy this.

How many commercially available instruments that rely on PC control use SCPI for communication? I doubt you'll find many. Can you name a PC-controlled oscilloscope that talks SCPI to the host PC over a standard protocol like USBTMC? Or a logic analyzer? Or a function generator? Some VXI cards maybe?

On the other hand, a lot of stuff used in automation will never be operated by hand once the programming is done. This means you dont need a screen, a keyboard, switches, ... all that stuff. What I do is using standard lab equipment (DMMs, generators, scopes, whatnot) during the design phase, and then I make special appliances that are SCPI controlled to do the automated stuff. With an Arduino its possible to make quite smart stuff that costs only a fraction of all the fancy gear with GUIs. That works great and its fun.

HP used to offer some equipment without front panel to be used only by GPIB control. It would not surprise me if that is still possible. But in general I highly doubt that removing the front panel is going to have that much impact on the BOM costs for anything but a very basic device like a simple bench power supply. How much of your saving is because the lack of front panel, and how much is because it's designed for a single purpose, less margins, less overhead, less compliance testing, etc? Modern front panels are using fewer mechanical switches and no expensive multi-turn pots, so if anything front panels are getting cheaper.

If you compare say an NI offering for a USB/PXI card to a bench instrument, then I doubt you'll be much cheaper. And good luck using it when NI stops supporting the device for newer OS versions or CPU architectures (x86 -> ARM). Their main advantage is density in a commercial production environment.

[jasonRF]

Headline : please don't purchase PC controlled test equipment
I mean of course ONLY controllable by pc

It is really driving me nuts, as a collector of funny old test equipment,
every time some stuff is controlled only by a pc, and a bit old,
it is impossible to get software, drivers, support
and operating systems changed 4 versions since...

this means such stuff (pc only control equipment)
come with a very short expected lifetime,
some of it is even just as expensive as real standalone equipment with remote capabilities,
this means in many years when the remote features are no more supported
you still have a working standalone unit, and not trash.

I believe everyone should buy the equipment that is best suited for their needs.  For some people and some needs that will include equipment with PC-only control, while for others it will not. 

Not long ago I had to be on travel and of course brought my laptop as always.  But I also wanted to bring a scope and packed a USB model with PC-only control.  It took virtually no space and weighed nothing.  Had I instead tried to bring a bench scope, I would have had to pack it to survive commercial air travel, and since I did not have room in my carry-on bag would have had to check a bag which would have meant an extra fee.  Are you really saying that someone who must travel like this often should never buy a usb scope, and instead lug around clunky bench scope and pay an extra baggage fee on many flights just so you can pick up the old bench scope in 20 years and make it work?  And even if the usb scope is junk sooner than the bench scope (not clear if you are traveling with it all the time), how many usb scopes can you throw out before you have generated the same amount of trash as one bench scope? 

jason

[JPortici]

Yes, we did.  While we also have equipment using industrial pc hardware like you describe, this unit was some oddball custom 386 base pc with MS Mobile Embedded in ROM on the board.  We even sent it out to a repair shop that claimed they could repair it (they have done many other repairs for us) and they couldn't resolve it either.  I'm sure from a technical pov it wouldn't be too hard to replace it with some generic pc system as the actual IO is run by PLC's in the cabinet, but apparently the manufacturer did not have any generic software that would interface with the PLC system as that software was all in the ROM.  So their solution was $40K+ to replace everything when the saw went for around $60K on the used market.

Well that's annoying. It illustrates the issue with equipment that should last many decades relying on computer technology that becomes obsolete after just a few years.

First of all, what manufacturer is claiming a CNC machine will last 25 years? Or a scope?.
They are very good if they provide any support more than 7 years after it was stopped being produced.

This is a good example how it would be better if they made a machine run on off the shelf DELL PC for which you could find replacement. But they used custom MB, so only course of action is that manufacturer should stock hundreds of spare motherboards for repair of machines in supported period..  Which they don't..

As I said, people can like or not PC based T&M equipment based on preferences or use case, but reparability is actually slightly better than for fully custom hardware.

It's not uncommon at all. Industrial machinery can cost tens to hundreds of thousands and is build to a certain degree of quality. Ten years ago when i dabbled in internships in that area it was no strange to find machinery with windows 3.11 PCs controlling them. The idea of scrapping a perfectly good (and costly) machine because the computer is outdated is ridiculous, but that means that you have to be able to keep the old pc running

[Wolfgang]

Disagree again.

Almost all "renowned" manufacturers offer SCPI-enabled instruments.

