severe grounding problems

[matrigs]

i am really getting confused with grounding - i guess i'll make my next box just out of encoders and buttons without any analog inputs because this s driving me crazy !!

i have 8 rows with 8 pots connected. the grounding and the +5 is connected serial horizontal in every row so when i connect the ground and +5 to just one - the whole row gets it.

but as i had problems with grounding before i've gone further and made 3 bridges vertical so i have connected all rows together.

now i have used 3 different grounding points on the core (one of the ain pins, one of the din pins and one at some random place) and guess what - random midi messages !!??

my question is - is so much grounding like i described above actually necessary? because maybe i'll ground myself to death and the reason will be something much more simply.

just tell me how much you have to ground your pots to make your box work.

[Sasha]

I`m not sure I understan exactly what is confusing you, but there is only one ground! It doesn`t metter where you took ground from. I use to solder piece of wire directly on the closest PCB if I need it for buttons or LEDs or whatever...

Only thing you have to know is that you need to ground all unused analog inputs to the ground no matter you are using muxed or unmuxed mode...

[matrigs]

it's confusing me that i have made a few dozen bridges to different grounding points so i'm 100 % sure that every single pot is grounded.

i HAVE all ains closed so i still don't get it.

so you have like all your pots linked serial and then just grounded them with one wire all ?

[Jidis]

i HAVE all ains closed so i still don't get it.

That's what I was going to suggest going after next (the path to the analog ins), but I guess you've ruled that out. I feel your pain on it though. I did a box a while back with a single, double sided board holding 8 pots, 24 buttons, and 32 lights, and the pots almost made me throw it off a bridge. I did the same thing with jumping additional new wires to ground and 5v points, as well as replacing a couple pots which were probably OK. Works fine now, with the exception of one pot that will throw out a jittery 127 sometimes at minimum (zero). Seems like basic "continuity" checks don't always cut it with some stuff.

Are all your prefs in the app set right too (AIN mux/unmux, no. of pots,etc.)?

Hope you get it fixed,

George

[matrigs]

i'm using the bare 64 ain 128 din 128 dout app so everything software wise should be fine.

there is one last option actually that i could check but this is quite work intensive and will take me a couple of hours.

the thing is - to make my connections more comfortable i don't solder the output pins of the pots directly to the ain boards but i use some special cables that are used in motherboards. one one side you have a connector like this:

and on the other side 3 cables. it's very comfortable because you have just to solder one end and then you put it on the according pins of the ain.

however - i remember that sometimes the cables were faulty and i had to replace it.

the problem is now that i have such a great amount of midi junk that it's impossible for me to check which cables are faulty and have to be replaced. i mean i can actually try to solder all pots directly to the ain without those cables but this will be a very time consuming part and if it won't work i assure you this midibox is going to crash.

it's weird though... is this problem so common? isn't there anything that can be done?

[Jidis]

Yeah, I'm in a habit of sticking mux and shift register chips on the boards with my controls, so I use salvaged internal computer cables when I can too.

Man, I skimmed through those details and saw the number '8'. Didn't realize you were talking 8x8! :o

Is there no way to isolate parts of the chain or anything to see how many work without the random mess? From the software angle, I know on that 8 pot box I've got, I'll sometimes compile for "number of pots = 1" and it ignores all the others, but I'm not sure if that would block out a more serious hardware issue. Do the random messages point to any one component's actual messages or are they not even stuff the box is supposed to send?

Good luck again! - off to fight with my own box now ;D

George

PS- We need to get some sort of Batman beacon light we can turn on to call Thorsten on stuff like this. ;)

[matrigs]

i guess i'll try to open just one ain and then check every singlle knob.

it would be cool if there would be some software app which would automatically shut down the knobs that got undgrounded. but i know that's impossible.

but one thing i'm thinking about - maybe a little "smooth" plugin could do wonders for a few midiboxes. like with your issue that it sometimes bursts 127 when you turn a knob beyond 0. the app wouldn't let peaks like this through.

of course the latency would be a little higher but not much.