Friday, September 30, 2011
solar generation
Two Years of Solar Generation! 20.8MWhs total generation so far. That's up 500KWhs from last year so that's great, especially considering we had two really wet months meaning very little sunlight. My meter is down to under 600 and still going down, we'll see how high it has to run up during the winter. Hopefully next year I'll roll it backwards.
Wednesday, September 28, 2011
some random computer stuff
A script to send email about netbackup tape changes. I got the original from a contractor I work with at DPL sometimes.
#!/usr/bin/perl -w
#who get mail to eject tapes from the library
#$operator="marti,joshw";
$operator="netbackupadmin";
#who get mail to return tapes from the vault
$courier="netbackupop";
# make associative array of offsite media
open(PIPE,"/usr/openv/netbackup/bin/goodies/available_media|") || die "Could not run available_media";
while() {
chop($_);
if( /^Off/ ) {
;
while() {
if(/^$/) { last; };
chop($_);
($evid,$mtype,$rtype,$robot,$slot,$face,$reten,$size,$status) = split(/\s+/,$_);
if($reten =~ /\d/) {
$volumes{$evid}{slot} = $slot;
}
}
}
}
close(PIPE);
# if a media id has a numeric slot number
# it needs to be pulled from the library
foreach $evid (sort keys %volumes) {
if( $volumes{$evid}{slot} =~ /\d/ ) {
push(@eject,$evid);
delete($volumes{$evid});
}
}
#
#
# we need to check the expiration date of the rest of the media to see if
# they need to be returned from the offsite tape vault
$now = time();
foreach $evid (sort keys %volumes) {
open(PIPE,"/usr/openv/netbackup/bin/admincmd/bpmedialist -ev $evid -l|") || die "Could not run: bpmedialist -ev $evid -l : $!";
while() {
($id, $j1, $j2, $j3, $time_alloc, $last_update, $expire, $j4) =
split(/\s+/,$_,8);
if( $evid eq $id ) {
if( $now > $expire ) {
push(@return,$evid);
}
}
}
close(PIPE);
}
if( defined(@eject) ) {
open(PIPE, "|/usr/lib/sendmail -t") || die "Could not run sendmail for eject\n$!";
print PIPE "To: $operator\n";
printf PIPE "%s %d %s","Subject: OffSite Tape Vault:", scalar(@eject), "Tape(s)\n";
print PIPE "\n\n";
print PIPE "Eject the following tapes from the C2 Autoloader and send to the Offsite storage:\n\n";
foreach $evid (@eject) {
print PIPE "\t$evid\n";
}
print PIPE "\n\n";
print PIPE "Detailed Tape Information:\n";
foreach $evid (@eject) {
print PIPE `/usr/openv/netbackup/bin/admincmd/bpmedialist -ev $evid`;
}
close(PIPE);
} else {
$no_tapes_to_eject++;
}
if( defined(@return) ) {
open(PIPE, "|/usr/lib/sendmail -t") || die "Could not run sendmail for return\n$!";
print PIPE "To: $courier\n";
printf PIPE "%s %d %s","Subject: OffSite Tape Retrieval:", scalar(@return), "Tape(s)\n";
print PIPE "\n\n";
print PIPE "Return the following tapes from the OffSite Storage:\n\n";
foreach $evid (@return) {
print PIPE "\t$evid\n";
}
print PIPE "\n";
close(PIPE);
} elsif( defined($no_tapes_to_eject) ) {
open(PIPE, "|/usr/lib/sendmail -t") || die "Could not run sendmail for notice\n$!";
print PIPE "To: $courier\n";
print PIPE "CC: $operator\n";
print PIPE "Subject: OffSite Tape Maintenance: no action needed\n";
print PIPE "\n\n";
print PIPE "No actions needed this week\n\n";
close(PIPE);
}
----------------------------------------------------------------------------------
Fixed the package problem by removing the ld.config file reverting to the default system library path, probably changed via crle to build php.
