Monday, June 27, 2016

NIR with Raspberry Pi


Pictures from a NoIR 5MP Raspberry Pi camera, not just for seeing in the dark...  You can do EL with a $100 camera/computer, power supply, and some image editing software.

Don't see anything in the raw image???

Change to greyscale and look really closely...

OK, digital time... auto-contrast with Photoshop Elements.  Same picture, I swear.  Do it yourself!
Probably should have covered the red LED on the front of the Pi NoIR camera.

Thursday, June 16, 2016

Treasured memories safe on CD/DVD? Think again...

Not along my normal lines of fixing, but I get into a little of everything.   In my quest to digitize our media library I've stumbled across a few scary things with CDs and DVDs that may change the way you think about these as a 20-50 year "archival" media.

If any of you know my work history of making solar panels, I've always been the "bricks and mortar" type.  Use thick sections of silicon instead of thin films because there's not much there, that's what I always say.  Apparently that transfers across technologies, as the thin films used to produce CDs and DVDs are subject to environmental factors that degrade them faster than you would believe, sometimes less than 10 years.

In digitizing probably over 300 CDs and 200 DVDs, I've found maybe 2% so far that have become unreadable by a modern computer/data grade drive.  There seems to be two primary failure mechanisms, either the lacquer coating to protect the data surface was damaged (and I'm not talking scratches) or the data layer itself has corroded.

On CDs, the data is pressed into a polycarbonate disc blank and then flashed with aluminum in a vacuum chamber before being spin-coated with a lacquer and made ready for silk-screening of the design.  The final disc is read through the bottom polycarbonate blank.  If that lacquer top coat does not seal the edges well or the aluminum is not masked or removed from the edge before lacquering, the aluminum is exposed and will oxidize and react with moisture causing the reflective layer in the CD to wrinkle.  Picture aluminum foil after you've crumpled it...

DVDs are a little different, and particularly 2-sided ones.  2-sided discs are pressed or laser cut both sides and a final lacquer applied.  In this case, the laser reads the disc through the lacquer coat, so if the coating is not applied properly, the disc not cleaned properly, or a sub-standard or contaminated lacquer is used, over time the layer can be permeable to moisture/oxygen and corroded the data layer, or the lacquer itself absorbs moisture and begins to take on an "orange-peel" texture.  Obviously not good for reading or focusing lasers to spot size ~0.8 microns.

So what to do??  Some brilliant people created an open source program for Linux called "ddrescue". Unlike "dd", which is a command line tool to write raw data to and from discs/files/whatever in sequential order, "ddrescue" works around a read error from the source and it skips ahead and keeps going, and flags that area for another attempt at some point.  It is exclusive of "dd", and uses it's own algorithm to attempt to retrieve as much data as possible.  It can even reconstruct a file/drive from two partials, say two drives of mirrored RAID setup that failed in different areas.  It has its limitations, but it's pretty darn good.

How to do it??  Understand what your doing, but essentially, within Linux you treat the CD/DVD as a data disc and recover it to the system hard drive sector by sector.  In my case, I had a side of DVD that was unreadable and truncated a 43 minutes program to 23 minutes and the following on from 43 minutes to 5 minutes.  After letting ddrescue chew on the disc for about an hour, it had recovered all but 69kB and 173kB from two source files, or less than a second of video.  The transcoding reader to digitize the files easily skipped over the error sections and there was hardly a minor blip.

NOTE:  All of my work for digitizing is for private/personal use only.  Whatever you do, understand copyright laws before you do your own work.



Monday, June 13, 2016

Is the A/C ready for summer?

It's been a while, my chances to get into things has dropped off recently with life intervening, but hey, it happens...

I had a useful one that's an easy fix (if that's the only thing that wrong!).   As things are heating up outside, I happened to be in the basement when our ground-source heat pump unit kicked on.  Instead of the familiar click and hum of the compressor starting, I heard a click and wub, no hum.  Thinking, "hmm that's not good" I  went over to the unit, which didn't show any errors, and shut it off on the control panel then turned off the local disconnect.  After waiting about 5 minutes for any energy to dissipate (capacitors), I removed the lid.

