Rick had some insulation left over from his own workshop project (like two bundles... his workshop is the size of Fragwell Manor, swear to God) so he sold it to me for about half price. Turns out it's enough to cover like 80% of my shop.
At the moment, I'm over halfway done and destroying the big shelf. I'll use the wood for other things, but it's in the way now.
A few hours later and I'm done with the insulation now. The process was nowhere near as annoying as I had thought it would be, thanks to advances in fiberglass technology since the 1980s, I'm sure.
Equipped with my Graywhale CD Exchange hoodie I bought in 1999 but only wore for the first time this fall while replacing a hub module on my car, a filter mask and (occasionally) safety goggles, I managed to get the job done and come away from it without being all itchy.
By my calculations, I have over 300 sq ft of wall covered now and only another 120 sq ft left uninsulated. The rest will be done with foam sheets, I think.
In the process of putting things back in order without the shelving I used to have, I decided to change my target layout. Instead of my desk being at the back of the shed with my back to the door while working, it will be against the south wall so I don't have the sun bouncing off my displays during summer evenings.
Saturday, December 27, 2008
Saturday, December 13, 2008
Destructively Productive
I'm out in the lab tonight, experimenting with the new thermal properties the recent snowfall has given the lab. I maxed out at 55F a day ago when I checked the thermometer inside the lab after work. This was with an SGI Octane, two Indigo2s, an O2 in idle mode in IRIX and an Athlon XP system running Folding@Home under Windows XP. With the snow, I saw 46F.
So what did I do? I started up my loud IBM eServers and am now working on a Pentium 4 system that will serve as another folding node.
After giving my wife four 512MB RAM sticks for her system to replace the quartet of 256MB sticks she formerly had in there, I was left with only 512MB per system. I can't really render on half a gigabyte, but I'm not really going to be doing any rendering for like a month, maybe longer. Regardless, I got tired of hearing her computer scream in pain during her Photoshop escapades and it would seem that doubling her RAM to 2 Gigs has helped a lot. It's worth the effort.
The Pentium 4 system I'm working on right now is kind of retarded. It has bad thermal compound that flakes off at low temperatures and I'm not sure where I put my syringe of Arctic Silver compound. I need to find it because I can't even get the OS to install at the moment. O wait... found it. Hopefully that will fix it.
I have been in a bit of a quandry regarding my network. It would seem that I don't have a way to run wire from the house to the lab until the ground thaws so I need a way to connect until that happens. What better solution than WiFi? That'll work and it just so happens that I have a Linksys WRT54G router that I'm not using. Some folks don't know this, but you can have more than one router on a network. You just plug your second one into a computer port on the first one. In this case, the main one is WiFi B while the WRT54G is WiFi G and very well-documented. Perfect for me since my wife only really connects via WiFi B on her laptop and connects her desktop via copper and I tend to be curious about new configurations and don't want to interfere with her work.
While I have a router that is second to none in terms of capabilities, I didn't think I had near enough adapters for the systems I connect to it until I remembered the four USB adapters I salvaged from trashed home control boxes. They're kind of sketchy and the drivers suck, but they're a way to connect and that's all I need.
the new Pentium 4 system is having issues. It appears it has page faults where none are expected. The RAM tests out just fine so that leaves me hanging for now. I will look into it further. I have more CPUs so if that's the issue, I can solve it. Right now, though, I would rather work on gettin my former desktop system running again since I moved it out to the lab and have yet to do so.
It's all in the name of getting the lab straightened out so I can start work on the 42" plasma display I picked up broken for $10 and have my own hunches regarding fixes. I'm just glad I did my dumpster diving (minus the dumpster) at work and got a nice piece of anti-static mat that's protecting the screen right now.
I've swapped the 3GHz Pentium 4 out of the system in favor of a 2GHz model. I'm going to see real soon whether that was the problem. So far, so good. I'd rather have a bad CPU than a bad motherboard as I have so few motherboards in working form.
Next things to do:
Uninstall the video cards from the IBM servers for later installation in the main frontend system.
