Dan East

The War

Dan East — Sun, 02 Jan 2011 08:10:00 +0000

I just finished watching The War, a seven part World War II documentary by Ken Burns. It originally aired on PBS in 2007. I watched the entire series, all 14 hours worth, over the course of two days. I have a couple comments on the series, which is why I’m going to the trouble to “review” it here.

I’ll start by saying that I enjoyed the series, and thought it was well done. I learned a lot, and there were many touching, poignant moments. But there are a couple things about it that troubled me.

First, and most important to me, is that it was a very negative series. Specifically, it seemed to dwell on every poor military decision, social issue, and horror of war (all as presented from the US side) that it possibly could. In comparison very little negative was presented about the Axis powers, particularly things like the bombing campaign the Nazis waged against the British citizens, the Holocaust, the atrocities the Japanese committed against the Chinese, etc. The series would point out US losses of a specific campaign, but not quote the enemy losses, which were of similar number. The series simply concentrated on the false starts, the failures, the tribulations of the US military, and added as a footnote that in each case victory eventually went to the Allies. Great detail was presented about the firebombing of Dresden and Japanese cities, but not the numbers of civilians that died as a result of the Axis powers. Foremost, the series spent a great deal of time covering the segregation of African Americans in the US military. In fact, if I were asked “what was the dominant theme of the series” then I would have to answer it was the segregation of African American and Japanese American troops during the war. That topic was discussed in detail in several episodes, and was mentioned time and time again throughout. I understand that was the state of society in the US at the point in time, and thus of course it was reflected in the military as well, but WW2 was a milestone and a turning point regarding segregation in general. By the end of the war African Americans were involved in direct combat alongside white troops, and the Japanese Americans had proven their valor, loyalty and allegiance to the United States time and time again, and had the most decorated regiment in the entire US military. Yet all in all the US was demonized in each of these areas – clearly the balance was toward the negative in all cases.

The second issue that bothered me, and perhaps it was more evident because I watched the series almost straight through, was that some specific footage was reused to represent entirely different things. This led me to question how much of the various footage throughout the entire series was actually from the battles or even the whole theater of war that was being discussed. As an example, the footage of two young girls sitting on a bench in a tent was used for refugees in Italy early on, and then again much later for the Philippines. Clearly the footage was misrepresented in one case or the other (or possibly both). Another example of obviously technically incorrect footage was a scene showing large German artillery firing, followed by the explosions supposedly resulting from said artillery. In one case, immediately after the firing of artillery, footage was shown of an obvious demolition of a water tower, with soldiers standing and watching the explosion. Again, this served to make me wonder how much of the series was simply a hodgepodge of WW2 footage pieced together that appeared to show the topics as narrated. That discredited what I was seeing from being authoritative and accurate to almost a sort of fiction.

All in all I enjoyed the series, but now I feel like I need to watch an actual historic documentary that portrays a balanced, neutral, objective perspective of the war and its various facets. This series left me with the impression of a subtle anti-American agenda, and now I feel my overall impression of the war is slightly skewed as a result.

Debunking “Lucky Track Inspector” Video

Dan East — Sat, 11 Dec 2010 18:49:14 +0000

A viral video has been floating around the various video sites for a couple years now. Usually it’s entitled something like “Lucky Train Inspector”. Of course none of the posts cite any sources of any kind, and most descriptions refer to the person in the video as a Railway Inspector, although in some he is simply referred to as a drunk. It does appear he (supposedly) works for the railroad, because he seems to inspect a switch, and he is wearing a hard-hat that falls off part way through the video. I assumed this video was a fake from the first time I saw it for a number of reasons, and I’ve taken the time to definitely prove that this video is indeed a fake. Read on for the details.

What happens in this 32 second video is a man is squatting down in the middle of a railroad track looking at something, then suddenly he stands up and leaps backwards off the track, just barely moving out of the way before a high-speed train shoots down the track he was just on. He trips and falls onto another track just a couple feet away, and of course a second train suddenly appears on that track moving in the opposite direction. He manages to stand up and jump off of that track in the nick of time, and as he stumbles towards the track he started out on, the first train ends, and he falls safely onto the now empty track.

I wrote this video off as a fake from the first time I saw it, for a number of reasons.

An inspector would know for certain when and where trains are at. They wouldn’t be inspecting a railway without knowing exactly when the next train is due. They do this day in and day out – it is not left to chance.
There are 3events that happen, each within a tolerance of less than 1 second, that have to unfold perfectly for him to have survived:
- Jumping off the first track.
- Jumping off the second track.
- The first train moving past and clearing the rail.
The inspector did not see or hear the first train until less than one second before it would have hit him, even though his body is angled in the direction the train comes from. This is one of the most obviously unbelievable things about the video.
The inspector did manage to see the second train, even though the first train would have been an incredibly loud and windy distraction, and even though he just made a big physical exertion and was laying on the tracks disoriented.
No sources of any kind, even though the video would have had to come from an official railway security camera.

