'Ghost' Key Presses

[Klomper]

I have almost completed my conversion of a nonworking NARC to Mame (will hopefully update my project page soon), with everything up and running and just doing some final setup and cleanup.  In the testing I have found that I am getting some ghost key presses. I had this almost identical setup in a 'temporary' panel for over a year and never had a problem.

I hoping people can point me in the right direction of what to check so I don't waste to much time trouble shooting.

Here is hopefully all the relevant information
1. IPAC-2 PS/2
2. 2 joys/w 7 buttons, 2x coin, P1/P2 start, ESC and Pause
3. Happens at least one time a game in 1942 (random loop) or other shooters where a bomb is randomly released. Did not notice a problem in NBA Jam or TMNT.
4. Windows XP
5. Stripped CAT 5 for the wiring (new to this setup)
6. All noticed random presses have been P1 B2 except one time it was Pause.

So far I have:
1. Resaved my configuration to the IPAC
2. Made sure P1 B1/B2 wires did not have contact.

[miles2912]

I would go through the checklist here and see if it helps.  Towards the bottom of the page is the troubleshooting.

http://www.ultimarc.com/ipac2.html

[ultrageek]

I wonder if the tristed pair wiring in the cat 5 is inducting a singal into the other pair(s).
I'v seen stranger. I know that this is extremely low voltage and shouldn't be an issue, but since the Cat 5 is new to the design, it's a likely contributor to the problem

[CrazyKongFan]

Isn't that why the twisted pairs in cat 5 are twisted in the first place, to prevent interference?

[bkenobi]

You are not inducing a signal by a button press.  I'm not an EE, but I know that inducing a current requires a high current on the source side.  You have (essentially) 0 amps running through the buttons, so you will have 0 amps running down any other conductors.  You may have 5 volts, but it's got (essentially) no current so you can't induce with it.

[Klomper]

My thoughts were the wires as well and I did some additional testing last night.

Did not work:
1. CP has two sets of buttons 1 & 2 at 8way and 4 way stick and it happened using either set. Both sets are wired on a set of twisted pair.
2. Thought I may have nicked the wires when removing the jackets.  Looked OK but added some electrical tape at the point to be sure.
3. Removed the quick disconnect from both button 2
4. Removed 4way b1/b2 wires from IPAC

Works but not the 'real' fix:
Mapped 'l-ctrl' to 1sw7 input on the ipac and 'c' to sw1 input since b7 is not used.  Was able to complete 2 games of 1942 without false loops. (these buttons are wired on the same cable but different twisted pairs)

Another note in IPAC test mode l-ctrl sometimes registers as r-ctrl? In fact when I switched the mapping and pressed l-ctrl on my keyboard it registered as r-ctrl the first time.

[Beretta]

i'd be SHOCKED if it's the ethernet.

wire normally used in ethernet is cat5+, very good quality wire.

in some cases people transmit power to power remote devices using PoE (power over ethernet) which calls for 48 volts.

the arcades controls calls for MUCH less then that usually around 5 volts or less, and like someone said almost no amp's


if i had to guess i'd say it's a wiring mistake.. or possibly a stray strand of wire.. you know need a tiny strand of wire to be on the loose.. htis is a problem when soldering also a single strand you missed creating a short.

i'd also check mame "input" settings... is it at all possible somehow mame's button configuration is off? you said you did not notice it in tmnt or nba jam