Fixed the jpeg/gd/php image resize problem by linking the libjpeg8 libraries to the libjpeg62 libraries so that php and gd were getting the same library regardless of what was called
keep_jobs_hours in bp.conf to set number of hours of job activity shown in netbackup monitor
disable svc:/application/graphical-login/cde-login:default to stop graphical console
#!/usr/bin/perl -w
#who get mail to eject tapes from the library
#$operator="marti,joshw";
$operator="netbackupadmin";
#who get mail to return tapes from the vault
$courier="netbackupop";
# make associative array of offsite media
open(PIPE,"/usr/openv/netbackup/bin/goodies/available_media|") || die "Could not run available_media";
while(
chop($_);
if( /^Off/ ) {
while(
if(/^$/) { last; };
chop($_);
($evid,$mtype,$rtype,$robot,$slot,$face,$reten,$size,$status) = split(/\s+/,$_);
if($reten =~ /\d/) {
$volumes{$evid}{slot} = $slot;
}
}
}
}
close(PIPE);
# if a media id has a numeric slot number
# it needs to be pulled from the library
foreach $evid (sort keys %volumes) {
if( $volumes{$evid}{slot} =~ /\d/ ) {
push(@eject,$evid);
delete($volumes{$evid});
}
}
#
#
# we need to check the expiration date of the rest of the media to see if
# they need to be returned from the offsite tape vault
$now = time();
foreach $evid (sort keys %volumes) {
open(PIPE,"/usr/openv/netbackup/bin/admincmd/bpmedialist -ev $evid -l|") || die "Could not run: bpmedialist -ev $evid -l : $!";
while(
($id, $j1, $j2, $j3, $time_alloc, $last_update, $expire, $j4) =
split(/\s+/,$_,8);
if( $evid eq $id ) {
if( $now > $expire ) {
push(@return,$evid);
}
}
}
close(PIPE);
}
if( defined(@eject) ) {
open(PIPE, "|/usr/lib/sendmail -t") || die "Could not run sendmail for eject\n$!";
print PIPE "To: $operator\n";
printf PIPE "%s %d %s","Subject: OffSite Tape Vault:", scalar(@eject), "Tape(s)\n";
print PIPE "\n\n";
print PIPE "Eject the following tapes from the C2 Autoloader and send to the Offsite storage:\n\n";
foreach $evid (@eject) {
print PIPE "\t$evid\n";
}
print PIPE "\n\n";
print PIPE "Detailed Tape Information:\n";
foreach $evid (@eject) {
print PIPE `/usr/openv/netbackup/bin/admincmd/bpmedialist -ev $evid`;
}
close(PIPE);
} else {
$no_tapes_to_eject++;
}
if( defined(@return) ) {
open(PIPE, "|/usr/lib/sendmail -t") || die "Could not run sendmail for return\n$!";
print PIPE "To: $courier\n";
printf PIPE "%s %d %s","Subject: OffSite Tape Retrieval:", scalar(@return), "Tape(s)\n";
print PIPE "\n\n";
print PIPE "Return the following tapes from the OffSite Storage:\n\n";
foreach $evid (@return) {
print PIPE "\t$evid\n";
}
print PIPE "\n";
close(PIPE);
} elsif( defined($no_tapes_to_eject) ) {
open(PIPE, "|/usr/lib/sendmail -t") || die "Could not run sendmail for notice\n$!";
print PIPE "To: $courier\n";
print PIPE "CC: $operator\n";
print PIPE "Subject: OffSite Tape Maintenance: no action needed\n";
print PIPE "\n\n";
print PIPE "No actions needed this week\n\n";
close(PIPE);
}
----------------------------------------------------------------------------------
Fixed the package problem by removing the ld.config file reverting to the default system library path, probably changed via crle to build php.
Fixed the jpeg/gd/php image resize problem by linking the libjpeg8 libraries to the libjpeg62 libraries so that php and gd were getting the same library regardless of what was called
keep_jobs_hours in bp.conf to set number of hours of job activity shown in netbackup monitor
disable svc:/application/graphical-login/cde-login:default to stop graphical console
Thursday, August 18, 2011
solar panel and emergency battery for camping
I bought a Duracell PowerPack 600 and an InstaPark 30W solar panel with the idea of using them as a power source/buffer system for charging electronics batteries/boombox/lights/etc while camping without running the car battery down.
The PowerPack comes with jumper cables, an AC charger, and 3 AC outlets and 1 DC outlet. The solar panel comes with a controller, cable, and charge controller. I also bought a male-male 12VDC extension cable.
After getting all the parts together, it appeared that I was supposed to run hardwired wires from the panel to the charge controller to the plug for the battery. That seemed like a bad idea to me, so I went to Ace Hardware and looked around for a bit trying to find a good plug-n-play connection method. Basically something to attach the charge controller to that would have room for connectors on each side.