So I had already written this off as a hoax, but I wanted hard proof. That is what I present here – hard evidence that this video has been fabricated and is a fake. I’ve identified two completely different errors in the video, each of which independently demonstrates that the video is a composite of multiple videos. Thus this video is comprised of at least 3, if not 4, different video clips, all edited together to unfold with perfect timing within in the span of a few seconds.

The truth lies in the shadows. There is only one source of light in this video, which is the sun. Thus every object in this video should cast a shadow based on the angle of the sun in the sky, and the angle of the sun is coupled to the time of day. So all we need to do is compare the shadow cast by the man, versus the shadows cast by the trains.

So without further ado, here are the measurements. Note that the angle measurement is the degrees from vertical, not the degrees elevation.

Shadow of the man:

So we have 4 measurements from both before and after the event. The angles differ because the video is low quality and low resolution, thus the precision isn’t perfect. So let’s average the 4 measurements together:

50.5 + 53.5 + 50.7 + 55.9 = 210.6

210.6 / 4 =52.65

So the average angle of the sun relative to the man in the video is 52.65 degrees.

Now let’s look at the trains:

The average of those measurements is:

41.3 + 39.3 + 38.7 + 37.5 + 37.4 + 39.0 = 233.2

233.2 / 6 = 38.9

So to compare the two, the average angle of the sun for the man is 52.65 degrees, and 38.9 degrees for the trains. That is a massive difference! What this means is the video of the man was taken at a different time of day than the video of the train. It is so large of a difference that it is readily identifiable without having to actually measure. Look at the very first photo above. Note that the man is standing just to the left of the track, almost exactly how far out the train extends over the track. Thus he is at the same position as the left edge of the train. His shadow is quite long, and extends well over onto the left track. Now look at the 5th picture, which is the first one showing a train. Notice how the train’s shadow doesn’t extend much further onto the left set of rails than the man’s shadow, even though the train is at least twice his height?

Fortunately there is a frame in the video where we can see both the train and man’s shadow at the same time. Here you can clearly see the difference in the angle of the sun:

Here is a simpler way to consider the shadows. The man casts a shadow longer than he is tall, that is because the sun is lower than 45 degrees in the sky when he was video taped. You can see he is 62 pixels tall, yet his shadow is 74 pixels long:

Yet the train shadow is much shorter, and is not as long as the train is tall, because the sun is higher than 45 degrees in the sky:

Now granted, these measurements are not perfect, but they do illustrate just how far off the shadows are in this video. The train’s shadow to the left is clearly not as long as the train is tall, and it would have to be if it was recorded at the same time of day as the video of the man.

I suspected that not only was the video of the man composited into the video of the trains, but that the trains were composited together as well. In other words, I wondered if those two trains really passed by one another with that exact timing in the first place. The answer to this is no, they did not pass by at the same time as depicted in the video. The first hint is their shadow angles. The average for the first train is 39.76 degrees, and for the second it is 38.0. That would indicate that the trains passed by about the same time of day, but not at the exact same time. However, that wasn’t good enough proof for me. I’m not a professional at analyzing videos, and that is not a large enough difference for me to be certain the problem isn’t with my measurements. So here’s the proof:

The red line shows the angle of the sun, from the top of the right train, to the edge of its shadow on the left tracks. Notice the shadow extends more than halfway across the other train’s tracks, yet there isn’t any shadow at all on the left train – it is brightly lit! The reason for this is because these videos were simply composited together. It would have required CGI or manual editing of the images to alter the left train to have a shadow on it, which requires a great deal more work and skill than a simple composite. Plus, the trains pass by so fast in the video that there are only a couple frames where the shadow problem is really obvious.

I suspect that the third train, which passes by on the far left track, has also been composited into the clip. The reason is it is a nice artistic touch, but again, the odds of 3 trains, passing by one another within 20 seconds of one another, which all happened to be at the exact point the man was inspecting the tracks, is nearly astronomical. So it is my gut feeling that the person responsible for fabricating this video threw that third train in just for fun, but I don’t have any proof of that.

Well, there you have it. Definitive proof that at least 3 videos, shot at different times of the day, were composited into a single video, supposedly depicting the miraculous cheating of death by a railroad inspector.

Moral of the story? Always assume everything you see on the internet is fake, until proven otherwise.

HTC Touch Vogue Review

Dan East — Sat, 13 Feb 2010 16:57:09 +0000

I recently evaluated the HTC Touch Vogue (US Cellular branded CDMA phone) for a few weeks. I tried my best to get this phone to a state so I could use it as my primary phone, and I just about succeeded. However, in the end, the hardware simply had too many handicaps conspiring against it.