For DMMs: Keysight/HP/Agilent, Rohde&Schwarz, Siglent, Rigol, Keithley, Tektronix, ...
For Scopes: Keysight/HP/Agilent, Rohde&Schwarz, Siglent, Rigol, Keithley, Tektronix, LeCroy, ...
For Signal Generators: Keysight/HP/Agilent, Rohde&Schwarz, Siglent, Rigol, Keithley, Tektronix,  ...

list may be continued. All of them now offer remote-only instruments. The market is clearly automated testing
in a factory environment, and this market is huge. Here you need the max of test functionality in a cramped space. All this PXI stuff is just doing that.

[horo]

Headline : please don't purchase PC controlled test equipment
I mean of course ONLY controllable by pc

It is really driving me nuts, as a collector of funny old test equipment,
every time some stuff is controlled only by a pc, and a bit old,
it is impossible to get software, drivers, support
and operating systems changed 4 versions since...

this means such stuff (pc only control equipment)
come with a very short expected lifetime,
some of it is even just as expensive as real standalone equipment with remote capabilities,
this means in many years when the remote features are no more supported
you still have a working standalone unit, and not trash.

Here, open source (OS), open firmware (OF) together with open schematics (OSch) would give your devices an eternal life.
One example I maintain is the >10 year old Hantek6022 DSO -> OpenHantek6022 (OS, OF, OSch).
Other examples are tinySA (OS, OF) and NanoVNA (OS, OF, OSch).
And if you are willing to give up Windows in your lab, then you can stop the artificial software rot, Linux supports old parallel port devices as well as historical GPIB and SCSI interfaces right out of the box.

[Berni]

There is a time and place for everything.

All reasonably modern reputable test gear supports SCPI without the need for special software. The PC control only versions of test gear is simply the same instrument with the front panel ripped away. Useful for packing as much gear as possible in an industrial test system rack.

The issue comes with gear that was originally designed to ONLY be PC controllable. Those are often optimized to offload as much of the work to the PC as possible, so that the hardware could be cheaper. This is impossible to do without having proprietary PC side software. For example there are realtime spectrum analyzers that are just a calibrated SDR in a box, then the PC uses there massive amount of processing power to actually FFT that very quickly.

Yeah i don't necessarily like this way of doing things, but it does make things more portable for people that need it. However if the piece of PC controlled gear is cheap it doesn't really matter if you have to throw it away once the software support is gone. Like for example i am a fan of those little USB logic analyzers. They are little dumb boxes that just barf the state of digital pins into the USB port and let the PC actually take care of any processing or storage. This makes the really cheep and tiny. At the same time the software is nicer to use with a mouse compared to a old huge boatanchor logic analyzer. It takes up very little space on your bench...etc. The cheap SDRs can also be very capable modulation analysis equipment.

So there are cases where a PC controlled instrument is a good idea.

[DavidKo]

The simple answer is cost.

In past buying a HPIB (GPIB) interface was not cheap. To sold single instrument which needed PC to operate to someone with its own PC was cheaper than pushing him to buy instrument (with PC inside), GPIB interface, cable etc. Now the situation is opposite and SW solutions tends to be cheaper than HW solutions (like software defined radio, which can substitute more devices at once) and powerful computers are cheap. Cheaper solutions wins for the buyer, since he need to pay price now, not after selling the device second hand.

[TomKatt]

It's not uncommon at all. Industrial machinery can cost tens to hundreds of thousands and is build to a certain degree of quality. Ten years ago when i dabbled in internships in that area it was no strange to find machinery with windows 3.11 PCs controlling them. The idea of scrapping a perfectly good (and costly) machine because the computer is outdated is ridiculous, but that means that you have to be able to keep the old pc running

LOL - We have a CNC running on Windows NT right now 

We have spares for most of the pc controller parts, but some items like ISA controller boards can be difficult to source.

And it is a real shame to basically scrap the machine because it's brain dead.  I still find it hard to believe that it can be cheaper to replace the entire machine rather than simply upgrade the controller, but in virtually every case we've encountered that's how it works out.  I'm sure given some time and resources we could refurb those things in-house, but in a commercial manufacturing environment, time is money and also floor space real estate is limited (these are large machines)...  It simply isn't feasible to revamp them ourselves.

I don't think there is really a good answer - pc control does offer the most flexiblity and bad as it is it's still easier to deal with than proprietary PLC controllers.  I'm sure someone must be refurbing these things on the used market, but we purchase used equipment frequently and I've rarely seen one that has been refurbed with new controllers.

