Time for a new power supply

After nearly 6 years, it seems the power supply in my computer has decided to go flaky.

The problem started about a week ago when the computer abruptly shut down on me. After a few seconds it powered back up, and didn't happen again for a while. Then the sudden shutdowns started to become more frequent. Sometimes the computer would stay running overnight, other times it would shut down 3 or 4 times in rapid succession. It would stay running for a few hours, then boom. Other times it would be up for a few minutes, then boom. Frequent but intermittent.

Primary suspect was the power supply. I reseated all the cables going from the power supply to the motherboard and wiped the BIOS, but that didn't help things. Then I noticed the main power switch on the power supply was broken and didn't toggle on/off anymore. Seems like a good candidate to cause such problems.

While the power supply was the main suspect, I wasn't able to rule out other problems too like something bad on the motherboard. Fortunately someone in the Charleston Tech Slack (if you're looking for geeky tech-y people to connect with, sign up here) had a couple of 650W power supplies he was willing to let me borrow. They're smaller than the 800W supply currently in the computer, but 650W is enough to run things.

Disconnected the old power supply from everything and connected the loaner. Since the power supply is just a loaner, I didn't bother with putting it in the case.


Computer's been up and running on the loaner power supply for a few hours now with no issues so far. Compared to what the computer had been doing before, that's a significant improvement and a good sign that something in the power supply is faulty. I'll keep it going for a couple more days to see how things go, and if everything's ok I'll order a new power supply to replace the original one.

New mailbox and post

At some point Thursday night, or very early Friday morning, the mailbox post got knocked down. No idea what happened or who knocked it over, but it was hit hard enough to snap the 4x4 post right off a few cm above the ground. I was just walking out of the house to the car Friday morning and saw it lying there on the ground.


*sigh*. Jerks.

So, it was off to Lowes to buy a new post and some concrete. This morning I dug out a hole next to the original post, cut what was left of the old post off at the ground and put in the new post. Got it nice and straight, set it in the ground with the concrete and filled in the rest of the hole with dirt.

We had been planning on replacing the mail box with a new one anyway, so I went ahead and put that on to the post. Shiny new mailbox post, shiny new mailbox. Maybe it will help add to the curb appeal of the house.


Tree nuts

The trees next to the house drop these large nut type things every other year that I've been told are pecans. This year the trees have dropped considerably more nuts than I've ever seen before.

Making things more interesting is that a bunch of them have split open and started sprouting, which is another thing I haven't seen these things do (or noticed anyway).




I planted some of them in some holes I dug around the yard, and the others I left on the ground where they sprouted. Since we won't be here much longer, I won't get to see what happens to them, but hopefully they end up sprouting into saplings and then into trees.

BarcampCHS 2015

Another Barcamp Charleston is over with. The seventh Barcamp Charleston was a lot of fun. Significantly smaller than previous barcamps, but still a good time. There were also fewer presentations pitched this year as a result, but there were some pretty good ones.


After the pitch sessions, I started the morning off hanging out with people in a BoF (Birds of a Feather) session. After that was my first presentation that I called Charleston Area Amateur Radio. Had a fairly decent sized group come to my talk where I talked a bit about the ham radio clubs in the area and the repeater networks around here that are available for use.

My session ended a bit early, and I was able to catch the Palmetto Scholars Academy kids working on their high altitude balloon launch demonstration. Their balloon didn't go all that high, but they did a good job of explaining and demonstrating the process and requirements behind launching high altitude balloons.


After the group photo and lunch was a session on graphs and graph theory by Denise Gosnell, one of the data scientists at Pokitdoc. Neat stuff. Kudos to Pokitdoc for being BarcampCHS sponsors and sending a great group of people to give some great presentations.

My second presentation of the day was a simple Ask a Medical Physicist session. Much to my surprise, I actually had a few (3) people show up, so it was just a very informal Q&A session where I gave a brief description of some of the things medical physicists do and answered any questions they had. It turned out to be a pretty decent session from my perspective.

The last two sessions I went to were on JSON and Clojure. They were both good sessions and I learned enough to dig a little more into them.