As with all devices, there are two distinct components that combine to produce the overall functionality and user experience of the device: hardware and software. As far as the software goes, I tried pretty much everything available. I experimented with Google’s Android, but eventually ended up with the most cutting edge version of Windows Mobile available – 6.5 running HTC’s TouchFlo shell in place of the default Windows home screen. Of course Windows 6.5 isn’t officially available for this device (nor is Android for that matter), but obviously it can be done. I found a good commercial 3rd party SIP (onscreen keyboard) that worked really well, and if it weren’t for that the device would have been a complete non-starter. Typing with a stylus is a ridiculous notion in this day and age.

So as I’ve already mentioned, the hardware was the ultimate reason for doing away with the device. In some ways it was simply inadequate, while in others it was poorly designed. Here are the specific hardware deficiencies I encountered.

Touchscreen. The touchscreen is designed for stylus use, plain and simple. It can be used for finger input, but the accuracy and sensitivity is poor compared to devices specifically designed for touch input. Typing on the device can be quite frustrating, and there is little tolerance for error. Again, the 3rd party SIP I utilized (in compact QWERTY mode) helped tremendously with its word prediction and completion, but it still doesn’t compare to typing on better touchscreen phones, like the iPhone or HTC Touch Pro 2.
Screen size. The screen is relatively small compared to most touchscreen phones. That makes finger typing on the onscreen keyboard even more difficult, because the keys are smaller.
Display quality. The visual quality of the LCD display is rather poor, particularly when it comes to viewing angle. Colors shift in hue, and black areas appear grey, if the display is not viewed straight on. In fact the problem is so bad that if the device is held in landscape orientation then each eye will perceive black regions as different shades of grey, resulting in a shimming sort of appearance which is quite annoying.
No standard headphone jack. This is a major pet peeve of mine. Devices capable of playing mp3s, videos, and streaming video and audio from services like YouTube and Pandora should have a headphone jack. Period.
Volume control. The design of the volume control on the side of the device is terrible. Instead of a rocker type control with an up and down button that provides good tactile feedback, the touch uses a sort of slider that moves up or down and springs back to the center when released. It is difficult if not impossible to get a feel for when the volume control is triggered up or down, since there is no good tactile feedback. With a standard push button volume control the user can push up or down any number of times to get an explicit number of steps up or down with the volume. That is very clunky with with the slider.
Buttons. The same design issues affect the other buttons on the device. The side buttons are totally flush with the case, making them difficult to locate by touch. The standard Call and End Call buttons on the front of the device are diminutive in size, especially considering their importance and frequency of use.
Indicator LEDs. At the top center of the device is the earphone speaker. The opening for the speaker is a small circle, about the diameter of a pencil. Recessed about a millimeter is the grill which covers the speaker. Behind this grill is hidden, and I literally mean hidden, the indicator LEDs. Their light is dim, obscured by the grill, and can only be seen when staring at the phone straight-on. Other phones, like Blackberries, put the indicator along the top edge, so it can be seen from as wide an angle as possible, even when the phone is in a holster at the waist. My hunch is some designer (not engineer) liked the aesthetics of combining the speaker and the indicator, and totally chose form over function.
Battery cover. The entire back half of the phone slides off to reveal the battery. There is no release mechanism, and the cover slides off with little resistance. In fact, it came off in my hand while talking on more than one occasion. Low impact drops, like off the bed onto a carpeted floor, onto the floor of the car, etc, nearly always results in the back cover falling off. And if the back cover comes off, the battery will fall out too. Significant drops, like from waist height, always result in the phone seemingly exploding into pieces – back cover, battery and stylus all go flying on impact.
Micro SD slot. This phone represented a first for me. After dealing with the absurd hiding places a dozen other phones obscure their SD slot, the HTC Touch was the first that I had to consult documentation to find the slot. Yes, it is that strange. At first I feared I was breaking the phone, as the entire right side is snapped off to reveal the slot. The one minor upside is that the SD card can be accessed without removing the battery, which can’t be said for most Blackberries and many other phones.
Light Sensor. Nothing can be said about the light sensor, because there isn’t one. I found this one of the most annoying things about the device. At night the screen would be too bright, so I’d go through several menus and options and darken the display. Then in daylight the screen couldn’t be seen at all. This was a real problem, and compared to my Blackberry Pearl, which seamlessly and automatically adjusted to the ambient light, it was like night and day.
Battery Life. Even with a brand new battery (which my carrier, U.S. Cellular provides for free), the phone seldom would last a whole day of just normal smartphone use. There simply aren’t enough amp-hours in that little battery to feed the demands of the phone.

None of the above issues could be worked around in a satisfactory way. Perhaps a case would have kept the back cover on. That is the only remedy that comes to mind to address any of the problems I experienced with the phone. Simply put, it was poorly designed, and gives the impression of a late 1990′s Pocket PC with phone connectivity grafted in.

So what is the upside? The main thing I liked about this phone was that the GPS worked really well. I’m not sure it is quite as good as my Garmin handheld GPS, but it was close. Perhaps the GPS stood out to me because that is one thing that never did work right on my Blackberry Pearl – I think I got the GPS to actually lock in on my position 3 times over the course of 2 years on that phone.

