I got mine last week and just got around to getting it in my cabinet. I have a Namco Cyberlead Japanese sitdown cabinet with 2 sticks and 6 buttons per player. That really takes up all of the space on the cp and I really don't want to drill through the plastic to add a mode switch button. My temporary solution for mode switching is to hang a button through the front by the coin door to use as a dedicated mode button. It's ugly but it works.
I have original Midway versions of the sticks from an NFL Blitz game. To get the sticks in I had to shorten them a bit as the full sticks hit the bottom of the panel and I couldn't close it or move the sticks. I thought I was going to have to cut the sticks but on looking closer, noticed that there was no spacer in the top like many of the Happ comp or ultimates. The stick extends past the bottom of the sliders about 1/2" so my first inclination was to cut that off and cut a new slot for the c-ring. After noticing the lack of a spacer I found all I had to do was cut about 1/2" off the long white bushing and put the little piece at the top of the stick shaft and the longer piece back on the bottom thus raising the stick enough to clear the panel and provide free movement. I cut it just short enough that the bottom bushing still makes full contact with the sliders through the full range of motion.
Long term, my goal is to have mode switching accomplished by the pc. The parallel port is the easiest way to accomplish that since it does not appear that the GP-49 can switch modes using any sort of USB communication. That's my one small gripe with the product, the use of a dedicated button to accomplish mode change. Anyways, it's not an insurmountable problem but rather an "Annoyance".
Hook up was fairly straight forward. I bought the solder version an made my own harness using a 12 position .100 header for connection to the sticks and straight wire runs for the switches. Here's another place I would actually like to see an alternate method of mode control available. because the GP-49 has to use it's buttons for a mode change, I am forced to use the buttons on the CP somewhere. Whether it be a dedicated mode switch button and dual use standard buttons or even just dedicating a set of buttons for mode change. Add in a second stick and you are now required to have 1 button for mode change (can be tied to both) and 1 button on each stick for each mode you want selectable. Since I have 6 buttons for each player (7 if you count player 1 & 2 start) this isn't a huge deal but having to re-setup my FE, MAME and the other programs to recognize the new control set is more than I was hoping to have to do. Not to mention that not every emulator or front end is as configurable for controls as MAME is. Again, not a terrible thing but another "Annoyance". Since I already have everything up and running using a JPAC, I was hoping to have to reconfigure just the joystick directions rather than all controls. The shortcoming for me is that the FE I am using supports most all functions using a gamepad except "Exit". Kind of a neat one to have when you are done playing a game. Of course, I can use RBJoy, joytokey, AutoIT or a couple of other utilities to map the buttons to keys. I just don't want to

I spent many hours eliminating stuff from windows xp, trimming my running progs, startup items, etc and just don't want to run more than I have to. Well, now I have to run joy2key or something like it until GameEx can exit from a gamepad by itself or I get the mode switching under PC control.
So how does it work? In a word, Awesome. I removed 2 t-stick+ balltops that are about a month old and am glad I did. Even with lighter cherry switches the t-sticks were just not to my liking. My kids noticed a big difference right away and all 3 said something about it. The 4 way & diagonals work great. Q*Bert is very playable. Pacman works better than with the t-sticks in 4 way mode and in 8 way, Robotron Rocks! Worth every dollar and hour of time I spent in installing and configuring. Even with the "Annoyances" for me it's a solid 10 on the 1 stick does all front.
PC Control. Here's where it gets fun. Parallel port control *could* be very easy. Especially since Randy has offered to do a variation on the chip to allow dedicated mode switch buttons. Even without dedicated mode switch buttons it is still easy to do, it may just cost more. Here are the scenarios I have put together. I expect that the various front ends could implement support for automatic mode switching very easily using the parallel port since it puts out +5V for a logic 1 and 0V for a 0. It's just a matter of sending the right combination of set bits to the port. Even in Win2000 and XP there are freeware drivers available that will let you send bits directly to the port.
1. Standard config using dedicated mode switch button. For this method, either a simple inverter circuit or a relay circuit is needed to change the way the logic works. Button presses are seen as logic 0 (ground) by the GP-49 and the parallel port starts off with all 8 data bits at this state. An inverter for each bit would make the logic work by setting the data bit active the button is pressed. assuming that data bit 0 is used as the mode switch button you only have to set bit 0 and any of the 1-7 to get to all 7 modes. After the mode is set, the bits can be cleared again. Another alternative to using a homebrew setup with an inverter circuit would be to use a parallel port relay board. They run about $30-$55 depending on where you get them and whether they are assembled, in a case, etc. Best price I have found is
http://www.electronickits.com/kit/complete/elec/ck1601.htm. The advantage of using the relay board is that it also provides some isolation between the PC and the GP-49. The LEDs on the board are only active when the bits are set so not very useful as is for providing mode info. Of course, you could expand that and add a latch circuit to the LED outputs and use that for mode ID as it would hold the last button pressed and keep the LED on. Should be easy to add if you buy the kit version.
2. Same as above but instead of using an inverter it is connected directly from the parallel port to the GP-49. The FE or whatever program you use to send the mode switch commands with has to set all 8 bits of the parallel port to logic 1. Mode switching is accomplished by sending logic 0 to data bits 0 & any of 1-7 to set the mode then back to all logic 1 (no need to keep them pressed).
3. Using a GP-49 with dedicated mode switch buttons (special chip from Randy). This is probably the easiest depending on how said support is written. It would depend on how Randy does this. I assume, based on what is in this thread, that this support would only use 7 buttons, 1 for each mode and would not require a dedicated mode switch button. Instead, each input would activate the mode. Here's where the "it depends" come in. If it is just the removal of the mode switch button and whenever a button press is detected on on one of the 7 mode pins then an inverter is needed (or a relay board as in #1). HOWEVER! IF Randy were to make this chip support "Normal" (not reverse) logic all that would be needed is to connect the parallel port data bits 1-7 directly to the GP-49 and whatever bit is set is which mode it is in. All you need is a parallel cable in this case! One thing I can think of that Randy might also want to do in this case would be to only accept the first active input as active in the case where 2 inputs were set active at the same time. Maybe he can protect us from software errors where a dummy like me tries to switch modes but doesn't clear the mode bit I've previously set

Add an LED driver board for even more geek excitement

Enough for now. There are my first impressions, my trials, my tribulations and a couple of ways to make real use of controls.dat.......
Sorry for the long winded post but I thought someone might pick info out of it that is useful.
Toonces