The crowd at this year's Barcamp Charleston was a lot smaller than previous years, but there were a lot of new faces there. Hopefully they'll spread the word and bring friends next year.

This is the official #BarcampCHS 7 group photo #CHS #Coc #community #Tech

A photo posted by Joseph Nienstedt (@joel8x) on

Two of the non-invasive meters I use at work can measure everything I need in a single shot: kV, exposure, exposure time, filtration, half value layer (HVL).

Even though the meters will calculate the HVL of the x-ray beam for me from one exposure, I still measure it with different thicknesses of aluminum and calculate the HVL. Just from first order impressions of using the meters over the years, I always felt that the HVL determined by the meter was never as accurate as doing the actual measurement.

It's not anything I've ever tested or looked closely at though. With all the HVL data I have from the past few years, I can take a closer look at how the meter HVL and my calculated HVL compare. There's at least a couple of years worth of HVL as a function of kV, so a plot of HVLmeter vs HVLcalc should give me a straight line with a slope of 1 if the meter HVL is useful.

I'll need to go back to my spreadsheets and extract the meter derived HVLs and collect the corresponding measured HVL so it could be a while before I find a big enough chunk of free time to do that.

If it turns out the meter derived HVL is close to my calculated HVL, that means I could probably modify my data acquisition and do my testing with fewer exposures.

November 7 (in addition to being BarcampCHS 7) is the third annual International Day of Medical Physics as well as being Marie Skłodowska Curie's birthday. The theme for this year's IDMP is "Better Medical Physics = Better Cancer Care in Radiation Oncology".

International Day of Medical Physics

Being a diagnostic medical physicist, I generally don't have a lot of involvement in radiation oncology. The use of imaging in radiation therapy is growing though, so I have more occasions to work with my therapy physicist colleagues these days.

Diagnostic medical physicists like me stay busy doing a great many things, including:

  • equipment performance tests to make sure the imaging equipment is performing according to state/federal guidelines and manufacturer specifications
  • quality control/quality assurance
  • teaching residents and technologists the physics behind how the equipment works and how the images are generated
  • calculating radiation dose estimates
  • creating lead shielding plans for x-ray rooms
  • working with radiologists and technologists to optimize imaging protocols and techniques
  • working with field service engineers to diagnose equipment related problems
  • troubleshooting image artifacts
  • research
  • collaborating with radiologists and other physicians on research projects

You probably won't see us if you need to get an x-ray or imaging study done, but we're there working behind the scenes making sure the images look good while using a minimal amount of radiation.

Questions? Feel free to ask!

Garmin StreetPilot 2610 guts

On the dissection table today is an old Garmin StreetPilot 26100 GPS unit. This is one I used in the car about 7 years ago, and then has been sitting in the console box since I got cell phoned.


Undoing about 5 screws was all it took to separate the two halves of the GPS unit. A couple of ribbon cables and a wire connected the two halves.

DSC02314.JPG DSC02315.JPG

Disconnecting the ribbon cables and wire gets us a good look at the boards. The board on the back half of the unit contains the CF card slot, mini-USB port and connections for power and computer connection.


The front half of the board is where all the interesting stuff is. Co-ax cables go from the patch antenna (top) and auxiliary antenna connector to the ICs that handle the GPS bits under the metal cover. Ribbon cables connect to the display, rear circuit board and the buttons at the front.


The patch antenna showed some discolouration and scratches. Not sure if that was during manufacturing, or if the unit was previously disassembled.


The patch antenna is soldered directly to the board and to the metal shield, so if I'm going to re-use it for something, I'll have to cut it off. Removing the cover from this part shows a 32.735 MHz crystal and a bunch of other unidentified components.


Underneath the other shielding section on the right side of the board is the brains of the unit, a 16/32 bit ARM core RISC microprocessor.


Removing a few screws separates the main board from the display module. The other side of the main board contains a few more large ICs underneath a shielded area (128 Mb SDRAM, 512 Mb flash RAM, 16 Mb flash memory, Garmin BIOS ROM) and components associated with the display module.