The other upside is that these phones can be bought unlocked for $50-$80 on ebay, which is a pretty good deal. I can think of all sorts of uses for a general purpose PDA with cellular connectivity and GPS for such a low price. However, as a primary use, mission-critical cell phone and PDA, the HTC Touch is simply inadequate across the board. It is hard to believe that the HTC Touch and iPhone hit the market within months of each other, because the Touch seems like it is hardware from an earlier era and generation of smartphone hardware.

Reviving water soaked electronic devices

Dan East — Tue, 01 Sep 2009 03:17:27 +0000

Unfortunately I’m becoming something of an expert on repairing or reviving electronic devices that have been submerged in liquids. Here’s a short list of what me, my wife, my 4 kids and our cats have been able to soak:

Kyocera Phenom – Ranch Dressing
Nintendo DS – Sprite
Motorola Razr – Cat pee
iPod Touch – Washing machine
Motorola Razr – Dr. Pepper
Samsung Cell Phone – Washing machine
FRS radio – Toilet
Blackberry Curve -Toilet

And those are just the ones I can remember off the top of my head. And yes, all those devices lived to tell the tale.

Now before I get into the various methods to “repair” a soaked electronic device, I’ll throw in the usual disclaimers. First, I make no guarantee or claims that what I say will work. Second, you’ll void any warranty that might actually be left on your device. Third, messing with electronic devices (especially lithium ion batteries) is dangerous, and could result in electrocution, burns, fire, toxic chemicals, etc. Fourth, if you open up any device then there’s a good chance that it will end up worse off than before you tried to fix it. Fifth, in short, consider what you’re about to read nothing more than entertainment, and don’t do any of the things I discuss in this article.

So let’s get into the nitty gritty. Say my wife Tracie has asked me to grab the clothes out of the washer and throw them into the dryer. In the process I see my Blackberry laying deathly quiet in the bottom of the washer. Here’s what *I* would do:

Try to keep from vomiting as my stomach locks up in knots as I fully realize what I’m seeing.
I do not turn on the device! No matter how strong the desire is for me to hit that power button and see if it works, I will not turn it on. Here’s why. Electricity is the number one enemy that can cause permanent damage. Water conducts electricity, and if the device is powered on then the water droplets inside the device can short and reroute electrical currents in ways that the components can’t handle. Even if the device were to work at first, as water droplets move around inside they can short and damage circuits.
Immediately remove the battery. This is absolutely critical. Modern electronics don’t even have real power switches. So when a device is “off” it is really just sleeping in a low power state, waiting for a button press or other event to wake it up. So as long as the battery is in the device, the potential for power to flow to where it doesn’t belong is there. But what about iPhones, iPods and other Apple devices that don’t have removable batteries? Well, there’s nothing I can do about that. So potentially those devices are more vulnerable to damage since the battery can’t be removed.
Do not plug the device in to recharge, or connect it to a computer or anything else. If I’ve been listening to anything I’ve said so far, then I should know that electricity = BAD, so plugging into a power source would be bad, bad, bad.
Open any access doors, panels, etc. Remove battery covers, USB covers, memory cards and covers, etc.
Physically remove as much liquid as possible. I do this by placing a towel on my leg, orienting the device so the largest opening faces down, and firmly hitting it against the towel over and over, physically flinging the liquid from the device.

Now, at about this stage in the game a decision has to be made. If the device was submerged in clean water then the remaining water will probably evaporate out, and if the device wasn’t damaged immediately when it was submerged (again, by the “electricity flowing through water to where it doesn’t belong” part) then there is a very good chance it will be fine. In this case I will bake it with a hairdryer. I heat the device evenly with a hairdryer. If the device gets too hot then it could be damaged, so I try to keep it about as hot as I can tolerate holding it in my hand.
However, if the device was soaked by something other than water (*cough* cat pee, Dr. Pepper, Ranch dressing) then drying it isn’t enough. Soft drinks will leave behind a sugary, sticky coating. Ranch dressing will simply dry into a sort of cement, and cat pee, well, that leaves behind urea and all kinds of nasty cat-pee-smelling-stuff (let me tell you from experience, that’s not exactly what you want to smell everytime you put your cellphone up against the side of your face). So in these cases I don’t have much choice – the device has to be completely disassembled. Now this leads to a completely new level of risk and commitment. Disassembling (and especially the mandatory reassembling part) pretty much any piece of modern electronics is not a task for the faint of heart. In fact, many devices, like cell phones and Nintendo products, cannot be dissembled without special tools. The Nintendo DS and DSi use screws that take a special bit, called a tri-wing. They look like a Phillips head bit but with only 3 wings instead of 4. I can’t run into Wal-Mart and buy a tool like that – in fact I don’t know of any brick-and-mortar store where I can buy that tool. Luckily I had already ordered one off of the internet for a previous DS repair job. Cell phones often use tiny Torx head bits, and those can be purchased at some hardware stores, but it is unlikely that most people would already have a tool like that (unless their cat peed on their cell phone at least once before).

