Issues with MC Cthulhu in an arcade stick

[Skivt68]

I just sent a message to Toodles on SRK about this but then I realized that he hasn't been active for 4 months.  Here is what I sent if anyone can help me with this:

Hi, I just bought one of your MC Cthulhu boards from focus attack and, first of all, it's great, but I have been having some problems. I have all of my buttons ground connections daisy-chained and connected to one GND terminal, and a separate daisy-chain for my joystick, hooked up to a second GND terminal. When I hook up my joystick it is fine, the same goes with hooking up buttons to 1p, 2p, and 3p. When I use more than 3 buttons, however, some buttons will resister as pressing two different buttons at once, such as pressing 2p registers as pressing 2p and 2k, or pressing start registers and pressing start, 3k, and 3p. Also, whenever I connect any button to the terminal for "select" the board no longer registers with the computer and when i remove the wire form the "select" terminal, the board works fine again and registers, save for the multiple button issue mentioned above. I was wondering what you thought the issue might be, and if it can be diagnosed/fixed or if I would just be better off returning it to Focus Attack to exchange it for another MC Cthulhu? I am in no rush, as I still have about 26 days left in the return/exchange window, but any answer would be much appreciated.

Thank you for your help and have a nice weekend,
Skivt68

I don't have pictures of my wiring right now, but I can take some if your guys think it might help in diagnosing the problem.

[PL1]

When in doubt, include pics. 

Have you ohmed out your harness wires and buttons to check for shorts?
(disconnect from the encoder first so you isolate your testing to just the wires and buttons)

Ohm from one wire to another using a pattern like this:

Gnd-1, Gnd-2, Gnd-3, Gnd-4, Gnd-5, Gnd-6
1-2, 1-3, 1-4, 1-5, 1-6
2-3, 2-4, 2-5, 2-6
3-4, 3-5, 3-6
4-5, 4-6
5-6

With no buttons pressed, it should show an open on all readings.
-----------
Connect one lead to the ground wire.

Hold one switch closed and ohm to each wire in turn -- it should show an open on all readings except the wire connected to the switch you are holding closed.

Repeat for each switch in turn.


Scott

[Skivt68]

Thank you for your advice, but I'm not sure i know what you mean by "ohm-out".  EDIT: I somewhat get it, but by 'ohm-out" do you mean continutiy test on a multimeter?
I do have plenty of pics of my highly unorthodox lighting controller gone arcade stick. 
It used to be in my school's auditorium, but we didn't need it anymore,and being one of the head student engineers I sometimes get surplus gear.  I took out the dimmer and put in the joystick while flipping the buttons to make them easier to press. After i get the buttons to work, i am planning to add three square buttons from the old dimmer board back into the holes by the joystick as select, home, and start.  that is on hold while i find a mounting solution for those.  In the end, 8 of the 9 grid buttons will be hooked up in addition to the three buttons that will be near the joystick as the MC Cthulhu has room for only 11 buttons and 9+3=12, so i'll just not hook up one of the 9.







The Joystick (white is ground, the rest are the 4 signal wires.  This works fine.)


how I have it working right now (sometimes doenst work if I nudge the encoder/wires as i'm screwing it back together.)


My ridiculously tedious wiring job.  I used a multimeter to check the daisy-chain, if I press the first button in the chain (top right) the ground is active on the last button in the chain (bottom left).  The pcb around the buttons is not part of the setup and is just there to hold them in place.  Should i cut the traces on it just to be safe?  (for this, the white and black are daisy chained ground and the green ad red are the various signal wires.

[PL1]

by 'ohm-out" do you mean continutiy test on a multimeter?

Yes.  Continuity means a resistance of less than ~2 Ohms between two points.

I phrased it that way since not all multimeters have a setting for "Continuity", but they all have a setting for "Ohms".

highly unorthodox lighting controller gone arcade stick. 

You are truly gifted in the fine art of understatement. 

how I have it working right now (sometimes doenst work if I nudge the encoder/wires as i'm screwing it back together.)

You will need better mechanical and electrical connections using solder and/or quick disconnects.

Also, you'll want to insulate the metal case to prevent accidentally shorting the wires -- some spray paint or adhesive shelf-liner should do a decent job.

The pcb around the buttons is not part of the setup and is just there to hold them in place.  Should i cut the traces on it just to be safe?

The red PCB holding the buttons looks like it has a built-in encoder circuit -- probably a matrix-type encoder -- that decodes buttons wired in rows and columns like this Atari 2600 Keypad.



if you can cut the traces so the only things electrically connecting to each button are your ground wire and signal wire, that would be the best way to proceed short of desoldering the buttons and remounting them on perfboard.


Scott
[/post #2000 ]

[Skivt68]

Thank you so much!  Congrats on 2000 posts! 
I'll have to get to work mutimeter testing tomorrow.  Also, cutting all of the traces near the switches by scoring the board with a utility knife should work, correct?
 I think that this might be the underlying issue, as the 7 button (labeled "S7") had it's solder pads lifted from the board when I was desoldering it to test which leads were ground and signal. (which in retrospect i shouldn't have desoldered it for, but oh well)  The middle row of buttons that are hoked up, and the 7 which had its pads lifted, all work because they are in a row, not columns, making them electrically separate. 
One last question, why do you think the computer will not recognize the board when "select" is connected, regardless of the button it is connected to?

[PL1]

cutting all of the traces near the switches by scoring the board with a utility knife should work, correct?

It SHOULD work, but it's much better to remove a small section of the trace to ensure there's no chance of reestablishing an accidental connection if something brushes against the cut trace and pushes it into contact again.

I remember years ago working on this one PCB.

When you'd test it with the ohmmeter, it looked normal, but when 5v was applied, it shorted.

Over the next month, several others attempted to repair the board, but weren't able to isolate the problem.

Turns out that one of my coworkers was a bit sloppy when doing the initial trace repair/soldering and epoxy backfill.

I found a series of microscopic solder globs in the epoxy that provided a path for the higher amperage 5v supply to arc over, but not for the very low amperage ohm-meter.

Once I carefully excavated the solder globs and re-epoxied it, the board worked fine.

One last question, why do you think the computer will not recognize the board when "select" is connected, regardless of the button it is connected to?

Does it still do that when you disconnect all other control wiring and use a piece of wire to short select to ground? (removes all other variables and tests just that input on the board)

I'm not familiar with the MC Cthulhu encoder, so I will leave any further in-depth troubleshooting to others.   


Scott

[Skivt68]

Alright, so i put only 1 button on the select and viola! it worked.  Now i just have to deal with the issue of cutting the traces. Do you think that instead of cutting the traces, (as they are tight and plentiful) I could remove all of the IC from their sockets?  I assume that that is what is keeping them encoded/linked.  i will try that now and get back to you guys if it works.  Let me know if you think that that would disrupt it though, as i can also remove any components on the outside of the board.
EDIT:
Update: removed to ICs, didn't do anything to help with the column issue, but now the select button works.  I'll get out the soldering irons and flush cutter later and remove all components on the board save for the buttons and if that doesn't work, i'll try to cut the traces.

[Skivt68]

Hey, sorry for the double post, but, i have removed all components save for the switches from the board, but in the process, i had to cut some joined resistor banks.  I didn't touch them much, but I was wondering if it is harmful to gut a resistor/ceramic capacitor.  I will have an update soon after I test it  just continuity tested the board, Scott, your are indeed right, the signal wires of all of the buttons in a column are indeed joined by traces.  i will be cutting the traces once I have the time.

[PL1]

If you cut all the row connections or all the column connections, you can use the uncut ones as a daisy-chain ground -- less soldering that way.   


Scott