Windows 10 switch

[ydna1977]

Hi ,
Anybody prepared or already switched to windows 10 :
•Do I have to upgrade?
•What are the benefits of upgrading now
•Is there an easy process I can follow

[Steve Reimer]

What are you switching?

[ydna1977]

Windows 7 support ends in January 2020 so update to windows 10
Anybody thinking of this ?

[bunnahabhain]

All office PCs with connection to the internet must be upgraded for security reasons, of course. Acquisition workstations can only be upgraded if the instrument control software will work under Windows 10. Ask your software supplier for Windows 10 support. If it is not available, separate your workstation from the internet and continue your work as usual.
In my eyes there is no need to upgrade a functional acquisition workstation just to be able to use internet.

[James_Ball]

All office PCs with connection to the internet must be upgraded for security reasons, of course. Acquisition workstations can only be upgraded if the instrument control software will work under Windows 10. Ask your software supplier for Windows 10 support. If it is not available, separate your workstation from the internet and continue your work as usual.
In my eyes there is no need to upgrade a functional acquisition workstation just to be able to use internet.

I agree, except my instrument acquisition workstation is also my "office" PC where I get the company email which is now "in the cloud" so it requires internet connection. Things would be so much easier if IT could go to the trouble of keeping email in house instead of migrating to the cloud.

[DR]

...my instrument acquisition workstation is also my "office" PC where I get the company email which is now "in the cloud"...

Does your IT department like malware and viruses?
I ask because that's a great way to get malware and viruses.

Most acquisition PCs require lots of open ports and most data systems do not work well in the presence of anti-malware programs. This results in port vulnerabilities and no effective means of scanning several folders on a given hard drive (or you close the ports and scan the CDS folders and get no data from your instruments).

You should really get another PC for your email and whatnot while leaving the instrument PCs behind a very restrictive firewall/proxy server or just make them part of a subnet with no direct outside access.

[James_Ball]

...my instrument acquisition workstation is also my "office" PC where I get the company email which is now "in the cloud"...

Does your IT department like malware and viruses?
I ask because that's a great way to get malware and viruses.

Most acquisition PCs require lots of open ports and most data systems do not work well in the presence of anti-malware programs. This results in port vulnerabilities and no effective means of scanning several folders on a given hard drive (or you close the ports and scan the CDS folders and get no data from your instruments).

You should really get another PC for your email and whatnot while leaving the instrument PCs behind a very restrictive firewall/proxy server or just make them part of a subnet with no direct outside access.

Most of the acquisition PCs run dual NICs one for the instrument and one for the house network. The internet connections are filtered through the main server as the firewall/proxy which is scanning and recording all inbound and outbound traffic. Certain individuals do not have access to internet when they log into a PC just in case.

In the past it was like pulling teeth just to get a PC that was a little faster to run the instrument, and we usually replaced the instrument PCs with the hand-me-downs from the front office(which makes no sense since a GC/MS or LC/MS/MS will use much more processing power than spreadsheets and email).

But things are changing around here and we now have someone who realizes the importance of the instruments and their needs, but small facilities will often have to make due with what is on hand.

Luckily I am finding that most of the Agilent software will run on W10 even though it isn't certified/supported on it. Though a few legacy instruments need tricks to get them to accept an IP address when they just do not communicate with the BootP as a service.

[benhutcherson]

I generally stick with "offline only" for the instrument computers under my control.

I have ONE GC computer where I can connect to WiFi on-demand, but I never leave the connection open. That one is running Windows 10, and it's controlling an SRI GC. Unlike other companies, SRI actually makes their acquisition software freely available on their website, and have a fully validated and ready to go version for Windows 10.

Not too long ago, I "fixed" a 7820/5975 that's not under my direct supervision(it belongs to a research group) but that I get asked to work on as needed(i.e. they break something on it). On this particular occasion, the GC and MS weren't communicating with the computer, and as soon as I looked at it, I knew why-they had attempted to connect it to the internet by running a wire from the ethernet hub that controls the instrument to a wall jack, and although the internet worked the school network had messed with the IP addresses on everything. After I disconnected it and set everything back to default, it was fine.

Interestingly enough, though, I was wearing a different hat today servicing an NMR. One of ours was being stubborn about tuning, and our NMR guy is out of town this week. Those systems all run RHEL, and are internet connected(I think they have dual NICs-one for instrument interface and one for the computer). I actually came to appreciate that as I was reading emailed instructions from the NMR manager and acting on them from the same computer rather than pulling up the email on my phone. I'm not sure whether or not RHEL does Flash(or if these computers have it installed) so about all the internet access on them gets used for is to email results, and probably more often to access the online scheduling system. Since vNMRj is now open source, presumably it won't be a security risk to keep updating both the OS and the software.