Uninstall Pentium 4 componenets from front-end in favor of AMD Sempron-based parts. (Reason: 2GB RAM vs 1GB and ATX vs uATX motherboard for more slots to accommodate the extra video cards)
I'm installing extra video cards in the frontend for the simple reason that I need them for my monitors. I have an ATI X800 AGP card in there right now and it's okay for one display, but I'm not about to run two off of it. I could use other video adapters, but I want the Matrox boards from the eServers because they have DVI ports and I'm not about to run analog signals to S-IPS LCDs ever again. For the record, I have a Dell 2001FP 20" panel as my center display and two Dell 1800FP panels on either side. Only one of the 1800FP panels happens to work, but I have a total of three non-working ones to mix and match from, or completely rebuild since it's generally a power supply fault anyway. I just want one perfect display when I'm done. No pressure spots or visible scratches. Odds are I have one. I just need to get it working. Based on simple power-up testing I just did, all of them have bad power supplies. Giggity. Easy fix and not too expensive.
The Pentium 4 system has managed to get into the Windows XP install screen in actual graphic mode. I reinstalled the case-back and the cooling fan.
And it's working. Proteins are folding and today is a happy day.
I just reached for something on a shelf near the ceiling and my hand became very warm. Looks like all my heat is heading toward the top, as predicted by that whole hot air rising thing. Time for a fan to move all that air.
Got it. If I'm going to freeze my ass off, I'm going to freeze my face off, too.
So what did I do? I started up my loud IBM eServers and am now working on a Pentium 4 system that will serve as another folding node.
After giving my wife four 512MB RAM sticks for her system to replace the quartet of 256MB sticks she formerly had in there, I was left with only 512MB per system. I can't really render on half a gigabyte, but I'm not really going to be doing any rendering for like a month, maybe longer. Regardless, I got tired of hearing her computer scream in pain during her Photoshop escapades and it would seem that doubling her RAM to 2 Gigs has helped a lot. It's worth the effort.
The Pentium 4 system I'm working on right now is kind of retarded. It has bad thermal compound that flakes off at low temperatures and I'm not sure where I put my syringe of Arctic Silver compound. I need to find it because I can't even get the OS to install at the moment. O wait... found it. Hopefully that will fix it.
I have been in a bit of a quandry regarding my network. It would seem that I don't have a way to run wire from the house to the lab until the ground thaws so I need a way to connect until that happens. What better solution than WiFi? That'll work and it just so happens that I have a Linksys WRT54G router that I'm not using. Some folks don't know this, but you can have more than one router on a network. You just plug your second one into a computer port on the first one. In this case, the main one is WiFi B while the WRT54G is WiFi G and very well-documented. Perfect for me since my wife only really connects via WiFi B on her laptop and connects her desktop via copper and I tend to be curious about new configurations and don't want to interfere with her work.
While I have a router that is second to none in terms of capabilities, I didn't think I had near enough adapters for the systems I connect to it until I remembered the four USB adapters I salvaged from trashed home control boxes. They're kind of sketchy and the drivers suck, but they're a way to connect and that's all I need.
the new Pentium 4 system is having issues. It appears it has page faults where none are expected. The RAM tests out just fine so that leaves me hanging for now. I will look into it further. I have more CPUs so if that's the issue, I can solve it. Right now, though, I would rather work on gettin my former desktop system running again since I moved it out to the lab and have yet to do so.
It's all in the name of getting the lab straightened out so I can start work on the 42" plasma display I picked up broken for $10 and have my own hunches regarding fixes. I'm just glad I did my dumpster diving (minus the dumpster) at work and got a nice piece of anti-static mat that's protecting the screen right now.
I've swapped the 3GHz Pentium 4 out of the system in favor of a 2GHz model. I'm going to see real soon whether that was the problem. So far, so good. I'd rather have a bad CPU than a bad motherboard as I have so few motherboards in working form.
Next things to do:
Uninstall the video cards from the IBM servers for later installation in the main frontend system.