I won’t get into the horrific discussion of dissembling mobile devices here, because the tools and techniques vary drastically from one device to another (don’t even dream of opening an iPod Touch / iPhone unless you’ve got a replacement glass bezel and touchscreen membrane handy, because you’ll almost certainly tear them both to pieces trying to pry them off). So I consult the internet to find specific disassembly instructions (there’s at least one YouTube video showing how to take apart any electronic device you can think of).

Now once I get the device apart, and separated into its constituent circuit boards, LCD panels, touch membranes, etc, I liberally apply 91% alcohol with a lint-free cloth, dissolving and wiping away whatever liquid is in there that doesn’t belong of the circuit boards, button membranes, etc. In the case of the Dr. Pepper soaked Razr, I actually soaked the circuit boards in a tray of alcohol. I clean all the plastic pieces, etc, really well. In the case of the DS that was soaked in Sprite, I had to remove each of the tiny ribbon cables from the connectors on the mainboard, and soak and wipe them thoroughly with alcohol to remove the sticky sugar that worked its way between the conductors. Note, do not apply alcohol to LCD panels! It either damages the LCD or leaves behind residue between the membranes – the panel will have permanent visual artifacts if alcohol gets inside it. Also, I make sure any alcohol has air-dried before applying heat with the hairdryer.

So far so good. Again, assuming the device wasn’t fried immediately due to shorts when it was first submerged then it might have a chance. However, the big IF at this point is the LCD display. Pretty much every mobile device in existence now has an LCD display, and those are finicky little suckers. See, the problem is that they are comprised of several layers of incredibly thin membranes, separated by incredibly thin gaps. When it comes to water, a little thing called Capillary Action comes into play. Remember seeing a thing called a meniscus at the surface of water in test tubes during chemistry lab? Well, me neither. But if I did remember that stuff then I would know that capillary action will literally suck water in between the thin layers that comprise an LCD panel, and there’s no way to actually dissemble an LCD panel to clean between them. If it is water, then most likely every bit will eventually evaporate and work its way out from between those layers (more on that below). However if the liquid was not water, then in all likelihood the LCD panel will end up with residue permanently left between those layers. In other words, it might work and be viewable, but it will look like crap.

So I’ve got my device all dried out. If I had to take it apart then I made sure I reassembled it correctly, and put everything back in place (sounds a lot easier than it is, that’s for sure). Finally I’m ready to power it up and see what happens. More than likely, the battery is dead. The act of shorting it out in a liquid allowed the current to flow unchecked, discharging the battery. So first I’ll try powering up by the battery, and if that doesn’t work, I’ll hook it up to external power.

Assuming it powers up and works, the first thing I’ll notice is the LCD display. It will look like fractal patterns of shimmering, strange shapes that change as the device is tilted. That is the liquid interfering with the polarizer film, causing the optical qualities of the LCD to be really whacked. There are two ways to quicken the removal of water from within the LCD panel. First is heat – back to the hairdryer. I’m always careful not to overheat the LCD, because that could damage it even worse than the liquid. It will take quite a while for the water to evaporate out of the LCD panel. It’s about like watching paint dry. When drying the iPod touch I took a picture of the pattern on the screen, and after half an hour of applying heat I could tell the water was dissipating when I compared the picture to the device. So it is a very slow process, but it does work.

The other technique is even slower, but it is better than not doing anything. Put some rice in a Ziploc bag and throw the device in with it. Have you ever wondered what those little baggies that sort of look like little tea bags are that are packed in with new electronic gadgets? They are called desiccants, and they basically help suck moisture out of the air, thus acting like a little dehumidifier. That’s what the rice does in this case – it pulls what little moisture there is out of the air in the bag, quickening the evaporation of water from within the LCD panel. Again, it is a slow process, and even without the rice it will evaporate on its own – eventually. My daughter’s Samsung cell phone took a few weeks for the moisture to evaporate out of the LCD in room air, and now the screen is visually perfect (that is if she can find it – she looses her cell phone weekly). And just to reiterate, if something other than water made it inside the LCD then it will likely leave residue behind that cannot be removed. That is the case with my 6-year-old’s DS – both LCDs work but they don’t look so good. The only alternative for perfection is to send it off for repair, or replace the panels yourself. The good news for me is that I have already replaced the LCD panels on both my 9 and 11 year old DS’s, so I’m already an expert in that area.

If you’re reading this half panicked after your toddler threw your iPhone into the toilet to see if it would sink or float, then don’t dispair. Make sure you don’t turn it on, lay a towel on your lap and start banging the water out of it. Follow up with hairdryer and a bag of rice, take two aspirin and don’t call me in the morning.

tolua bug / limitation

Dan East — Fri, 02 Jan 2009 18:15:51 +0000

Minor update: I now use swig for my C++ / Lua binding, and I like it much better.

I’ve been using the Lua programming language for several years now as an embedded scripting language in a game engine. In order to bind C++ classes into the Lua environment I use a 3rd-party library called tolua.