The instruments under my supervision are a nice anthology of older versions of Windows. Most of my stuff is on Windows 7, aside from the SRIs on Windows 10, but I have a decent number on XP still, a few on 2000(including the 5890 FID/NPD and a Finnigan LC-MS that's a bit of a work in progress), the 5890/5971 is on 98, and until a couple of things gave up the ghost at the same time I even had a Nicolet FTIR/Raman on Windows 3.11(I'm trying to get the money to both fix the instrument and buy a newer version of the software that will run on Windows 2000). I also can't forget the home-built NMR that runs on Mac OS 9-I provide tech support on that computer since I have relatively "fresh" OS 9 knowledge and also can support the hardware. Offline definitely rules the roost, though.

[DR]

Most of the acquisition PCs run dual NICs one for the instrument and one for the house network. The internet connections are filtered through the main server as the firewall/proxy which is scanning and recording all inbound and outbound traffic. Certain individuals do not have access to internet when they log into a PC just in case.

In the past it was like pulling teeth just to get a PC that was a little faster to run the instrument, and we usually replaced the instrument PCs with the hand-me-downs from the front office(which makes no sense since a GC/MS or LC/MS/MS will use much more processing power than spreadsheets and email).

But things are changing around here and we now have someone who realizes the importance of the instruments and their needs, but small facilities will often have to make due with what is on hand.

Luckily I am finding that most of the Agilent software will run on W10 even though it isn't certified/supported on it. Though a few legacy instruments need tricks to get them to accept an IP address when they just do not communicate with the BootP as a service.

While it is a good idea to limit access on instrument PCs, there are still several open ports that are required in order for the acquisition PC to communicate with the server (assuming a C/S environment). This is where the risk comes in. Also, all you need is a badly hung Outlook process to force a reboot and take out several active runs to get things going again. As cheap as PCs are these days, they really should keep a "hands off" policy on instrument PCs.

@ benhutcherson - 3.11 and 2000 machines? You have my condolences. I rather liked 3.11 but was a much bigger fan of 98, XP and 7. I get that 10 works, but I think it is change for change's sake and they always manage to take away things that I used whenever there's a big upgrade. Credit for adding in a virtual drive mounting utility for people who d/l an occasional ISO, but I miss things like photomanager and integrated LAME.

[James_Ball]

Most of the acquisition PCs run dual NICs one for the instrument and one for the house network. The internet connections are filtered through the main server as the firewall/proxy which is scanning and recording all inbound and outbound traffic. Certain individuals do not have access to internet when they log into a PC just in case.

In the past it was like pulling teeth just to get a PC that was a little faster to run the instrument, and we usually replaced the instrument PCs with the hand-me-downs from the front office(which makes no sense since a GC/MS or LC/MS/MS will use much more processing power than spreadsheets and email).

But things are changing around here and we now have someone who realizes the importance of the instruments and their needs, but small facilities will often have to make due with what is on hand.

Luckily I am finding that most of the Agilent software will run on W10 even though it isn't certified/supported on it. Though a few legacy instruments need tricks to get them to accept an IP address when they just do not communicate with the BootP as a service.

While it is a good idea to limit access on instrument PCs, there are still several open ports that are required in order for the acquisition PC to communicate with the server (assuming a C/S environment). This is where the risk comes in. Also, all you need is a badly hung Outlook process to force a reboot and take out several active runs to get things going again. As cheap as PCs are these days, they really should keep a "hands off" policy on instrument PCs.

@ benhutcherson - 3.11 and 2000 machines? You have my condolences. I rather liked 3.11 but was a much bigger fan of 98, XP and 7. I get that 10 works, but I think it is change for change's sake and they always manage to take away things that I used whenever there's a big upgrade. Credit for adding in a virtual drive mounting utility for people who d/l an occasional ISO, but I miss things like photomanager and integrated LAME.

I have to say my best running instrument control PC ever was MSDChemstation running on NT4. I had a few with battery backups that ran over a year between reboots with no slow down or problems at all. Of course there was no USB support, but those were just rock steady machines. W7 running MSDChemstation experiences a memory leak that wasn't there in XP, as about once a week I get the error that "this computer is running out of memory" every time a new run in a sequence starts. Have to reboot to get rid of the error. We are getting ready to upgrade one of our last XP systems to W10, so wish us luck that everything will work, if the thing wasn't causing files to lose ownership and lock us out of them I would not even be thinking of upgrading, but it seems to be a problem with the OS and I really don't have a disk handy to reload it.