Uninstall Pentium 4 componenets from front-end in favor of AMD Sempron-based parts. (Reason: 2GB RAM vs 1GB and ATX vs uATX motherboard for more slots to accommodate the extra video cards)
I'm installing extra video cards in the frontend for the simple reason that I need them for my monitors. I have an ATI X800 AGP card in there right now and it's okay for one display, but I'm not about to run two off of it. I could use other video adapters, but I want the Matrox boards from the eServers because they have DVI ports and I'm not about to run analog signals to S-IPS LCDs ever again. For the record, I have a Dell 2001FP 20" panel as my center display and two Dell 1800FP panels on either side. Only one of the 1800FP panels happens to work, but I have a total of three non-working ones to mix and match from, or completely rebuild since it's generally a power supply fault anyway. I just want one perfect display when I'm done. No pressure spots or visible scratches. Odds are I have one. I just need to get it working. Based on simple power-up testing I just did, all of them have bad power supplies. Giggity. Easy fix and not too expensive.
The Pentium 4 system has managed to get into the Windows XP install screen in actual graphic mode. I reinstalled the case-back and the cooling fan.
And it's working. Proteins are folding and today is a happy day.
I just reached for something on a shelf near the ceiling and my hand became very warm. Looks like all my heat is heading toward the top, as predicted by that whole hot air rising thing. Time for a fan to move all that air.
Got it. If I'm going to freeze my ass off, I'm going to freeze my face off, too.
Monday, December 1, 2008
Danger Danger! High Voltage!
I've been wanting to experiment with vacuum tubes for quite a long time, but there was always the problem of proper power supplies, especially for the high voltage side. Not anymore!
At work, we throw out a bunch of SAFT LS-series "AA" 3.6 volt Lithium thionyl chloride cells that are used in crash data recorders for six months at a time. When they come back, the cells still carry 80% of their useful charge. Wasteful, I know, but that's the price of reliability. Anyway, now that the demand for the units the cells are installed in has increased so much, there are more and more cells to dispose of. True to my treehugger nature, I don't want to see Lithium enter a landfill or go to a recycler in a dangerous form. While these same cells were one commonly used for clock backup in computers, this is now rare so it's low-volume.
So what do I do with the 300 cells I already have? Use them.
The advantage with Lithium cells is they have a relatively flat discharge curve, dropping like a rock at the very end of their life. This is a good thing since a consistent supply is always helpful.
Radio Shack, and other such dealers, sell holders for 8 AA size cells at a cost of about $2 each. While normal AA cells in such a holder would result in 12VDC at the terminals, Lithiums give 28.8VDC. My plan is to build a few different supplies. The first will use four banks (32 cells total) for 115VDC output while others will be built for up to 400 volts (14 banks or 112 cells). All of these supplies will be housed in insulated cases with isolated keyswitch interlocks and .063A fuses to guard against shorts. I'm only likely to be pulling a maximum of 30mA current anyway, and that's with a lot of tubes running. Most of the time, I'll be seeing a maximum of 10mA. The cells are rated for a maximum pulse current of 270mA. Continuous drain carries a 70mA rating. I'm sure the internal resistance of so many cells in series is going to limit that.
At work, we throw out a bunch of SAFT LS-series "AA" 3.6 volt Lithium thionyl chloride cells that are used in crash data recorders for six months at a time. When they come back, the cells still carry 80% of their useful charge. Wasteful, I know, but that's the price of reliability. Anyway, now that the demand for the units the cells are installed in has increased so much, there are more and more cells to dispose of. True to my treehugger nature, I don't want to see Lithium enter a landfill or go to a recycler in a dangerous form. While these same cells were one commonly used for clock backup in computers, this is now rare so it's low-volume.
So what do I do with the 300 cells I already have? Use them.
The advantage with Lithium cells is they have a relatively flat discharge curve, dropping like a rock at the very end of their life. This is a good thing since a consistent supply is always helpful.
Radio Shack, and other such dealers, sell holders for 8 AA size cells at a cost of about $2 each. While normal AA cells in such a holder would result in 12VDC at the terminals, Lithiums give 28.8VDC. My plan is to build a few different supplies. The first will use four banks (32 cells total) for 115VDC output while others will be built for up to 400 volts (14 banks or 112 cells). All of these supplies will be housed in insulated cases with isolated keyswitch interlocks and .063A fuses to guard against shorts. I'm only likely to be pulling a maximum of 30mA current anyway, and that's with a lot of tubes running. Most of the time, I'll be seeing a maximum of 10mA. The cells are rated for a maximum pulse current of 270mA. Continuous drain carries a 70mA rating. I'm sure the internal resistance of so many cells in series is going to limit that.
Subscribe to:
Comments (Atom)