Recently I kept running into a bug with a new class interfaced via tolua that would manifest differently between Release and Debug builds. The symptom was apparent corruption of other unrelated objects within Lua. The specific situation was accessing a member of a class that was also a class. Here’s an example to make things clearer:

class B { int n; }; class A { B b; }; //create an object a of type A A a;

Then in lua when accessing the nested class the bug would be triggered:

a.b.n=5

tolua passes instances of C++ classes into lua using simple user data, which is basically just a pointer. To reduce memory overhead there is no metadata passed into lua – just the pointer. So whenever an C++ object is accessed in lua, tolua has to look the pointer up in a table or list in order to determine what the object’s type is. This is a design flaw, because memory pointers are not necessarily unique to a specific class. Specifically, the first member variable in a C++ class has the same pointer as its parent class. In the above example, &a and &a.b will have the exact same memory address, even though they are two different objects of different types! So the limitation is that, for two classes interfaced via tolua, one cannot be the very first member variable of the other.

The workaround is straightforward – either rearrange the order of member variables in your class so the interfaced nested class is no longer first, or place a dummy variable in front of it. If you do not have the flexibility to modify the class you are exporting then you may be in trouble – this would require fixing the design flaw in tolua itself.

I’m posting this in the hopes that it will keep someone else from wasting their time.

CAPTCHA

Dan East — Wed, 10 Dec 2008 02:29:59 +0000

It appears the CAPTCHA used to keep spam comments out of my blog was broken, which prevented any commenting at all. Oops. Well, that explains the total dearth of comments for the last several months.

Exploding Duracell 9V Battery

Dan East — Sun, 07 Dec 2008 04:32:28 +0000

Exploded 9V Duracell Battery. Click for full-size.

Well, a matter of minutes ago, while sitting here reading the news at my desk on my computer, I heard a loud “POP” sound, like a firecracker going off. It was pretty loud – enough to make me jump. A foot away from my arm was a Duracell 9V battery, and apparently one of the cells in it exploded. It warped the top where the connectors are, and popped the bottom loose. I felt it immediately and it was room temperature. It was not connected to anything or touching anything – it was simply laying on its side on my desk. Then less than two minutes later another cell exploded. I’ve got my Blackberry sitting here video recording the battery in case a third cell explodes. So far its up to 11 minutes and no third POP.

This is really disconcerting. I could understand if the terminals were shorted, and the battery got really hot causing it to rupture. But this thing was just sitting there, disconnected, not touching anything. It’s been on my desk like that for weeks – I don’t even know if it has been used or not. The fact that two different cells exploded within minutes of each other is very odd. I wonder if the barometric pressure is falling, causing pressure inside the battery to stress the sealed cell and pop it open. Right now the pressure is 29.88. I just consulted a history chart. The pressure was 30.18 24 hours ago, and it has been dropping steadily. However that can’t be the cause, because the barometric pressure was even lower than that 5 days ago on December the 1st.

This reminds me that a couple months ago I found another 9V battery on my desk (I don’t remember the brand) that also had also spontaneously exploded at some point, without being connected to anything.

Well, my Blackberry has been recording for 30 minutes now and no third explosion. So maybe the fireworks are over for tonight.

Nintendo DS Lite Surgery: It Lives!

Dan East — Thu, 17 Apr 2008 14:26:57 +0000

My son and daughter both got new DS Lites for Christmas. We purchased Nerf Armor for them both, knowing that they would be heavily used and dropped often. Well, around two months ago, my son dropped his DS, while it was open, onto solid concrete. One quick glance at it was all it took to get that sinking feeling in my stomach – the top LCD panel was toast. A crack ran diagonally from one corner to the other, and only a sliver of screen at the bottom actually rendered pixels. The Nerf Armor didn’t save it, but the fact that the screen was open when it was dropped was probably the main reason it was damaged.

So I began searching for replacement LCD panels. Having done quite a bit of electronic repairs in the past, I decided I would try swapping out the panels myself. I found some at various online retailers for $40-$50, but most were out of stock, and all were priced too high. By the time I would pay shipping and buy the special screwdriver required to open a NDS Lite, I would have had $70 in parts just to attempt a repair on a $129 product. So my son made due with only the bottom screen for a few weeks (hopefully that served as a good lesson in being extra careful with expensive electronic devices). I eventually started looking for LCD panels again, and found one for sale on eBay for $4.98. Yes, you read right. $4.98 USD, with $10 shipping from Hong Kong. I read through the user’s feedback, and it was good, however I learned the shipping would take around two weeks, and that buyers received no communication from the seller at all. So a couple clicks of the mouse, and less than 10 days later the new LCD panel arrived (didn’t receive one single email from the seller as expected).

I had done enough research to know that the DS Lite, along with the Wii, uses special security screws designed to keep the average person from opening up their hardware. The screwdriver that fits is a “tri-wing”, which is like a Phillips head screwdriver, except it has three wings instead of four. So I ordered and received the screwdriver for $8 (including shipping) before the LCD panel arrived.