[Berni]

Yeah CNC machines are often run by computers from the stone age.

I am guessing the reason they don't get upgraded is logistical difficulties of having the space for it and hauling it around. If it is in a factory it needs to be worked on in place, being in the way. If someone else wants to fix one up and sell it then they need to haul the massive metal behemoth over to there place, fix it up, then haul it to the costumer. Also when they get the machine, they might not have any idea of the condition of the mechanical parts. It is possible the ways,bearings,ballscrews..etc are worn to hell and have heaps of slop, maybe the machine had a bad crash in its history that bent stuff etc.. So it might not just need new electronics but also some special precision mechanical parts that are also near unobtainable by now too. Even then companies will want someone to call up when the machine breaks, so the guy who refurbishes these things would also have to be the one fixing them. Since they are refurbs every machine would be slightly different so maintenance could get more complicated to provide, especially when the company expects to have the machine back up and running ASAP.

Id guess the dicking around is not worth for companies that just want a reliable machine that will pump out parts all day to make money. After a while the parts pay off the cost of the machine anyway. If it breaks down 30 years later they figure it made plenty of profit anyway, it might be worn to hell by now, so time for a new machine.

Still an excellent project to take on for getting a cheap CNC into the home shop. That is if you have the space for one and find a machine small enough (They are often too massive for what you need). That way the large amount of labor of fixing it up and figuring out how to do it is part of the fun rather than the cost.

[TomKatt]

Even then companies will want someone to call up when the machine breaks, so the guy who refurbishes these things would also have to be the one fixing them. Since they are refurbs every machine would be slightly different so maintenance could get more complicated to provide, especially when the company expects to have the machine back up and running ASAP.

I know that at least for us, support is rated nearly equally to a machine's ability / features.  Down time costs a lot of money.  So unless a refurbished machine comes with support (including a stocked replacement parts supply), we're not likely to purchase it.  I suspect that many businesses operate with similar requirements.

[zrq]

Isn't a PC based embedded controller runing some Windows usually easier to service and replace than a customized digital board running some obscure long-dead RTOS? There isn't much one can do if such customized thing break for example because of storage card corruption.

At least one can always do the disk imaging in a relatively standard way and in case some software break, you know how to load the exe to IDA or GHIDRA. And it will be the better documented x86 rather than PPC or  some weird shit.

[m k]

Easier is relative, one is hobbying and other one is starving.

[pdenisowski]

Disagree again.

Almost all "renowned" manufacturers offer SCPI-enabled instruments.

For DMMs: Keysight/HP/Agilent, Rohde&Schwarz, Siglent, Rigol, Keithley, Tektronix, ...
For Scopes: Keysight/HP/Agilent, Rohde&Schwarz, Siglent, Rigol, Keithley, Tektronix, LeCroy, ...
For Signal Generators: Keysight/HP/Agilent, Rohde&Schwarz, Siglent, Rigol, Keithley, Tektronix,  ...

Yes - almost all of our instruments are controllable via SCPI, and many of them have remote GUI interfaces that only require a web browser or a standard VNC / RDP client. 

I've worked for test and measurement equipment manufacturers for over 25 years, and the trend has very definitely been to move away from "proprietary" solutions for controlling instruments remotely.  In the old days, working as a field applications engineer in T&M meant having dozens of programs installed on your laptop for all the products you supported.  There was nothing quite like being on site and finding out that you didn't have the right software package to talk to an instrument.

That said, there are a lot of both "professional" and "hobbyist" instruments that come with a proprietary remote control software package, but I agree strongly with the OP that these should not be the ONLY way to communicate with or use a modern test and measurement instrument.

[zrq]

Really liked the official built-in VNC server of many recent R&S gears  .
We are also using R&S SGMA series compact generators without front panel, which works nicely for us.

[JPortici]

Isn't a PC based embedded controller runing some Windows usually easier to service and replace than a customized digital board running some obscure long-dead RTOS? There isn't much one can do if such customized thing break for example because of storage card corruption.

At least one can always do the disk imaging in a relatively standard way and in case some software break, you know how to load the exe to IDA or GHIDRA. And it will be the better documented x86 rather than PPC or  some weird shit.

It then depends on what the custom application requires, and what the custom driver for the custom hardware card requires.
It maybe easier to reverse engineer but it still will cost a lot in term of time/money

[TomKatt]

With the complexity of modern gear, there is simply no avoiding one kind of obsolescence over another...  In 15 years any system will be difficult to locate software or components for operation or repair.