Moving on to the display module, bending out a few tabs allows it to be removed from its metal frame. The touch screen display module consists of the front glass, the LCD module, some plastic gratings/diffuser sheets and the backlight unit. It's an interesting little module. The backlight unit consists of a U-shaped light that shines into a block of plastic. One side of the plastic has a matted finish to help diffuse the light.


On top of the light are some plastic diffuser sheets that polarize, scatter and block some of the light.


Finally comes the LCD unit. Seems to be a pretty run of the mill LCD.


Not a whole lot of re-usable components in here. I'll clip off the patch antenna for the junk bin and keep the buttons and mini-USB module. I might try to remove some of the larger components off the boards. Aside from that, the rest of it will probably end up in the electronics recycle box eventually.

Phone repaired

My Nexus 5 arrived a couple of days ago after I shipped it back to LG to have the power button issue repaired. After it was received at LG, the post-assessment email said the phone was out of warranty by date code, and that I could have the phone repaired for $45, or have them do a full refurb for $179. I opted for just the repair and a few days later it was shipped out. Total turn around time from mailing out to coming back was 13 days.

Unpacked the phone, popped the SIM card back in and booted it up. Took a while, but the phone worked just fine, just like new. It came back with 5.1.1 (Lollipop) installed, but quickly picked up two 5.1.1 updates, and the next morning my phone was telling me the 6.0 update was ready for downloading.

Getting all the apps reinstalled was a bit of a long process, but went pretty smoothly. Fortunately I had recently backed up all the important stuff, so data wise I didn't really lose much.

While the Nexus 5 was off being repaired, I put my S2 back into service as a wifi only device, which worked out reasonably well. Using Android Beam to transfer (via NFC) the few files I had collected/modified back to the Nexus 5 worked out pretty well, and was nice and easy.

Now the Nexus 5 is pretty much fully restored and practically feels like a new device.

Lollipop updates

A while back, I took the plunge and rooted my Transformer Prime tablet. Found an Android 4.2 (Jelly Bean) ROM to install and installed it along with a new recovery ROM. Worked pretty well and made the tablet fun to use again.

The Galaxy S2 got the Cyanogenmod treatment with CM11, but for the past few months, ROM updates have been relatively infrequent, and there were no signs of Android 5 (Lollipop) ever making it to the S2.

With Android 6 (Marshmallow) being pushed out to phones, I decided it was time to bring the tablet and S2 a little further into the future. The XDA Developers forums is a good place to find builds of new ROMS for all kinds of devices. The Transformer Prime got the KatKiss 5.1.1 ROM along with a fresh TWRP recovery ROM while the S2 got the Candy5 5.1.1 ROM as well as the TWRP recovery.

The ROM install on the S2 was pretty painless, and seems to be running pretty well so far. Thought the S2 might have some issues running Lollipop, but so far so good.

Well phooey. My Nexus 5 appears to have suffered the same hardware failure my S2 suffered a couple years ago. The phone doesn't do anything when I push the power button, although sometimes if I push it hard enough it will come on. Plugging the phone in causes it to power up, flash the Google logo on the screen, and then it immediately shuts off and begins the cycle all over again. It doesn't even stay on long enough to get into recovery mode. That also means I can't back up or copy anything off the phone.

None of the usual incantations have helped, so it looks like a call to the Google help center is in store. It's a couple months out of warranty now, so I doubt there will be anything they can do aside from have me send it back to LG and pay for them to repair the phone.

Another alternative is to see if any of the local cell phone fix-it places around here can do that kind of repair (or just replace the board the switch is on) or to find a switch or mother board and replace it myself.

Off to deal with people on the phone.

Update: Decided to ship it back to LG for repair. After speaking with a Google tech support person and doing a bit of troubleshooting with him, he established that it was out of Google's warranty coverage period and put me in touch with LG support. At LG, the first person I spoke to said their warranty coverage was 15 months from purchase, and the repair would probably be covered under warranty. Had to dig around my email archive to find the proof of purchase they needed (it had to have the IMEI number), and when I called back the second person said the warranty coverage was 15 months from the manufacture date. So now I have no idea if the repair is going to be covered under the manufacturer warranty or not, but they have the info they need so I'll let them sort it out there. I'll have the repair done regardless of whether it's covered or not.