It was not looking good for that...tons of options, but none that looked solid and yet re-useable and easily attached/detached. Then it hit me..phone cord is still rated for the old Ring Voltage level.
So I bought a phone wall jack box and front plate with two RJ11 ports, connected the solar panel to a male RJ11 plug and then ran into a problem. The 12VDC connector for the battery has standard 12VDC wires, about 5 times larger than what a standard RJ11 can usefully be attached to. So I just ran it straight in through opening at the back of the RJ11 port and connected it directly to the back for the time being. Later I'll do a down convert on the wiring since it doesn't need to be that high gauge since it's carrying at most 12V@1.6A. I'll try to remember to take some pictures of it tonight.
Then when I get back from Burning Man I'll have to work out a better connector/plug combination.
Tested out the outputs with my new multimeter and even at twilight the panel is producing 12V@0.1A. The charge controller is only rated for 3A max, but then the panel only produces 1.67A max, so it should be a good reasonably fast charge in bright daylight.
The PowerPack comes with jumper cables, an AC charger, and 3 AC outlets and 1 DC outlet. The solar panel comes with a controller, cable, and charge controller. I also bought a male-male 12VDC extension cable.
After getting all the parts together, it appeared that I was supposed to run hardwired wires from the panel to the charge controller to the plug for the battery. That seemed like a bad idea to me, so I went to Ace Hardware and looked around for a bit trying to find a good plug-n-play connection method. Basically something to attach the charge controller to that would have room for connectors on each side.
It was not looking good for that...tons of options, but none that looked solid and yet re-useable and easily attached/detached. Then it hit me..phone cord is still rated for the old Ring Voltage level.
So I bought a phone wall jack box and front plate with two RJ11 ports, connected the solar panel to a male RJ11 plug and then ran into a problem. The 12VDC connector for the battery has standard 12VDC wires, about 5 times larger than what a standard RJ11 can usefully be attached to. So I just ran it straight in through opening at the back of the RJ11 port and connected it directly to the back for the time being. Later I'll do a down convert on the wiring since it doesn't need to be that high gauge since it's carrying at most 12V@1.6A. I'll try to remember to take some pictures of it tonight.
Then when I get back from Burning Man I'll have to work out a better connector/plug combination.
Tested out the outputs with my new multimeter and even at twilight the panel is producing 12V@0.1A. The charge controller is only rated for 3A max, but then the panel only produces 1.67A max, so it should be a good reasonably fast charge in bright daylight.
Sunday, July 31, 2011
July 2011 PV totals
18.7MWhs generated so far for a total of 1100 last month. I'm really hoping to get a real monitor for the system soon. Then I'll be able to get daily totals, graphing, web access, etc. I wanted that when I bought the system, but they failed to provide it to me and I haven't had the cash since then.
Thursday, June 30, 2011
PV totals
17.6mWhs generated so far, meaning I generated 2kWhs in June this year. I want my meter to spin faster backwards, and I want a better way to track the information. Maybe I'll be able to afford the web interface this year.
Sunday, June 26, 2011
Full spectrum solar
Full Spectrum Solar or at least more full than it is now appears to be on the horizon. This should drastically increase the efficiency of solar power generation. So many break throughs in PV, I can't wait to see which ones work out.
Tuesday, June 14, 2011
pmconfig cannot find/create /etc/.cpr_config
Now here's a bizarre one. I created the meta device to start mirroring two drives on this x2270 running Solaris. When I finished and let it finish syncing and rebooted, it worked fine. I added the other metadevices for the other slices and rebooted and I got a drive error.
I wish I could have copied down what that initial drive error was, because it might be important, but unfortunately Oracles ILOM java based remote console doesn't have any way to copy text to or from it. So...anyway, drive error. I reboot to DVD and run fsck. It finds some problems, but nothing major. I reboot and get drive errors again. So then I'm wondering if there was some problem in creating the meta device so I switch back to booting from the disk slice 0.
That's when I start getting the error in the subject line. Again, I wish I could have copied it all exactly but this is the main error:
/usr/sbin/pmconfig: cannot open/create "/etc/.cpr_config", Bad file number.
