Tuesday, May 31, 2011

All Work and Little Play....

Well, we've made good progress getting some systems running again. The Launch Control Center is now operable, and I was able to get Management to dump some of the old, antiquated software we were using, like PCAnywhere, in favor of a new program I use here at home called TightVNC. We use it for remotely connecting to, and controlling, a bunch of servers that provide all the display screens in "Mission Control".
Quite frankly, PCAnywhere is a PIG for resources, and extremely slooooow. It would take about 40 seconds from when you hit the "connect" button until the remote desktop appeared. With TightVNC it happens so fast you barely have time to count "One....Two..." and BAM, you're connected and displaying the remote desktop.
On top of that, PCAnywhere is $200 per pair of PC's you want to use it with, and TightVNC is free, as it's licensed under the GPL, like Linux. It's faster, can be more secure, and very simple to set up and use.
People's jaws dropped when I did the demo, as they were used to starting PCAnywhere, and then going for a cup of coffee.
Considering we use this type of software to control several dozen computers, I think I just earned my salary for the month.
We still have to figure out the clock system, as it's been changed several times without proper documentation, and nobody knows who did what. It's a robust system using several GPS units that spit out a time code to a "server" box (basically a NMEA-0183-to-RS-485 converter), and it's one of those things that "just runs", but if it ever broke, we'd be lacking a paddle to get back down sh1t creek without proper documentation!
Today we started on one of my 'specialties', the Weather Radar System. When we first powered it up and tried to command the antenna, things went nuts, and the PC we use to control it shut everything down to prevent damage. If you've ever seen a fully-steerable 3 Meter dish go bonkers, you know it's not pretty. We found corroded slip rings, and loose connections to the servos, synchros, and tachs. When we first powered it up, it took about 15 minutes to run "fsck", as the last person to shut it down just pulled the plug, NOT something you do to a UNIX Operating System unless it can't be avoided.
Tomorrow I'll look through the logs to see when the last time was that it was powered up and exercised, and that will probably explain why the loose connections weren't noted in the written log ("Antenna won't sync" or something similar), and the lack of operation would also explain why the slip rings got all crummy from just sitting.
After that we'll check the installed TWTA, and swap in the spare for testing.
Compared to all the PC bashing I've done lately, doing some "Real RF" work will be fun!

10 comments:

  1. I've used TightVNC to control a remote PC that has some U451 USB relays on it, running a local program. Now, though, the program can run on my machine and communicate with the remote relays. But TightVNC worked great for that, and it was free.

    ReplyDelete
  2. Yeah, it's a great piece of software; lightweight and fast!

    ReplyDelete
  3. Sounds like you're having FUN! :-) And getting to do something that is actually productive!

    ReplyDelete
  4. I guess "productive" can be viewed (sometimes) as a long-term thing!
    We (me, actually) found a problem that's been bugging us ever since the radar was installed. Once in a while the antenna controller would squawk about "Elevation Tach Inconsistent" or "Synchros Out-of-Range", or any of several other related antenna control issues. Well, today I suspected perhaps *all* the connectors needed to be cleaned and reseated, and I got a big surprise when I pulled off the 25-pin connector that goes from the rack up to the pedestal.
    One of the pins was about half way back in the shell!
    It's obviously been like this since it was installed, as *nobody* has ever pulled that connector and checked it. And the pin is the one that carries the AC Return for the servo system, meaning it would lose its ground connection occasionally, potentially causing all manner of weirdness in the servo system.
    The pin would go into the shell, and 'kinda-sorta' snap into place, but was very easy to push back out.
    Tomorrow we'll be replacing the connector with one that has permanently install pins in the shell. This connector was one that used crimp-on pins that have to be inserted into the shell, and it looks like the vendor messed up when they put that pin into place, damaging the plastic retention mechanism inside the shell, and causing the pin to be loose.
    AFAIC, that type of connector should NEVER be used in a high vibration environment, as the possibility of the pins coming loose is just too great.
    If this fixes the phantom servo problems we've been having, I'm going to insist we replace ALL of the crimp-pin style connectors with solder cup type.
    They never should have been allowed into the system in the first place......

    ReplyDelete
  5. Ah... it's amazing what happens when one 'knows' the history of the system...LOL And yeah, those canon plugs SUCK!

    Try using them on an airplane... sigh... BTDT and damn near went nuts!

    ReplyDelete
  6. These aren't Canon plugs....they're standard DE-25 "computer" connectors. The vendor never should have used them in a shipboard environment, but too late to change them.
    Got them replaced today (ALL day!) and we're still having problems. Looks like it might be the I/O card in the computer driving the antenna. Disconnect the one cable from the panel to the I/O card, and everything is happy. Hook it back up and Weird Things start happening. Just glad we have a spare, brand-new, still in the anti-static bag with unbroken seals from 2005!

    ReplyDelete
  7. Thanks for not calling them DB-25.

    Grounding issue?

    ReplyDelete
  8. Yeah, it bugs me, too, when people call them a "DB-25"! I knew one guy who always called his Ethernet cards "NIC cards". He didn't 'get it' when I asked him what brand of "Network Interface Card Cards" he was buying!
    I thought "lost ground" when we found the loose pin. The loose pin was the AC return/ground for the 110VAC power to the servos, and I thought perhaps if it was losing its ground, it might explain all the trouble we're having getting it to communicate with the pedestal electronics. Now it looks like it's the I/O card in the PC, as when we disconnect the cable to the PC, everything seems OK. Hook up the cable, and funny stuff starts to happen.
    Running the 110/120 VAC servo power *in the same cable* as the control signals is another BIG no-no. If I would have been involved more in the design and installation process I would have really squawked about doing that, but I was assigned to a different group at the time.
    I'd recommend pulling a separate cable for it, but I'm sure I'd get shot down. The new owners are really watching their nickles and dimes until we get at least one bird in orbit, and get paid for it.

    ReplyDelete
  9. "Running the 110/120 VAC servo power *in the same cable* as the control signals"

    Ugh! Are the control signals at least differential? RS422/RS485 or the like?

    ReplyDelete
  10. No, they're single-ended signals, another boo-boo, as the I/O card supplied in the PC can be configured to use differential signals.
    This was the first shipboard radar the vendor built, and it's becoming obvious that they didn't understand certain things!
    What's amazing is that it worked extremely well until we shut it down 2 years ago.
    We found some more corroded connections this morning and repaired them, but we still can't get the servo system to work.

    ReplyDelete

Keep it civil, please....