From another thread:
Pic of a WS with onboard IDE resistors soldered in:
http://linux-hacker.net/websurfer/images/ide-resistors1.jpg
I would use wirewrap wire if I were to short across them. The pic indicates which way the resistors would need to be shorted.
Easy way to tell, one side connects to the connector, the other to the controller, when you find 5 or more like that you get the general idea.
Look at "RN11" for a example, four traces come from the IDE connector, jumper them across to the controller side.
R99 and R100 are pull-ups to +5V, so anything between 1K and 20K would be OK for what we are doing. That would translate to "102" up to "203" in little tiny part numbers.
The packs can be from about 22 ohm to maybe 100? 33 was a good choice someone suggested, the IO uses 22 on their IDE. They add RFI "protection" and maybe if you plug it in backwards with power on it will save you.
That translates to "220" up to "101" in little tiny part numbers.
You can find these on old dead $1 swapmeet hard drives and boards if you look. That is why I am saying that the values can vary, it's not all that important since the cable run is very short. This is all 5V logic signals.
The ones you see are 33 and 1K, but engineers have this stuff laying around.
COM1 is attached and working in the pic, with the addition of U10, it's the same as U17. Same pinout at the connector as a normal serial.