I wasn't familiar with the symptom, but knew enough about the motor and compressor setup to understand there was a problem someplace.  A quick online search brought up some similar issues in a WaterFurnace chat forum that the compressor motor run capacitor could be bad.  If you don't know what that is, it looks sort of like a V-8 juice can, right circular cylinder, that is connected with large wires into the motor/compressor circuit.  Here's the control panel side view of it (with tan and blue wires attached).


Usually when they go bad, they short and the can overheats and puffs out on the ends.  This one was very slightly domed, but when I disconnected the spade terminals and took it out, it tested open with an Ohm meter, and no reading on a capacitance meter (the can said 370VAC and 80uF or microFarad).

The chat forum had also mentioned the manufacturer went from a 440VAC rated capacitor down to a 370VAC for home units, which is normally fine since they run 240VAC.  We had recently had a power surge that took out a bunch of stuff in the house (that the power company paid to have fixed!) and could have damaged this too.  I was able to find this capacitor (440VAC, 80uF) at my local Grainger for $53, in stock.  I drove down to get it and the tech there also verified with their capacitor tester that the old one was open.

Came home, popped in the new one after removing the adapter plate which had a 2" opening instead of 2.5" for the larger 440VAC rated cap, closed it up, turned it on....  and the compressor started!

While I was in there however, I also noticed that there was quite a bit of dust underneath the compressor start contactor, and that the control board looked a little cooked under a  transistor and power resistor.  Rather than get that far into it, I called my local service guys that installed the unit (HEY, EVEN I HAVE LIMITS), and they came out and replaced the control board and contactor FOR FREE UNDER WARRANTY.  I was 8 years into a 10 year warranty.  Nice.  The contactor was severely pitted from cycling for 8 years, plus cycling for at least several days with a locked-rotor compressor.  They also checked out the unit and topped up the outside loop pressure, and verified it was delivering about 63,000BTU for a 5 ton (60,000BTU) rating.  All good.

Technical note:  For those of you not familiar with large single-phase motors, they typically have capacitor in series with one motor winding to create a "fake" second phase that makes the motor run much more efficiently.  With that capacitor open, the compressor was trying to turn two windings with only one connected, and that no worky.  See the compressor and cap in the schematic below:




Sunday, May 1, 2016

Helping out (and a free cup of coffee)

My local favorite coffee hang out, Dublin Roasters, had a problem... I thought they had closed when I didn't see their OPEN sign lit.  I stopped in and they were open, and asked about it (along with a cup of Volt Blend and a kitchen sink brownie).  Apparently the yoga class had crushed the plug.  They had replaced it but still wouldn't work.  I offered to take a look at it for them.

There were two problems, the new transformer they had found was the wrong voltage (6V) and way under-powered (like 300mA).  Getting it home, I opened the back and hooked a bench supply to the power contacts.  With a 1A current limit set, I slowly took the voltage from 6V up, the sign turned on at about 10V, at 12V it was going good, and pulling about 600mA for an LED sign.

As luck would have it, I had a leftover power supply with 12VDC and 3.15A laying around which worked perfectly.  I also happened to notice part of the second "e" in coffee was out.  I found that the PCB trace had broken on that small substring of lights, the LEDs were 5-6 in series for the 12V with a 150 Ohm current limiting resistor in the string.  A small scrape of the conformal coating and touch with the soldering iron, and it was back to full "e" status.


Thanks Dublin Roasters for the free cup of coffee and oatmeal bar on return!

2004 Toyota Prius DTC codes P0A40, P0A41, U0100, U0111

After a short trip, returning home is usually a welcome and pleasant time.  However on getting home and finding my Prius dash lit up like a Christmas tree with master warning and hybrid system warning lights, my enthusiasm quickly faded...

The last time this happened a rambunctious squirrel had its way with a the motor generator position sensor (generator resolver), as evidenced by some bare copper strands swaying in the breeze when looking straight down the front of the engine behind the radiator.  I suspected something similar, but this was a little harder to find.

After getting the car up on ramps, and crawling underneath, I finally found the offending connector, the motor position sensor (motor resolver), conveniently located halfway up the engine behind the drivers side wheel well.  I already had the Toyota part number and the tool I needed from the last time (made from a precision hex driver and a grinding wheel), so I ordered another 10 pack for about $7 each.