It also listed a missing inode. So I thought maybe that file was corrupted or something. So I did the reboot to DVD thing, changed the files name, found a valid version of it and copied it into the directory, etc. Should lock down the inode to the renamed file.
No good.
So after a bunch of googling and waiting for Oracle to get me an OS/software guy to call me he found the solution. And it was a stupid stupid thing on my part.
When I switched back to booting from the disk/slice, I forgot to change /etc/system so that it didn't include the reference to the metadisk as the boot disk at the bottom of the file. Very stupid mistake.
Now...I still don't know why it failed the mirroring, but now that the system boots and is stable again, I can start troubleshooting that part.
I wish I could have copied down what that initial drive error was, because it might be important, but unfortunately Oracles ILOM java based remote console doesn't have any way to copy text to or from it. So...anyway, drive error. I reboot to DVD and run fsck. It finds some problems, but nothing major. I reboot and get drive errors again. So then I'm wondering if there was some problem in creating the meta device so I switch back to booting from the disk slice 0.
That's when I start getting the error in the subject line. Again, I wish I could have copied it all exactly but this is the main error:
/usr/sbin/pmconfig: cannot open/create "/etc/.cpr_config", Bad file number.
It also listed a missing inode. So I thought maybe that file was corrupted or something. So I did the reboot to DVD thing, changed the files name, found a valid version of it and copied it into the directory, etc. Should lock down the inode to the renamed file.
No good.
So after a bunch of googling and waiting for Oracle to get me an OS/software guy to call me he found the solution. And it was a stupid stupid thing on my part.
When I switched back to booting from the disk/slice, I forgot to change /etc/system so that it didn't include the reference to the metadisk as the boot disk at the bottom of the file. Very stupid mistake.
Now...I still don't know why it failed the mirroring, but now that the system boots and is stable again, I can start troubleshooting that part.
Monday, May 30, 2011
homelite 24v cordless is a failure
In 2008 I bought a Homelite 24v cordless lawn mower. It worked beautifully and would often cut my yard twice per charge. That was even with the mulcher in, so it was mulching the whole yard without any problems.
Then in 2009 it wouldn't even charge up far enough to cut the yard once. So I got warranty service on it, and got a free battery out of it. Oh, but they decided it wasn't the battery, it was the battery charger which somehow wasn't covered in the warranty. Replacing the charger and battery cost me 80% of the original cost of the mower when it was new, but of course I didn't find that out until after the work had been done. I put in all of the complaints about that I reasonably could, and just hoped the mower would work for a while.
But no...even the same summer I'd had it replaced the new battery didn't work as well as the original. But it worked well enough to cut the yard. In 2010 I got the blade sharpened and figured out that if I took the mulcher out and was very careful about how I mowed and made sure to mow often and when it was as dry as possible, the mower would reliably cut the entire yard.
Now it's 2011. The first time this year I mowed, I noticed that the mower lost power much more quickly than it should. But it made it all the way through the yard before giving out. So I thought maybe just the first charge of the year was the problem. Unfortunately it rained pretty much all of May and I didn't get a chance to mow, so today when I tried to mow the grass it was a foot high and while I waited as long as I could for it to dry out, it was still pretty wet. The mower wouldn't even finish the front 1/8th of the yard. Now admittedly this is a worst case scenario and would have been a problem even for a gas mower, but I don't think there is any chance it'll work much better later in the summer when it's dry but hopefully it'll get me through this summer.
Next year I'll go get a corded electric mower. I really don't want a corded electric mower, I hate using my corded snow thrower it's such a pain in the ass. but I don't see that I've got much choice since I hate gas mowers even more. I certainly won't be buying a homelite mower again regardless...their service is absolute shit.
Then in 2009 it wouldn't even charge up far enough to cut the yard once. So I got warranty service on it, and got a free battery out of it. Oh, but they decided it wasn't the battery, it was the battery charger which somehow wasn't covered in the warranty. Replacing the charger and battery cost me 80% of the original cost of the mower when it was new, but of course I didn't find that out until after the work had been done. I put in all of the complaints about that I reasonably could, and just hoped the mower would work for a while.
But no...even the same summer I'd had it replaced the new battery didn't work as well as the original. But it worked well enough to cut the yard. In 2010 I got the blade sharpened and figured out that if I took the mulcher out and was very careful about how I mowed and made sure to mow often and when it was as dry as possible, the mower would reliably cut the entire yard.