On the day the panel arrived I did some googling for DS Lite repair info. I found and watched a couple videos, and it became clear that I was mistaken in believing that I would only have to disassemble to top half of the DS. The ribbon cable linking the LCD panel to the main board of the DS is made into the LCD panel, thus the bottom half of the DS must be disassembled so the ribbon cable can be routed through the hinge, and connected to the main board. It was also clear that some soldering was involved. This wasn’t a problem for me – I have a solder station with digital temperature control, and have done quite a bit of soldering on various projects in the past. However, it was something I overlooked during my preparation, and probably would have dissuaded me from attempting the repair if I didn’t already have that capability.

I’m not going to get into the details of the DS Lite repair. There are plenty of tutorials and videos on the net covering that. I will say that, as someone who has significant experience in repairing similar types of devices, it was a pretty tough job. There are a few small ribbon cables to deal with. There are tiny springs and metal posts that hold the shoulder buttons in place, that are tough to position properly and extremely easy to lose. Several types of screws of varying length are used, and it can be tough too keep track of exactly where they go when reassembling. This certainly would not be a good repair to attempt for a first-timer. The ribbon cable for the top panel has to twist and pass through one of the hinges, which at first glance appears to completely impossible to manage. I actually damaged the touch-screen ribbon cable during the repair. The bottom screen has two ribbon cables – a skinny 4-conductor cable for the touch screen, and a wider data cable. The two ribbon cables route around opposite sides of the mainboard, and they sort of hold the LCD panel in place when the unit is disassembled. I wasn’t careful enough moving the bottom half around, and the smaller ribbon cable was bent sideways within the connector, and part of it tore off. I extracted the tiny fragment with an Exacto knife, and was able to cut off a couple millimeters of the cable so that the edge was nice and straight. Otherwise I would have had to purchase an additional touch panel.

Now, as I mentioned earlier, the replacement LCD was cheap – $4.98 plus shipping. When I first fired the unit up I was quite excited because – quite frankly – it simply worked. After playing homebrew Lemmings for a while something began to grab my attention. The black areas of the top display had something like a diagonal sheen to it. As I tilted the display the sheen would move. I also noticed that the vertical viewing angle wasn’t so good – as I slightly tilted the DS so I was looking at the top panel from above, colors would quickly shift. Like reds would become greens.

So I fired up my daughter’s DS and did a side-by-side comparison.

The first thing I noticed is the new LCD has four distinct backlights along the bottom edge. On the OEM screen the backlighting is perfectly even. Now this only stands out if I look along the display, in an almost parallel line of sight. When holding it normally I don’t notice any uneven backlighting.
The biggest difference is the vertical viewing angle. The OEM display has a much better tolerance for viewing from less than 90° angles. The colors would quickly shift on the new display, while they changed more slowly on the OEM display as I tilted the DS. The horizontal viewing angle (from left to right) is good.
There is a diagonal sheen visible in black colors that moves up and down as the screen is slightly tilted.

However, to an 8 year old, none of these issues would be apparent. My son never missed a beat with the new screen, and has not mentioned any difference in quality whatsoever.

So, the repair cost me $15 for the display and $8 for the screwdriver, for a total of $23 to replace the top panel. The next time around I won’t have to buy the screwdriver, so it will only cost $15 total. Of course there won’t be a next time though. Yeah, right.

KLH R7000 No-Audio Repair

Dan East — Sun, 06 Apr 2008 20:52:52 +0000

For those of you looking for the actual R7000 repair, I get to that at the very bottom of this blog. No, the power cord was not the repair.

I am the proud owner of a KLH R7000 tuner / receiver. It is a 600 watt receiver with Dolby Digital / DTS / Dolby Pro Logic II surround sound decoding, and a plethora of video and audio inputs and outputs. It has composite, super and even component video inputs and outputs, to allow easy switching of various video sources without having to resort to an A/B switch.

The reason I’m a proud owner of this receiver isn’t because I am in possession of this low-end, budget piece of hardware. It’s because it didn’t cost me a cent. Well, actually, it cost me $1.95, but I’ll get to that later. This receiver gets good reviews – when it works. It sold for as low as $99 at Best Buy (based on my googling), and has very good audio qualities for the price. The problem is they seem to die pretty quickly – usually in a matter of several months. So how did I happen across one of these receivers? Well, one day I was hauling off the trash (around here we have refuse centers where we take our trash and recyclables) and happened to notice a pristine-looking tuner / receiver in one of the trash bins. The power cord had been cut off, but besides that, it looked brand new. I snagged it up and took it home, where it sat for a good month before I finally decided to check the thing out.

