Run your test instruments on Windows? No thanks. / General Chatting / Forums

Forums

4hv.org :: Forums :: General Chatting

« Previous topic | Next topic »

Run your test instruments on Windows? No thanks.

1 2 3 next

Move Thread

LAN_403

2Spoons

Tue Nov 08 2011, 01:52AM

Registered Member #2939 Joined: Fri Jun 25 2010, 04:25AM
Location:
Posts: 615

Whoever thought using windows as an OS to run an oscilloscope should be shot. Repeatedly.
Our fancy Tektronix DPO 'scope (worth about $20k) had a virus. Got that cleaned out. (as a work mate commented "USB sticks are just like genitals" - the ensuing conversation was hilarious, and unrepeatable.)
Then the 'scope application decides to crap out, and wont run. So I've just wasted half a day reinstalling everything from the backup partition. Now its all working again, except the factory calibration file is now missing so its running on defaults. ARRGH!

On a similar note we have a Rode and Schwartz RF spectrum analyser, running windows. It likes to get stuck in an endless reboot cycle (a heat issue I believe).
And it takes 5 minutes to boot. My old Tek 492 (circa 1980) was always up and running in 20 seconds - the time it takes for a crt filament to reach operating temp.

End of rant.

Patrick

Tue Nov 08 2011, 02:28AM

Registered Member #2431 Joined: Tue Oct 13 2009, 09:47PM
Location: Chico, CA. USA
Posts: 5639

2Spoons wrote ...

Whoever thought using windows as an OS to run an oscilloscope should be shot. Repeatedly.

I wholly agree, windows (all of its variants) are evil. Everyone at Micro$oft should be prosecuted for crimes against humanity.

To the Hague !!!

Mattski

Tue Nov 08 2011, 08:35AM

Registered Member #1792 Joined: Fri Oct 31 2008, 08:12PM
Location: University of California
Posts: 527

Linux would be better yes... but I've seen well-made instruments running Windows and crappy instruments running windows. In my limited experience when instruments first started using Windows they had bad UI's and longer boot times, but more recent ones work pretty well. What I hate is when they assume you're using a keyboard or mouse instead of the front panel, so controlling it from the front panel ends up being incredibly awkward. I've used instruments running a custom (or at least not readily apparent) OS which took forever to boot up, 5 minutes or so. Some fairly modern instruments I've used run WinXP and boot up in 10-20 seconds.

General rules to keep Windows instruments happy: Turn auto-run off if it isn't already off, don't let anybody stupid within 10 feet of it (applies to all instruments, not just Windows ones), don't give it a publicly routable IP address, and if you want to be even better firewall it off from any PC's except the ones which need to communicate with it. If I remember correctly one Windows instrument I use came with Symantec antivirus, though you shouldn't install an antivirus program without talking to the manufacturer generall.

Steve Conner

Tue Nov 08 2011, 09:52AM

Registered Member #30 Joined: Fri Feb 03 2006, 10:52AM
Location: Glasgow, Scotland
Posts: 6706

I've designed a couple of instruments that run Windows CE. Not had any complaints about viruses yet. Indeed the hardest bit is getting it all to run from a custom front panel instead of a keyboard and mouse. To make the "File Open" dialog usable I had to do some truly disgraceful Win32 hacks.

HP's old high end instruments ran HP Unix. I'm not sure what Tektronix used, but I bet it was something similar.

Jrz126

Tue Nov 08 2011, 03:14PM

Registered Member #242 Joined: Thu Feb 23 2006, 11:37PM
Location: Erie PA
Posts: 210

There's quite a few benefits to running windows. keyboard/mouse easier to navigate than buttons for changing settings, Saving files, etc.
Also being able to remote desktop is a great feature. I have a logic analyzer set up in the opposite corner of my cube. No need to turn around to look at signals on the tiny 10" screen.

2Spoons

Wed Nov 09 2011, 01:39AM

Registered Member #2939 Joined: Fri Jun 25 2010, 04:25AM
Location:
Posts: 615

There's nothing that Windows does for a test instrument that can't be done with proprietary code on proprietary hardware (probably with better speed and stability). Using a desktop OS is just sheer laziness. Windows was never designed to run real-time hardware.

Maybe I'm just getting old and cranky (at 43, lol) but I like having a front panel full of knobs and buttons. I could walk up to any 'scope like that and be using it in seconds, not waste half a day clicking through menus trying to find the setting I want.

Besides - knobs and buttons just look cool.

Carbon_Rod

Wed Nov 09 2011, 03:00AM

Registered Member #65 Joined: Thu Feb 09 2006, 06:43AM
Location:
Posts: 1155

Unfortunately, people still owe Microsoft a licensing fee if they plan to provide long file name support on FAT formatted USB drives. More than one Linux based manufacturer has been hit by this old licensing annoyance.

Linux running off romfs with direct hardware DSP interfaced kernel modules is more common than most people would guess. These special kernels even have a special diagnostics patch to map around faulty areas of RAM.

However, it is still quite common to find Windows CE, VXWorks, or freeRTOS on consumer equipment.

And for $20K... one would expect the unit would have a self-calibration routine like the $350 "overclocked" Rigol DS1052E.

Steve Conner

Wed Nov 09 2011, 03:54PM

Registered Member #30 Joined: Fri Feb 03 2006, 10:52AM
Location: Glasgow, Scotland
Posts: 6706

2Spoons wrote ...

There's nothing that Windows does for a test instrument that can't be done with proprietary code on proprietary hardware

Well, except get the thing to market quicker with the ability to save results to a FAT formatted USB flash drive and upload them to a Windows PC.

For one of our instruments, I wrote a kernel mode driver to connect a DSP to the main processor, an ARM running WinCE. It can transfer data at about 5-10MByte/sec between the two chips, and since it goes through the DSP's DMA controller, it can be working on something else in the meantime. The DSP does all the nasty real-time stuff (and very nicely too) while CE handles the user interface and results display. I was worried about refresh rate, but it can happily draw a scope trace at 50fps using the standard Win32 GDI routines. (another reason to buy an OS- if rolling your own you'd have to rewrite all the graphics routines)

I am still rather fond of our lab's Tek 2465B. But knobs and buttons cost more. (We use those nasty membrane keypads to reduce assembly labour and keep the instrument water resistant.)

2Spoons

Thu Nov 10 2011, 01:45AM

Registered Member #2939 Joined: Fri Jun 25 2010, 04:25AM
Location:
Posts: 615

I found the self-cal routine for the 'scope. Then bloody Windows Update killed the scope software again. Turns out the anti-virus turns on auto update without permission. So here I am again, reinstalling the software. And auto update IS disabled this time. At least I'm getting paid to do this.

Patrick

Thu Nov 10 2011, 04:14AM

Registered Member #2431 Joined: Tue Oct 13 2009, 09:47PM
Location: Chico, CA. USA
Posts: 5639

2Spoons wrote ...

... At least I'm getting paid to do this.

Im not.

The US federal government, Sally Mae, and my parents are paying while i have screw around wasting my time with bad instrumentation on campus.

And in my student evaluations, which we are required to fill out, i said so.

1 2 3 next

Moderator(s): Chris Russell, Noelle, Alex, Tesladownunder, Dave Marshall, Dave Billington, Bjørn, Steve Conner, Wolfram, Kizmo, Mads Barnkob