Now it's 2011. The first time this year I mowed, I noticed that the mower lost power much more quickly than it should. But it made it all the way through the yard before giving out. So I thought maybe just the first charge of the year was the problem. Unfortunately it rained pretty much all of May and I didn't get a chance to mow, so today when I tried to mow the grass it was a foot high and while I waited as long as I could for it to dry out, it was still pretty wet. The mower wouldn't even finish the front 1/8th of the yard. Now admittedly this is a worst case scenario and would have been a problem even for a gas mower, but I don't think there is any chance it'll work much better later in the summer when it's dry but hopefully it'll get me through this summer.
Next year I'll go get a corded electric mower. I really don't want a corded electric mower, I hate using my corded snow thrower it's such a pain in the ass. but I don't see that I've got much choice since I hate gas mowers even more. I certainly won't be buying a homelite mower again regardless...their service is absolute shit.
Sunday, May 29, 2011
SMIT tensile solar
Flexible cloth-like solar. I'd be interesting in finding out how it attaches to the grid, or what it uses for level regulation, etc. Didn't find that information in a quick search though.
Friday, May 13, 2011
Some horrible horrible clues for php with pdo support on solaris
I have a drive going bad in one of my servers. Well, in the jbox attached to it. It's part of a mirror so as long as the other half holds until the new drive gets here, this is the swap plan.
Physically swap the drives.
Use the prtvtoc -> fmthard trick to push the vtoc from the other drive to the new drive
metareplace -e d50 c2t2d0s0
And then wait for it to sync. I can use the vtoc from the other half of the mirror since they are identical so I don't have to do that before the physical swap.
--------------
Now here's another horrible problem I worked through with Al Marti yesterday.
I needed php5.3.6 compiled with pdo and the extensions for it built. I grabbed the latest sfw package and installed it, but it didn't have options for enabling pdo in the php.ini. Sure enough, taking a look at the compile line used for the package, it didn't have pdo enabled. That means I have to compile php. I hate compiling php.
Compiling php is an amazingly annoying process for me. It never goes right, and the things that fail are things that should be problems in a big popular package. For the versions of 5.2 that I worked with before, I had to edit the makefile after running configure but before running make because of a typo in the makefile that wasn't in the configure line.
For 5.3.6 I had to cross link a library into /usr/lib even though the configure found the right directory with the file in it during the exploration phase of configure. Anyway...once I got configure working I ran into a problem with the system being unable to find libfontconfig.la. Sure enough, I couldn't find it either. Googling around produced nothing useful for me. Happily while Al was here, he caught a small misunderstanding. I had thought that on SunFreeWare if there was not a specific version number listed for a dependency for a package, then any reasonably recent version of that package would work. I was wrong, so I had missed some of the dependencies because the newer versions of the packages had different libraries. So that solved the libfontconfig.la problem.
Now on to the horrible...
So I get php compiled. It's got pdo enabled in the build. I can confirm that. Still no extensions for it. Al is looking around, I'm looking around. My brain fails so I decide to take a short break and build the apc module. Al sees that and asks about it, and suddenly we wonder "Could the pdo extensions be a pear/pecl module, not something built with the software?" So we try it.
pecl install pdo
Pdo exists! It starts to build! Yay! It fails! Boo!
pecl install pdo_mysql
Exists! Fails!
So we go back to googling around. After quite a while we both find the same bit of information pretty much at the same time. It seems ridiculous. Surely this isn't a reasonable solution just to get a module to build from a well known well used software package. But no...it is. We have to modify the code of the php package in order to get it to work.
From http://pecl.php.net/bugs/bug.php?id=18837
// /usr/include/php5/Zend/zend.h:320 add lines
#define refcount refcount__gc
#define is_ref is_ref__gc
// /usr/include/php5/Zend/zend_API.h:52 add lines
#define object_pp object_ptr
Completely ridiculous, but it works. The pdo package builds.
