Most of you should already be familiar with this problem. Randomly during play, both of the trigger buttons RT and LT will just stop working until the stick is replugged. Annoying, and you can never tell when it will happen, and no one knows how to deal with it. It'd be nice to change that.
The reason I'm starting this thread is because it's come up a couple of times in the past couple of days, so it's gotten me thinking about the problem. Thinking led to ideas, and now I finally have something solid to offer.
In short, I can now recreate the problem on whim, and Im sure any of you techs can do the same. If we can recreate the problem, we can solve it. I'll spare the random stuff I tried before.
Recreation: In order to recreate the problem, I had to apply a voltage of 8v to the trigger signal line. I used a CR2032 coin cell battery, put the negative side connected to the USBVCC point (+5v) from the main board, and lightly brushed the signal lines for RT and LT against the positive side of the coin cell. This quickly placed a voltage of 8v on the signal line. When I did this to the RT signal line, the same 'triggers no workie' problem occured. When I did this to the LT signal line, all buttons, triggers, and directions stopped working, but the controller remained connected to the Xbox360. Replugging resolved all problems. (Applying -3v to signal: no problem, so I believe it's too high voltages causing the problem, not too low. )
Recreation prevention: Using a 10k ohm resistor in series with the 8v source stopped the issue from occuring. Using a 3.0v diode (white LED actually) as a shunt for the signal line did NOT prevent it from occuring. But, in case the charge buildup causing this problem is on the board and not from the buttons or external static through the case, I'd still feel a lot better about a shunt diode being in place as well. A 'belt and suspenders' solution; you may only need one, but it doesn't hurt to have both.
Proposed solution: Use a shunting diode of 3-5volts (zeners and white or blue LEDs are acceptible, zeners are WAY recommended over other diodes though. Its kinda what they're made for.) between trigger signals and ground, and a resistor of 10k ohm in series with the signal to the buttons.
The closer to the 'black blob' on the main board the better, so that best and cleanest place to put the resistors is on the BE3 and BE4 surface mount pads on the top of the main TE board. Use an Xacto or VERY fine Dremel tip to cut the trace between the two halves, and solder a 10k ohm, 0603 size resistor to the pads over the now-cut trace. 0805 is too large, and frankly, 0603 is cutting it close. 0402 is probably the best fit, but sucks to solder that small.
If you can't or won't use the surface mount spots on the top of the board, you can put the resistor on the distribution board. Cut the trace between the CN connector and the where the QD is soldered, and solder the resistor in place, one leg on the botton of the CN connector pin, other left to where the QD is soldered. Remember, do this twice, once for each trigger.
The shunting diodes are recommended, but since I cant tell where the charge is building up 100%, I cant be certain if they will help or not. So please do it anyways, and mark it up to 'Toodles says it might help'. The best recommendation I can provide is to install it on the bottom of the button distribution board, between the two large blobs. If using a zener diode like recommended, the cathode end must be connected to the signal, and the anode to ground; yes, this is counter intuitive. If you're using a normal diode or LED, cathode end must be connected to the ground, anode to the signal. The easiest way to do this I can think of is between the two QDs on the bottom of the distribution board.
Suggested, but not tested, parts for this:
Zener diode shunt: http://search.digikey.com/us/en/products/1N5226BTR/1N5226BFSCT-ND/1992218
Resistor, surface mount 0603: http://search.digikey.com/us/en/products/RMCF0603JT10K0/RMCF0603JT10K0CT-ND/1943191
Resistor, surface mount 0402: http://search.digikey.com/us/en/products/RMCF0402JT10K0/RMCF0402JT10K0CT-ND/1942936
Resistor, through hole for non-surface mount: http://search.digikey.com/us/en/products/CF18JT10K0/CF18JT10K0CT-ND/2022766
If anyone with more analog experience has anything to add, I'm all for hearing it, good or bad. I may be off base, but since I can recreate the symptoms, and the fix prevents it from occuring, Im fairly confident this will prevent the issue commonly seen in the wild.