[LALman]

I have a very small lab. I have my own router behind the providers router. I run Win7 Pro in the front office on a PC that has two NIC's one for internet, one for internal net. On the internal net, I have a few gig switches to route the traffic. I have two Win98 PC's running 5890/5971 and 5890 FID. Then I have one PC running WinNT 4.0 (ICP-MS) and three instruments controlled by WinXP Pro (Chromeleon ICS-2000, 6890 GC-FID, 6890N/5973inert), and an OI Analytical TIC/TOC analyzer running WinXX from a sim card. The 5973 forces me to have static IP on the internal net.

Win3.11 runs fine in OS/2 if you dare. I think it might also run well in W2KPro but not so well in NT4. The problem for me is not really OS but that the 5890 series GPIB cards run on an ISA bus. Everything runs fine except that Win98 needs to be rebooted after I tune and also if I generate reports. So, I copy the 5890/5971 data over to the 6890/5973 system and process it with that much faster software and NIST library.

[LALman]

James Ball, I agree with you about NT4. That was the most stable OS ever for its time. It even seduced me to abandon OS/2 when IBM abandoned it.

I find that the refurbished HP 6005 Pro SFF and HP 4000 Pro SFF were about $150 refurbished with 8GB of DDR, 500 GB drive, and Win7Pro. I bought a couple of each for backups when my first HP4000 that came with my 5973 instrument bit the dust. I copy the drive to a 256GB SSD and that makes for a fine machine for Chemstation D.XX.XX The HP4000 is a little slow for Chemstation E.XX.XX. The HP 4000 has an intel Pentium E6700 at 3.2GHz and the HP 6005 has an Athlon II X2 at 3.2GHz.

[James_Ball]

James Ball, I agree with you about NT4. That was the most stable OS ever for its time. It even seduced me to abandon OS/2 when IBM abandoned it.

I find that the refurbished HP 6005 Pro SFF and HP 4000 Pro SFF were about $150 refurbished with 8GB of DDR, 500 GB drive, and Win7Pro. I bought a couple of each for backups when my first HP4000 that came with my 5973 instrument bit the dust. I copy the drive to a 256GB SSD and that makes for a fine machine for Chemstation D.XX.XX The HP4000 is a little slow for Chemstation E.XX.XX. The HP 4000 has an intel Pentium E6700 at 3.2GHz and the HP 6005 has an Athlon II X2 at 3.2GHz.

I missed out on OS/2, I went from a home computer that was an Atari800xl to DOS/Windows when I started working in the lab. Well there was the RTE-A on the HP1000 with two scan boxes for the 5995 and 5970 and 8 dumb terminals to do the data workup and print out reports. It was so fun to need two hours just to quant and print the reports from the overnight run on those, but backing up the data on a file cabinet sized reel to reel tape drive made it feel like I was working at NASA We had three of those setups in the lab when I started in 1991, and one W3.11 5971 that nobody would touch because they had no clue what Windows was lol. I learned that one while I was working night shift, reading manuals and making methods from scratch and losing them because I didn't save at the right time. Oh how far we have come in 30 years!

[LALman]

About 1992, I had my first Environmental Lab job. I was running two dual column (DB-1701 & DB-5 IIRC) Perkin Elmer GC-ECD for pesticides and PCB's. I had a single PC running OS/2 and in OS/2 I was running two copies of Win3.11 for Workgroups one for each GC-ECD setup. So I was acquiring 60 cps from 4 detectors on two GC-ECD instruments. OS/2 was so smooth that I could format a floppy disk or QC a data set while acquiring without dropping any data. I think the software was Chromperfect Direct and I had two acquisition cards in the PC.

I tried running two instances in Win3.11 but that crashed over and over. It was a superb demonstration of multitasking and I was pretty stoked that it worked. It was nuts to see Win3.11 get top reviews in "multitasking" benchmarks compared to OS/2.

[James_Ball]

Best OS I ever used for multitasking was BeOS. But it came out about the same time as Windows95 and versus the hype and ad money of MS and Intel combined it just couldn't compete. I saw it first when Leo Leporte did a demo of it on ZDTV, the demo had 6 videos displayed one on each side of a rotating cube, and it didn't even slow down the system. Windows was no where close to being able to do that back then.

I would be happy now to be able to run the Unix version of Enviroquant, at least then the instrument would get almost all of the processing power instead of most of it going to support Windows.