Now onto getting pdo_mysql to build. It fails because it can't find mysql_config. It finds the directory in which mysql_config exists, but fails to find the file. The file is right there. Baffling! More googling around turns up other people with the same problem but not a lot of solutions until we find this:
http://www.spiration.co.uk/post/1332/install%20PDO_MYSQL%20-%20mysql_config%20and%20header%20files%20problem
Which tells us to create a link from where it exists to another place, and the build will find it. Even though it's already finding the correct directory. Is that reasonable? NO! But we can't find anything else, so we try it. Sure enough...it works.
ln -s /usr/local/mysql/bin/mysql_config /usr/bin/mysql_config
Absolute madness. I hate building php. I really do. So far with us trying to use drupal I've been able to use a pre-built package successfully exactly once on 3 brand new machines when the installed version of php for Solaris was the newest, and we hadn't yet moved on to the newest version of drupal.
Physically swap the drives.
Use the prtvtoc -> fmthard trick to push the vtoc from the other drive to the new drive
metareplace -e d50 c2t2d0s0
And then wait for it to sync. I can use the vtoc from the other half of the mirror since they are identical so I don't have to do that before the physical swap.
--------------
Now here's another horrible problem I worked through with Al Marti yesterday.
I needed php5.3.6 compiled with pdo and the extensions for it built. I grabbed the latest sfw package and installed it, but it didn't have options for enabling pdo in the php.ini. Sure enough, taking a look at the compile line used for the package, it didn't have pdo enabled. That means I have to compile php. I hate compiling php.
Compiling php is an amazingly annoying process for me. It never goes right, and the things that fail are things that should be problems in a big popular package. For the versions of 5.2 that I worked with before, I had to edit the makefile after running configure but before running make because of a typo in the makefile that wasn't in the configure line.
For 5.3.6 I had to cross link a library into /usr/lib even though the configure found the right directory with the file in it during the exploration phase of configure. Anyway...once I got configure working I ran into a problem with the system being unable to find libfontconfig.la. Sure enough, I couldn't find it either. Googling around produced nothing useful for me. Happily while Al was here, he caught a small misunderstanding. I had thought that on SunFreeWare if there was not a specific version number listed for a dependency for a package, then any reasonably recent version of that package would work. I was wrong, so I had missed some of the dependencies because the newer versions of the packages had different libraries. So that solved the libfontconfig.la problem.
Now on to the horrible...
So I get php compiled. It's got pdo enabled in the build. I can confirm that. Still no extensions for it. Al is looking around, I'm looking around. My brain fails so I decide to take a short break and build the apc module. Al sees that and asks about it, and suddenly we wonder "Could the pdo extensions be a pear/pecl module, not something built with the software?" So we try it.
pecl install pdo
Pdo exists! It starts to build! Yay! It fails! Boo!
pecl install pdo_mysql
Exists! Fails!
So we go back to googling around. After quite a while we both find the same bit of information pretty much at the same time. It seems ridiculous. Surely this isn't a reasonable solution just to get a module to build from a well known well used software package. But no...it is. We have to modify the code of the php package in order to get it to work.
From http://pecl.php.net/bugs/bug.php?id=18837
// /usr/include/php5/Zend/zend.h:320 add lines
#define refcount refcount__gc
#define is_ref is_ref__gc
// /usr/include/php5/Zend/zend_API.h:52 add lines
#define object_pp object_ptr
Completely ridiculous, but it works. The pdo package builds.
Now onto getting pdo_mysql to build. It fails because it can't find mysql_config. It finds the directory in which mysql_config exists, but fails to find the file. The file is right there. Baffling! More googling around turns up other people with the same problem but not a lot of solutions until we find this:
http://www.spiration.co.uk/post/1332/install%20PDO_MYSQL%20-%20mysql_config%20and%20header%20files%20problem
Which tells us to create a link from where it exists to another place, and the build will find it. Even though it's already finding the correct directory. Is that reasonable? NO! But we can't find anything else, so we try it. Sure enough...it works.
ln -s /usr/local/mysql/bin/mysql_config /usr/bin/mysql_config
Absolute madness. I hate building php. I really do. So far with us trying to use drupal I've been able to use a pre-built package successfully exactly once on 3 brand new machines when the installed version of php for Solaris was the newest, and we hadn't yet moved on to the newest version of drupal.
Wednesday, May 11, 2011
solaris drive info
Why can I never remember this kind of stuff? It's very frustrating.
To find out all drives attached to a solaris box and their sizes/serial numbers/etc
iostat -E
To find out all drives attached to a solaris box and their sizes/serial numbers/etc
iostat -E
Friday, April 15, 2011
Electricity from solar power via magnetism
Now this is just amazing to me. The idea is to use the magnetic component of light to generate energy without absorbing the light.
http://ns.umich.edu/htdocs/releases/story.php?id=8368
http://ns.umich.edu/htdocs/releases/story.php?id=8368
Friday, March 11, 2011
Stanford sees significant increase in solar panel efficiency.