Edit 5/17/2010: Linked and description updated to 10k ohm resistors from 4.7k ohm based on report from beefmissile.
The reason I'm starting this thread is because it's come up a couple of times in the past couple of days, so it's gotten me thinking about the problem. Thinking led to ideas, and now I finally have something solid to offer.
In short, I can now recreate the problem on whim, and Im sure any of you techs can do the same. If we can recreate the problem, we can solve it. I'll spare the random stuff I tried before.
Recreation: In order to recreate the problem, I had to apply a voltage of 8v to the trigger signal line. I used a CR2032 coin cell battery, put the negative side connected to the USBVCC point (+5v) from the main board, and lightly brushed the signal lines for RT and LT against the positive side of the coin cell. This quickly placed a voltage of 8v on the signal line. When I did this to the RT signal line, the same 'triggers no workie' problem occured. When I did this to the LT signal line, all buttons, triggers, and directions stopped working, but the controller remained connected to the Xbox360. Replugging resolved all problems. (Applying -3v to signal: no problem, so I believe it's too high voltages causing the problem, not too low. )
Recreation prevention: Using a 10k ohm resistor in series with the 8v source stopped the issue from occuring. Using a 3.0v diode (white LED actually) as a shunt for the signal line did NOT prevent it from occuring. But, in case the charge buildup causing this problem is on the board and not from the buttons or external static through the case, I'd still feel a lot better about a shunt diode being in place as well. A 'belt and suspenders' solution; you may only need one, but it doesn't hurt to have both.
Proposed solution: Use a shunting diode of 3-5volts (zeners and white or blue LEDs are acceptible, zeners are WAY recommended over other diodes though. Its kinda what they're made for.) between trigger signals and ground, and a resistor of 10k ohm in series with the signal to the buttons.
The closer to the 'black blob' on the main board the better, so that best and cleanest place to put the resistors is on the BE3 and BE4 surface mount pads on the top of the main TE board. Use an Xacto or VERY fine Dremel tip to cut the trace between the two halves, and solder a 10k ohm, 0603 size resistor to the pads over the now-cut trace. 0805 is too large, and frankly, 0603 is cutting it close. 0402 is probably the best fit, but sucks to solder that small.
If you can't or won't use the surface mount spots on the top of the board, you can put the resistor on the distribution board. Cut the trace between the CN connector and the where the QD is soldered, and solder the resistor in place, one leg on the botton of the CN connector pin, other left to where the QD is soldered. Remember, do this twice, once for each trigger.
The shunting diodes are recommended, but since I cant tell where the charge is building up 100%, I cant be certain if they will help or not. So please do it anyways, and mark it up to 'Toodles says it might help'. The best recommendation I can provide is to install it on the bottom of the button distribution board, between the two large blobs. If using a zener diode like recommended, the cathode end must be connected to the signal, and the anode to ground; yes, this is counter intuitive. If you're using a normal diode or LED, cathode end must be connected to the ground, anode to the signal. The easiest way to do this I can think of is between the two QDs on the bottom of the distribution board.
Suggested, but not tested, parts for this:
Zener diode shunt: http://search.digikey.com/us/en/products/1N5226BTR/1N5226BFSCT-ND/1992218
Resistor, surface mount 0603: http://search.digikey.com/us/en/products/RMCF0603JT10K0/RMCF0603JT10K0CT-ND/1943191
Resistor, surface mount 0402: http://search.digikey.com/us/en/products/RMCF0402JT10K0/RMCF0402JT10K0CT-ND/1942936
Resistor, through hole for non-surface mount: http://search.digikey.com/us/en/products/CF18JT10K0/CF18JT10K0CT-ND/2022766
If anyone with more analog experience has anything to add, I'm all for hearing it, good or bad. I may be off base, but since I can recreate the symptoms, and the fix prevents it from occuring, Im fairly confident this will prevent the issue commonly seen in the wild.
Edit 5/17/2010: Linked and description updated to 10k ohm resistors from 4.7k ohm based on report from beefmissile.