Sunday, March 20, 2011

The Mystery of the Non-Waterproof Hot Glue Solved

Six years ago I had a bad experience with hot glue.  I expected it to be waterproof and it wasn't.  I did an immersion test and within 3 days the glue turned into a whitish gelatin and fell off.  That's barely water resistant and certainly not waterproof.

For waterproofing I usually lean towards silicone (RTV, silastic, aquarium cement).  It works well in every application I've ever used it on.  But hot glue is so convenient and sets so fast that I kept thinking about it and wishing I could use it.  A few weeks ago I decided to learn more about hot glue.

There are many different kinds of hot glues used in industry and most of them are available in standard 11mm sticks, but I would much prefer to use the common household clear bluish or yellowish stuff if possible.  The standard home-use glue is made from EVA (ethylene vinyl acetate) plus various additives to adjust its characteristics.  Nearly every article I read says that "hobby" hot glue is waterproof, and indeed, EVA is a waterproof thermoplastic.  This disagreement with my experience really got me studying.  Something is not adding up here.  I learned about all kinds of hot glues and their characteristics and just as I was getting ready to compose emails with specific questions to hot glue manufacturers, I ran across the answer.

What happened to me six years ago was a "materials compatibility problem".  Six years ago, what I had done was to waterproof some electronics by building it inside of sealed PVC pipe.  The plan was to bring a couple of small insulated wires out through a small hole in the pipe and then seal it up by pumping hot glue into it.  This is the application that failed the immersion test.  Naturally I assumed it was the water that caused the failure, but it was not the water.  It was the PVC pipe material and especially the PVC insulation of the wires.  Actually, it was not the PVC itself but the pthallate plasticizer that's mixed with the PVC to make it flexible and not brittle.  PVC wire insulation has an especially high amount of pthallate.  Pthallate attacks EVA and causes it to decompose, and that's why my test failed.

Recently I've done more experiments with EVA hot glue.  Yes the material is waterproof.  I've had a glob of hot glue immersed in water for weeks and it's  unaffected.  I did another experiment where I used EVA to hot glue polyurethane to painted ABS.  It sticks like a weld.  You have to destroy the pieces to get them apart and it's unaffected by water.  So it all depends on materials compatibility.  I just wish I had known this sooner.

Tuesday, February 22, 2011

The IEC Power Cord Mystery

The ubiquitous IEC power cord has been a part of everyone's life for almost 30 years.  Nearly every device capable of running on different mains voltages and frequencies has one so the manufacturer can ship to different countries by simply changing the power cord to match the plug type used in that country.  Great idea.

Every device that needs an IEC cord comes with one from the manufacturer.  When a device is no longer wanted and we dispose of it, we always keep the cord "because it might come in handy".  Thus, there should be an ever-increasing number of IEC power cords in the world.  There should be billions of them in existence.  After 30 years of this, even the most non-technical person should have at least 20 cords stashed in their closet.  Yet this is not the case.  When you need one, there's none to be found.  Search the entire premises and there's not one unused IEC cord.

How can this be?  I've never thrown one away in my life.  Do unused cords simply vanish?  Do they automatically return to the manufacturer to be shipped to someone else?  Do little gnomes steal them at night?  Where are all my cords?

Tuesday, February 8, 2011

Simplest Way to Play Raw PCM Audio on Ubuntu: libao

There are a zillion ways for a Linux programmer to play audio through the sound card.  This is the problem.  There are many layers to the audio system, many ways to go, and most of them are very complex because multimedia is very complex.

But what if you have the simplest of all cases where you have a buffer in memory containing raw PCM samples, ready to play, and you just want to pump the data out to the sound card and play it at a certain sample rate?  In many cases you're looking at hundreds of lines of code, writing your own plugin, etc.

After two days of asking questions, Googling, and reading, I finally found what I was looking for.  Libao is part of the standard Ubuntu distribution and it does the job without writing tons of code.

I found a couple of examples but both had problems compiling cleanly.  After using Synaptic to install the libao development files the following will compile cleanly on Ubuntu 10.04 using the gcc command shown in the comments below:

  * ao_example.c  
  *   Written by Stan Seibert - July 2001  
  *   Modified slightly by Phil Landmeier - February 2011  
  * Legal Terms:  
  *   This source file is released into the public domain. It is  
  *   distributed without any warranty; without even the implied  
  *   warranty * of merchantability or fitness for a particular  
  *   purpose.  
  * Function:  
  *   This program opens the default driver and plays a 440 Hz tone for  
  *   one second.  
  * Compilation command line (for Linux systems):  
  *   gcc -lao -ldl -lm -o ao_example ao_example.c  
 #include <stdio.h>  
 #include <string.h>  
 #include <ao/ao.h>  
 #include <math.h>  
 #define BUF_SIZE 4096  
 int main(int argc, char **argv)  
     ao_device *device;  
     ao_sample_format format;  
     int default_driver;  
     char *buffer;  
     int buf_size;  
     int sample;  
     float freq = 440.0;  
     int i;  
     /* -- Initialize -- */  
     fprintf(stderr, "libao example program\n");  
     /* -- Setup for default driver -- */  
     default_driver = ao_default_driver_id();  
     memset(&format, 0, sizeof(format));  
     format.bits = 16;  
     format.channels = 2;  
     format.rate = 44100;  
     format.byte_format = AO_FMT_LITTLE;  
     /* -- Open driver -- */  
     device = ao_open_live(default_driver, &format, NULL /* no options */);  
     if (device == NULL) {  
         fprintf(stderr, "Error opening device.\n");  
         return 1;  
     /* -- Play some stuff -- */  
     buf_size = format.bits/8 * format.channels * format.rate;  
     buffer = calloc(buf_size,  
     for (i = 0; i < format.rate; i++) {  
         sample = (int)(0.75 * 32768.0 *  
             sin(2 * M_PI * freq * ((float) i/format.rate)));  
         /* Put the same stuff in left and right channel */  
         buffer[4*i] = buffer[4*i+2] = sample & 0xff;  
         buffer[4*i+1] = buffer[4*i+3] = (sample >> 8) & 0xff;  
     ao_play(device, buffer, buf_size);  
     /* -- Close and shutdown -- */  
  return (0);  