Standford is using quantum dots to drastically increase the efficiency of solar panels. The best part is it seems like a very easy cheap technique to use.
Tuesday, March 1, 2011
2/11 generation and a story about self-healing batteries.
13.5mwhs total generated so far for a total of 700kwhs for February. So it looks like I barely break even on generation in November, totally fail to generate enough in December and January, break even in February then I'm back to making more power than I use for the rest of the year. That's not too bad.
http://www.wired.com/wiredscience/2011/02/self-healing-batteries/
Here's an article about potentially self-healing batteries.
http://www.wired.com/wiredscience/2011/02/self-healing-batteries/
Here's an article about potentially self-healing batteries.
Wednesday, February 2, 2011
January 2011 power generation (edited)
Wow, I have to recheck this when I get home, but I'm pretty sure the meter said I'd generated a total of 12.3mWh. That'd only be 100kWh generated this last month. I know I'm not generating much because my meter keeps going up, but damn I was hoping it was more than that.
I was generating 100W this morning when I left. Doesn't seem like much, but it's only 1.5 hours after sunrise AND there was 4" of snow on the panels. Now that's impressive.
I re-checked this later and it was 12.8MWhs not 12.3MWhs. That's much more reasonable.
I was generating 100W this morning when I left. Doesn't seem like much, but it's only 1.5 hours after sunrise AND there was 4" of snow on the panels. Now that's impressive.
I re-checked this later and it was 12.8MWhs not 12.3MWhs. That's much more reasonable.
Sunday, January 30, 2011
solar panel payoff
Ok, I have all the information I need to figure out a pretty accurate total pay off time for my solar panels. Unfortunately to make it actually accurate would require a ton of data mining and calculations. So instead of that I'm going to do a quick and dirty approximation.
The 6.72kWh system cost $42672.
Xcel bought 20 years of my carbon credits for $23520 in the form of a rebate off the total price.
Then I paid the $19152 that was left.
I received $5444 in tax credit in 2009.
I received $5188 in tax credit in 2010.
I'll receive $3334 in tax credit in 2011 unless something changes.
And I hope to receive approximately $1111 in 2012.
I will get about $20 per year back from Xcel for generating more electricity than I use and letting them fuck me up the ass with a barbed pole for it.
Now the simplistic part of this calculation. I checked and my average power bill now is about $50 less per month than my average power bill was before I got the solar installed. While this is true, it's not really a very accurate assessment of the solar panels effect on my power bills. But until I do the work to mine all of the rest of the data required, we'll settle with that.
So...assuming I get this year and next years projected tax credits, and using the simplistic power bill difference above the out of pocket cost of these solar panels for me is going to be about $4000. So it will pay for itself in about 6.5 years. In that 6.5 years I will never pay for electricity on my power bill. (I will pay for electricity because I have to pay for what I use, then Xcel pays for what I generate, then charges twice what they pay me to someone else for the power I generated. It really is a giant FUCK YOU from Xcel.)
6.5 years total. 5.2 years from now. March 2016. Then I'll get another 13.5 years of solar power generation out of the system. Sure the system will slowly degrade, but they guarantee it'll be generating at least 85% of the power it generates now in 20 years, they actually expect it'll be generating more than 90% of it's rating at install for 25 years. And unless I begin using significantly more electricity than I currently do during that time, I'll never have to pay for electricity on my power bill.
Now...being the geek that I am, and having figured out that I'm not calculating this quite correctly, it's likely sometime this year I'll get annoyed with that and be bored and I'll end up data mining my power bills and doing a bunch of more accurate calculations.
And speaking of using more power...I'm planning on buying an EV next year, so that'll take up a crapload of power, and of course it'll never be plugged in during the day so that'll kill the balance. Of course I'm also intending to continue making my house more energy efficient and if I can figure out how to do it I'll be installing solar thermal and more solar panels. If I can find a way to do it I'll be installing a big assed storage system so that I store the excess power I generate during the day to use at night, but Xcel is kinda pissy about that so I'm not sure how I'll do it even if I can come up with the money. Which is considerable.