Once I finally got around to trying out the new receiver, I headed on over to the local Goodwill, where I purchased a cute little ceramic Potpourri warmer for $1.95. Although it had a nice country look to it – blue enamel glazed in pretty patterns, what really caught my eye was the heavy-gauge power cord with polarized plug. So after returning home with my prize I removed the metal plate from the bottom of the pot, and found that I had done someone a big favor. The leads going to the heating coil were corroded and burnt through, and were completely shorted. At best this thing would have blown someone’s breaker, and at worst burnt down their house.

So, new power cord in hand, I opened up the receiver, grafted it on in a quite professional way, fired it up, and… it worked! All the inputs, all the outputs, the thing worked perfectly. Yes, I still had this nagging feeling in the back of my mind, “exactly why was this thing laying in a garbage bin with the power cord cut off?” I inspected the inside thoroughly, checking the top of the mainboard for anything that looked burnt or browned from overheating, and everything looked new. After sorting out the rats nest at the back of my entertainment center, I had all my speakers hooked up, all my audio and video from my DSS receiver, video game console, DVD, etc routed in and back out of the receiver, and I was ready to go. I fired it up, and couldn’t help but grin at how well it worked.

My next concern was the remote. Unfortunately the remote was not to be found in the trash bin on that fateful day. I consulted the on-screen manual for my DirecTV DVR, and tried both the KLH codes it listed. Neither would operate the receiver. Now I was really concerned. If I couldn’t control the volume from my DirecTV remote, then the thing was almost worthless. I thought maybe I would have to invest a good universal remote that would control the KLH. But before I had to do anything that dramatic, I first consulted Google. Lo and behold, I came across this page, and if you’ll scroll down to the KLH brand, you will see three codes. Yes, it contained the Lost KLH Code of the Ancients, hitherto thought lost to posterity: 31428. And, you guessed it, that third code operates the KHL R7000 receiver. So now I’m really on a roll. I mean, this must be meant to be, right? Not so fast…

After a solid week of use disaster struck. With a loud “pop” out the speakers, the R7000 became deathly quiet. Regardless of the input source, or the volume setting, or anything else I would try, there was silence. The LED display was as lively as ever, but no sound emanated forth from the speakers. More googling, and I found others that had the same problem, but no solution. So I decided to try and fix this thing, once and for all.

The first step any good repairman will use is to bang on the item that is not working properly. This

Three power transistors with open solder joints. Click for large image.

simple technique illuminated the problem immediately – after a few good whacks with my fist, the tell-tale “pop” came out of the speakers. I knew right then I had a “cold” or open solder joint. I removed the cover, and using an insulated probe continued my “banging” in a more refined manner, by tapping on individual components. I started with the relays, because they sometimes stick. Next I moved onto the power transistors, because I knew they were the stars in this show – they switch a lot of power, get really hot, and determine what wattage the amplifier can handle. The stress these components endure is why they are one of the main things to go out in any power amplifier. All the actual audio transistors are mounted on a large heatsink on this receiver, which runs the entire width of the case. They all looked good, and had plenty of thermal paste binding them to the heatsink. So I moved on to three other power transistors located at the center of the mainboard that were mounted to their own little heatsink – there I struck gold.

Solder joints to reflow. Click for large image.

After tapping on one of the transistors the speakers “popped” and I saw a spark at the base where the leads go through the mainboard. This thing was as good as fixed already. The next problem was how to get at the bottom of the mainboard, where all the solder joints are at. I nearly goofed at this point, and was about to start removing over a dozen screws and connectors and various other hard-to-remove-things to try and get the mainboard out of the chassis. Fortunately I took a look at the bottom of the receiver, where I found a convenient access door held on by seven screws. After removing the door my hunch was proven correct – there in the center of the mainboard, at the base of the three power transistors, was obvious thermal damage. The mainboard had been so hot that it blackened, and some of the circuit board coating had blistered. A quick inspection revealed that all three legs of the middle transistor had open

Click for large image.

solder joints. The components had gotten so hot that the solder had been softened enough to fall away from the legs (see 3rd picture). So now the task was simple – I fired up my soldering station, cranked it up to 850F, and re-soldered all the solder joints in that area of the circuit board. I tried to simply reflow the existing solder but there wasn’t enough, so I just added some new to the existing.

I put the thing back together, fired it up, and had an even bigger grin on my face the second time around, because that nagging question “exactly why was this thing laying in a garbage bin with the power cord cut off?” was answered – because of a little solder joint that needed to be reheated.

Click for large image.

When I get around to it I will blog about the various free items I have repaired over the years that we use daily. The total cost of the items would have been several thousand dollars if they were purchased.

Until then, maybe some R7000 owner will stumble across this page, and be able to breath new life into their little receiver.

Update: It’s been 8 months since the repair, and the receiver has worked flawlessly, day in and day out. It’s been operated somewhere in the neighborhood of 1000 hours since the repair.

Update: Now it’s been 16 months, and the problem reoccurred. This is no surprise, as the R7000 is improperly designed, and cannot dissipate enough heat from these transistors (they need heat sinks!). A couple minutes worth of soldering and everything is back to normal.