[alm]

I think oz2cpu's point is that the problem is still easier if all the necessary hardware and software is contained in a single box, rather than if you need to get the instrument, cable, computer, software, etc all from different sources, and inevitably separated from each other over time. Bonus points if the software has some annoying copy-protection like a hardware dongle or online activation. If the PC is embedded in the instrument, like modern high-end scopes, then at least it will stay with the instrument over its life. But if you don't even have the software, and the instrument is fully reliant on that software to do anything useful, then it's game over for that instrument.

For DMMs: Keysight/HP/Agilent, Rohde&Schwarz, Siglent, Rigol, Keithley, Tektronix, ...
For Scopes: Keysight/HP/Agilent, Rohde&Schwarz, Siglent, Rigol, Keithley, Tektronix, LeCroy, ...
For Signal Generators: Keysight/HP/Agilent, Rohde&Schwarz, Siglent, Rigol, Keithley, Tektronix,  ...

How many DMMs and signal generators do Tektronix, Rigol and Siglent make that rely on PC-control? How many scopes and signal generators does Keithley make at all?

list may be continued. All of them now offer remote-only instruments. The market is clearly automated testing
in a factory environment, and this market is huge. Here you need the max of test functionality in a cramped space. All this PXI stuff is just doing that.

Sure, this has been going on for decades. NI has been doing that or a long time, but their hardware is generally very reliant on their software. Can you talk to PXI devices without any instrument-specific driver, like a stand-alone scope that supports LAN and SCPI over TCP/IP? Is for example the R&S PXI line SCPI? The only thing I find in their documentation is how to use their drivers, which will no doubt stop working at some future Windows / LabView update.

[pdenisowski]

Is for example the R&S PXI line SCPI?

Well, we (R&S) don't really make PXI instruments   But with very rare (like, < 5%) exceptions, all of our remotely-controllable products are SCPI-compliant.

The only thing I find in their documentation is how to use their drivers, which will no doubt stop working at some future Windows / LabView update.

Remote control (SCPI / VISA) is extremely important to many of our customers, so I can pretty much guarantee uninterrupted support for our drivers, etc.   I've also personally covered a very wide range of R&S products (spec ans, sig gens, VNAs, scopes, power sensors, power supplies, specialty products like direction finders, avionics testers, etc.) for over 15 years and I can't remember a single time that remote control was "broken" for any of them.

As a matter of fact, many of our instruments can even work with SCPI (or even non-SCPI) code written for older, obsolete instruments, both from R&S and from our competitors (!).  And we don't charge for that

https://www.rohde-schwarz.com/us/solutions/test-and-measurement/aerospace-defense/r-s-legacypro/legacypro_254458.html

[jasonRF]

I think oz2cpu's point is that the problem is still easier if all the necessary hardware and software is contained in a single box, rather than if you need to get the instrument, cable, computer, software, etc all from different sources, and inevitably separated from each other over time. Bonus points if the software has some annoying copy-protection like a hardware dongle or online activation. If the PC is embedded in the instrument, like modern high-end scopes, then at least it will stay with the instrument over its life. But if you don't even have the software, and the instrument is fully reliant on that software to do anything useful, then it's game over for that instrument.

For DMMs: Keysight/HP/Agilent, Rohde&Schwarz, Siglent, Rigol, Keithley, Tektronix, ...
For Scopes: Keysight/HP/Agilent, Rohde&Schwarz, Siglent, Rigol, Keithley, Tektronix, LeCroy, ...
For Signal Generators: Keysight/HP/Agilent, Rohde&Schwarz, Siglent, Rigol, Keithley, Tektronix,  ...

How many DMMs and signal generators do Tektronix, Rigol and Siglent make that rely on PC-control? How many scopes and signal generators does Keithley make at all?

list may be continued. All of them now offer remote-only instruments. The market is clearly automated testing
in a factory environment, and this market is huge. Here you need the max of test functionality in a cramped space. All this PXI stuff is just doing that.

Sure, this has been going on for decades. NI has been doing that or a long time, but their hardware is generally very reliant on their software. Can you talk to PXI devices without any instrument-specific driver, like a stand-alone scope that supports LAN and SCPI over TCP/IP? Is for example the R&S PXI line SCPI? The only thing I find in their documentation is how to use their drivers, which will no doubt stop working at some future Windows / LabView update.

Yup.  We built up a system based on a PXI chassis for a customer that had equipment from NI, Agilent, etc. in it.  Whether or not it would still work in 20 years was of no consideration.   Portability of a system designed for a specific set of automated measurements was the issue.  It was mounted in a portable, shock-mounted mini-rack, and the PXI form-factor PC ran some version of windows and Labview.  It was shipped a bunch of different places over the course of a few years as a single unit and just worked every time with just a few minutes of setup, no complicated cabling to reproduce every time, etc. 