So that's some damned good news in my opinion, and hopefully will help convince some more people to invest in putting solar on their house. Especially since the panels you can get now are already much better than the ones I've got.
The 6.72kWh system cost $42672.
Xcel bought 20 years of my carbon credits for $23520 in the form of a rebate off the total price.
Then I paid the $19152 that was left.
I received $5444 in tax credit in 2009.
I received $5188 in tax credit in 2010.
I'll receive $3334 in tax credit in 2011 unless something changes.
And I hope to receive approximately $1111 in 2012.
I will get about $20 per year back from Xcel for generating more electricity than I use and letting them fuck me up the ass with a barbed pole for it.
Now the simplistic part of this calculation. I checked and my average power bill now is about $50 less per month than my average power bill was before I got the solar installed. While this is true, it's not really a very accurate assessment of the solar panels effect on my power bills. But until I do the work to mine all of the rest of the data required, we'll settle with that.
So...assuming I get this year and next years projected tax credits, and using the simplistic power bill difference above the out of pocket cost of these solar panels for me is going to be about $4000. So it will pay for itself in about 6.5 years. In that 6.5 years I will never pay for electricity on my power bill. (I will pay for electricity because I have to pay for what I use, then Xcel pays for what I generate, then charges twice what they pay me to someone else for the power I generated. It really is a giant FUCK YOU from Xcel.)
6.5 years total. 5.2 years from now. March 2016. Then I'll get another 13.5 years of solar power generation out of the system. Sure the system will slowly degrade, but they guarantee it'll be generating at least 85% of the power it generates now in 20 years, they actually expect it'll be generating more than 90% of it's rating at install for 25 years. And unless I begin using significantly more electricity than I currently do during that time, I'll never have to pay for electricity on my power bill.
Now...being the geek that I am, and having figured out that I'm not calculating this quite correctly, it's likely sometime this year I'll get annoyed with that and be bored and I'll end up data mining my power bills and doing a bunch of more accurate calculations.
And speaking of using more power...I'm planning on buying an EV next year, so that'll take up a crapload of power, and of course it'll never be plugged in during the day so that'll kill the balance. Of course I'm also intending to continue making my house more energy efficient and if I can figure out how to do it I'll be installing solar thermal and more solar panels. If I can find a way to do it I'll be installing a big assed storage system so that I store the excess power I generate during the day to use at night, but Xcel is kinda pissy about that so I'm not sure how I'll do it even if I can come up with the money. Which is considerable.
So that's some damned good news in my opinion, and hopefully will help convince some more people to invest in putting solar on their house. Especially since the panels you can get now are already much better than the ones I've got.
Thursday, January 27, 2011
Xcel Solar Rewards
Well, this is just incredibly disappointing. It's not like I expected a lot, but this undershot that by magnitudes. According to Xcel they are paying me $0.02857/kWh and I generated 680kWhs more than I used last year. So my Solar Rewards for this year are a whopping $19.43. Pays for the mandatory connection fee for two months. So...that just sucks. Once I've done my taxes for this year I'll be able to calculate the actual expected payoff for the solar panels and how much it's saving me a year and all that. With any luck it will be of more value each year for a while as I refine my ability to use less electricity and generate more.
By the way, they are charging me close to twice what they are paying me for electricity. Fuckers.
By the way, they are charging me close to twice what they are paying me for electricity. Fuckers.
Thursday, January 20, 2011
T25 exceptional EEC40 crash test results.
Gordon Murrays EV concept car T25 survives it's EEC40 deformable half barrier impact crash test with incredible results. Makes some sense given that it was designed around a safety cell as in the type of technology that keeps racers safe. This isn't actually the EV like the title says, but the only change between the two would be the drive train.
Saturday, January 1, 2011
December 2010 solar generation
500 kwhs this month. Should be the lowest month each year both because of the tilt of the planet and because December is generally kind of cloudy here. Still quite respectable. 12.2 MWHs generated total this year, and I'm pretty sure I'll never have to pay for electricity again. I generated 110kwhs more this december than I did last december, possibly because I managed to get the tree cut back. It's still not cut back far enough, but to do that I'd pretty much have to top it and cut it down. I really don't want to have to cut down my tree, and it doesn't obscure more than a panel at a time in the winter, so I think I'll leave it.
Subscribe to:
Posts (Atom)