keyboard hack and electrical resistance

[mravett]

I'm working on a keyboard hack  (yes I know an Ipac or hagstrom encoder would be MUCH easier, but the budgetary commitee/my wife is not giving much funding to the project  ).

Anyway, due to the dodgy nature of my soldering job, some of the pins on my encoder are just barely connecting to each other.  How much resistance between two pins is enough to keep one from firing them both.  On most of these neighbor pins I see something like 500kOhms, but it varies from about 100kOhms upto "infinite" on those few I soldered correctly  

I've been scraping the areas between pins out with an awl, but again due to the weakness of some of these solder bonds, I'm afraid to do too much.  I've already popped a couple off.  I plan to put large amounts of epoxy around my soldering to strengthen it once I'm done.

[Chris]

It might be worthwhile to desolder the bad joints and re-solder them... a cheap desoldering pump from Radio Shack will do...

--Chris

[mravett]

Actually the problem is that I'm had to effectively surface mount my wires to small fingers of copper on the keyboard control board.  Because of this, it's almost impossible to get a decent solder weld and I had to use a silver-bearing paste for my solder to get into the small places.  In a couple places apparently a miniscule amount of this paste is connecting between two pins.  I've been able to scratch most of this away, but on a couple I can't quite seem to get all of it.

I think I've got the worst up to at least half a MOhm now, with the rest showing "infinite" on my multimeter.  My main concern is that while epoxying these leads down, a small metal dust particle will bridge two pin and bring down the resistance to something in around 100 KOhms.  

[RandyT]

It's actually much easier to solder the wires directly to the pins on the decoder chip, unless it is surface mount.  Most keyboards use a big 40+ pin chip and the pins are pretty accesible.  You could also go to the bottom of the circuit board.

But in your current case, I wouldn't trust ANY connection between the pins to work properly.  It may work once in a while, but certain button combinations might lead to erratic behaviour.  This will all depend on what type of encoder the keyboard uses and and how much voltage it requires to think that a key has been pressed.

If the PC board is a single layer (most of them are) get yourself an Xacto type knife with a sharp point and thin blade.  Carefully dig down into the board (not too deep!)  between the crossed connections.  Using a high power magnifying glass or jewelers lupe can help here as well.  If you see a little channels between the traces and no shiny stuff, you should be all set.

RandyT

 

[mravett]

The decoder chip is surface mounted under a big black glob unfortunately...

I've scratched out little troughs between all of the offending pins already and it seems to have worked for most of them.  I haven't checked in a while but there was still a 1.5-2MOhm connection between two of them, but I think one of those two pins was one that I can't use anyway now so it's not a big problem.

Actually that leads me to another question.  I mapped out the matrix on my old OCed p-366/550 and found some of my keys were repeated on the matrix (two x's, two g's etc.)  When I hooked the encoder to my laptop to test my mapping I found that one row and one column, both of which had mostly repeated keys in them, didn't work like normal.  For example, instead of getting a 'g' the laptop would act like I pressed the "Mute" button.  Can anyone explain why my two computer's disagree on input from a keyboard encoder?  Is my desktop just turning a funky scan-code into letters for the heck of it?

Ryan