Wednesday, February 2, 2011

Major Flickr Accident

Photo sharing service Flickr accidentally deleted a user's photos.  Five years and 4,000 photos are gone and there's no recovery, no backups.  Read about it here:

If just the thought of this gives you a wave of nausea, there are solutions to the problem such as this one:

How about your Gmail account?  Google Docs?  I have 1,500 spreadsheets and 300 important documents plus miscellaneous stuff stored on Google Docs and can't afford to lose them.  What about all your posts on Facebook?  A service like Backupify can give you a level of safety and control you don't have now.

Google catches Microsoft with pants down, copying search results

Funniest thing I've read in a while:

Friday, January 14, 2011

Install Fails, Files Out of Date

The other day I was trying to install the software for a digital oscilloscope and ran into a problem.  The software is written in Visual Basic.  The installer said that "certain files are out of date" and offered to update the files before rebooting the system.  If you answer no, the installer quits.  If you answer yes, the system reboots and when you run the installer again it repeats the "files out of date" message.  There's no way to get past this, but I have found a workaround.  The problem occurs with both the Microsoft and Nullsoft installers, and it happens on Windows 2000 and XP.

How To Fix It

Your installable program will consist of at least a setup program, a CAB file, and a file called SETUP.LST.  SETUP.LST is structured like an INI file so you can open it with a text editor like Notepad.  Don't use a word processor.  If you open SETUP.LST you'll see the file organized into sections that look something like this:

SetupText=Copying Files, please stand by.

[Bootstrap Files]
File1=@VB6STKIT.DLL,$(WinSysPathSysFile),,,7/15/00 12:00:00 AM,101888,
File2=@COMCAT.DLL,$(WinSysPathSysFile),$(DLLSelfRegister),,5/31/98 12:00:00 AM,22288,4.71.1460.1
File3=@asycfilt.dll,$(WinSysPathSysFile),,,3/5/10 7:37:40 AM,65536,5.1.2600.5949
File4=@olepro32.dll,$(WinSysPathSysFile),$(DLLSelfRegister),,4/14/08 6:42:04 AM,84992,5.1.2600.5512
File5=@oleaut32.dll,$(WinSysPathSysFile),$(DLLSelfRegister),,4/14/08 6:42:04 AM,551936,5.1.2600.5512
File6=@stdole2.tlb,$(WinSysPathSysFile),$(TLBRegister),,4/13/08 10:12:08 PM,16896,5.1.2600.5512
File7=@msvbvm60.dll,$(WinSysPathSysFile),$(DLLSelfRegister),,4/14/08 6:42:02 AM,1384479,

The problem is in the [Bootstrap Files] section.  Several of the files in that section are unnecessary.  If you comment them out (by putting a semicolon at the start of the line) your install will proceed without any problems.  You may have to experiment with how many files to comment out but in general you should keep the first two or three files in the list and comment out the rest.  The configuration that worked for me is below.  Your SETUP.LST file will probably be slightly different.

[Bootstrap Files]
File1=@VB6STKIT.DLL,$(WinSysPathSysFile),,,7/15/00 12:00:00 AM,101888,
File2=@COMCAT.DLL,$(WinSysPathSysFile),$(DLLSelfRegister),,5/31/98 12:00:00 AM,22288,4.71.1460.1
;File3=@asycfilt.dll,$(WinSysPathSysFile),,,3/5/10 7:37:40 AM,65536,5.1.2600.5949
;File4=@olepro32.dll,$(WinSysPathSysFile),$(DLLSelfRegister),,4/14/08 6:42:04 AM,84992,5.1.2600.5512
;File5=@oleaut32.dll,$(WinSysPathSysFile),$(DLLSelfRegister),,4/14/08 6:42:04 AM,551936,5.1.2600.5512
;File6=@stdole2.tlb,$(WinSysPathSysFile),$(TLBRegister),,4/13/08 10:12:08 PM,16896,5.1.2600.5512
;File7=@msvbvm60.dll,$(WinSysPathSysFile),$(DLLSelfRegister),,4/14/08 6:42:02 AM,1384479,

Good luck!