After removing the driver's side splash pan (have extra plastic body fasteners on hand!), I was able to get one hand in through the wheel well and another up underneath to wiggle both the motor temperature sensor and resolver plugs free.  The little rascals had chewed clean through two wires and nicked a third, but for the life of me couldn't remember how to get those darn terminals out.  After nearly an hour of unpleasant language, I conceded to cut the entire plug off (I had six new terminals and splices, well 10 actually).  After working at the kitchen table for almost another hour (dinner anyone?), here's the trick for my ailing short term memory and all you others that need to do it....

...remove the white plastic locking ring FIRST!!!  With a gentle pry up on both short edges of the white locking ring (exposed at the edge of the central plug), it pulls up and completely off the plug, exposing the plastic catches that you need to move away from the terminal to pull them out the back.

I'm waiting for the rain to subside so I can put it back on and DRIVE my Prius off the ramps.

End note, the U codes from the scan tool have to do with CAN bus communication to the battery, I think these are incidental to the furry friend incident, as they could be cleared from scan tool.

Thursday, April 7, 2016

When internal combustion no longer combusts...

HomeLite UT8514 30cc leaf blower & Homelite UT80522D 2700psi 2.3gpm pressure washer

Shaking the cobwebs out of the garden tools this year, I ran into two problems.  The trusty old leaf Homelite leaf blower that has been working faithfully since we moved out into the trees ran for about 3 minutes and then died suddenly.  As I pulled the cord to start it again, I noticed the head and spark plug wore seemed to be moving more than I would expect.  On closer inspection, I found one of the the three (YES THREE) head bolts snapped and the cylinder head was lose.  No compression, no engine.

If you haven't found it already, check out http://www.ereplacementparts.com (I have not received any compensation for their mention).  They actually have useful repair diagrams for a lot of tools and usually have reasonable prices and decent shipping.  Well, a look for the blower and I find new head bolts are $0.99 each, one broke, order three new and replace them all!!!
Cover removed, crankcase cover open, missing (snapped) head bolt just below the crankcase.

Snapped head bolt, they're only M5!
So, about a 30 minute job later (and luckily the head bolt snapped where I could turn it out with needle nose pliers without a whole tear down, bolt it all back together and it fired right up!

The other one for the power tool closet, the pressure washer.   OK, so it was giving me some fits to keep running last fall and I kinda winterized and forgot about it.  After 45 minutes of pulling, purging the fuel bowl on the carburetor, and using starting ether, I couldn't keep it running longer than about 10 seconds, and only then with the choke full on...  Grrrr....

Once again (the same order actually), ereplacementparts did it.  This time an ENTIRE new carburetor for a Homelite pressure washer was $23 (and this is a $250 pressure washer about two years old).  Other than the fact you had to unbolt the engine and plastic facade from the frame, this would have been a 5 minute job to replace, as it was, it ended up being about 25 minutes.  It looked from the old carburetor that the fuel jet was damaged (or not protruding into the intake far enough to get into the venturi stream) somehow and not putting enough fuel in, makes sense it would run on choke, eh?  
Note the hemostats to clamp the fuel artery.
Button it all up, five pulls (three after I remembered to attach the spark plug wire again!) later, and ran like it was new.

Fix it, don't throw it out! 

Tuesday, March 22, 2016

iOS 9.x Text Message on all phones sharing iCloud account

I've been experimenting in the iOS 9.3 Beta program to see if it fixes some of the phone hang up button and Bluetooth hand off issues I was having with a iPhone 5 on 9.2.1.  During one of the updates, it asked for iCloud signin again and my wife concurrently received a message asking to allow text messages from me.  She said yes, and thereafter both our phones were receiving text messages from each others phone.

It may be obvious to some, but I'll share because it drove us both nuts for about a week.

Go to Settings>>Messages>>Send&Receive and check which emails/phone numbers on the AppleID or iCloud account should receive messages to each device.  Somehow, all 6 were checked on mine and my number was checked on my wife's phone.

NOTE: We haven't gone through setup of the Family Sharing program yet because it requires a credit card, so far we've been living in the AppStore world with redeemed change jar gift cards.




What it should look like on a phone with only your number checked.

The other annoyance was that the it changed the contact info associated with with phone, as viewed through Phone>>Contacts, it listed my wife's name with my number.  I needed to change this back by going to Settings>>Mail, Contacts, and Calendars>>CONTACTS-My Info.