I don't understand why folks here cannot wrap their heads around the fact that there are a gazillion different applications that can have very different priorities.  None of us could possibly know all the different ways all of these instruments are used.  Insisting that it is always a bad idea to buy a specific kind of equipment shows a kind of arrogance that is laughable is silly. 

jason

[Wolfgang]

I think oz2cpu's point is that the problem is still easier if all the necessary hardware and software is contained in a single box, rather than if you need to get the instrument, cable, computer, software, etc all from different sources, and inevitably separated from each other over time. Bonus points if the software has some annoying copy-protection like a hardware dongle or online activation. If the PC is embedded in the instrument, like modern high-end scopes, then at least it will stay with the instrument over its life. But if you don't even have the software, and the instrument is fully reliant on that software to do anything useful, then it's game over for that instrument.

For DMMs: Keysight/HP/Agilent, Rohde&Schwarz, Siglent, Rigol, Keithley, Tektronix, ...
For Scopes: Keysight/HP/Agilent, Rohde&Schwarz, Siglent, Rigol, Keithley, Tektronix, LeCroy, ...
For Signal Generators: Keysight/HP/Agilent, Rohde&Schwarz, Siglent, Rigol, Keithley, Tektronix,  ...

How many DMMs and signal generators do Tektronix, Rigol and Siglent make that rely on PC-control? How many scopes and signal generators does Keithley make at all?

list may be continued. All of them now offer remote-only instruments. The market is clearly automated testing
in a factory environment, and this market is huge. Here you need the max of test functionality in a cramped space. All this PXI stuff is just doing that.

Sure, this has been going on for decades. NI has been doing that or a long time, but their hardware is generally very reliant on their software. Can you talk to PXI devices without any instrument-specific driver, like a stand-alone scope that supports LAN and SCPI over TCP/IP? Is for example the R&S PXI line SCPI? The only thing I find in their documentation is how to use their drivers, which will no doubt stop working at some future Windows / LabView update.

After some research:
Rigol does have PC-only signal generators, oscilloscopes
Siglent has PC-only oscilloscopes
Tektronix has PC-only oscilloscopes, spectrum analyzers
Keithley makes PC-only SMUs and multimeters

Tek/Keithley are now the same company.

 I think that the demand for remote-only is huge. PXI is fine as long as standard interfaces are used, otherwise you are again hooked to a specific manufacturer.

[alm]

I think that the demand for remote-only is huge. PXI is fine as long as standard interfaces are used, otherwise you are again hooked to a specific manufacturer.

No argument the demand is huge for all kinds of automated setups. But it is a niche market. For bench use, a more traditional unit with a front panel and remote control is more versatile for the majority of people.

Is PXI like PCI in that you need a driver for every instrument? Or is it like GPIB where you only need a driver for the interface and the rest is standard SCPI?

[Wolfgang]

I think that the demand for remote-only is huge. PXI is fine as long as standard interfaces are used, otherwise you are again hooked to a specific manufacturer.

No argument the demand is huge for all kinds of automated setups. But it is a niche market. For bench use, a more traditional unit with a front panel and remote control is more versatile for the majority of people.

Is PXI like PCI in that you need a driver for every instrument? Or is it like GPIB where you only need a driver for the interface and the rest is standard SCPI?

Regarding PXI I suppose that there are good (SCPI, ...) and bad (proprietary protocol) guys. I dont have any PXI stuff so I cannot tell from personal experience. In my lab everything from Keysight, R&S, Rigol, ... has a user interface (expect a PicoScope) plus SCPI programmability, but the instruments for automation I built myself, still using SCPI interfaces.

[zrq]

If so many people don't like PC controlled test equipment, please sell them at once one get hands on them, on ebay or this forum. They are going to loose value quickly, maybe after a Windoze update it will break, so better put a lower asking price to get rid of them quickly.

So that I can pick them up  .

[switchabl]

Is PXI like PCI in that you need a driver for every instrument? Or is it like GPIB where you only need a driver for the interface and the rest is standard SCPI?

PXI is essentially just PCI with a different form factor and some extra signals (trigger and so on). Some PXI cards expose a SCPI interface through VISA but you still need a vendor-specific driver. PXI also supports DMA transfers for high-throughput applications and there isn't really a standard interface for that (maybe IIO on Linux) so that will always be proprietary.