www.megadrum.info

Hello all!

I am getting weird crosstalk problems on most of my inputs (an input triggers every other that has a cable connected, but not the disconnected ones). After a lot of testing, it so happens that on most of the pads (snare, all toms, some auxes, crash, ride) I'm reading a 10.9K - 11Kohm resistance between tip and ring (and between rings and tips of any two inputs which have this problem) with the digital board connected and no power. Without it, I measure no continuity (as I think it should be).

I'm 99% sure it's due to poor etching (hairline shorts on the digital board), since I've scrubbed everything with isopropyl alcohol a few times and checked visually and with a tester for shorts but could find none. The question I have is, on which pins of the digital to analog board connector should I check for this resistance, in order to find the hairline short (if it is indeed a hairline short)?

I don't want to have to re-etch the digital board pcb since I'm pretty sure it's just something really small and I'm too dumb to find it - and I risk having the same problem all over again. This is quite an expensive project here in Brazil, so I'm trying to avoid having to buy most of the components again (especially the IC and processor sockets) and I'm pretty sure I'd ruin them if I tried to salvage them again (they were on another digital board that had a track sucked out along with the solder).

If it's not a short, what could it be?

My build is the current Dmitri dual boards, latest firmwares with ATMEGA644@20MHz, PIC@12MHz, no 6n138 since I coudn't find it here. Everything else works, LCD, buttons, programming, inputs individually work very well.

It's likely that I'll find it sooner or later, but if someone has a light to shed on this please help, I'm dying to put this baby to work and make envy to all my drummer friends!

EDIT: ATMEGA644, not 622! Sorry!

Take out 4051 ICs out of their sockets.
Do you still get any reading for resistance between tips and rings with and without the Digital board connected?

Ok, tested it again.

With 4051 and digital board connected = around 11Kohm
With 4051 and digital board disconnected = around 50Kohm
Without 4051 and digital board connected = around 20Kohm
Without 4051 and digital board disconnected = around 60Kohm

Using common sense when measuring resistance between any two inputs with 4051's out I would expect that a tester would show not less than MOhm resistance if all BAT85s are soldered correctly (there is a path between any two inputs over two BAT85 connected in series and in reverse polarity to each other). But I might be wrong because different testers would show different resistance on the same non-linear devices (such as diodes) depending on what voltage they apply. In theory, there should be infinite resistance between any two inputs with or without 4051s in place.

There is indeed a very large (above 20Mohm) resistance between some of the inputs' rings and tips.
The ones that have low resistance are:
Hihat pedal ring and tip (pins 1 and 5 - lower than the other inputs, across 1 and 3 it's ok)
Hihat
Snare
Ride
Crash
Tom1,2,3 and 4
Aux1,4,5,6 and 7

24 inputs out of 32 (25 counting the hihat input). Too round a number to be a coincidence, so maybe it is related to tracks on the PCB or the BAT85s.

I've tested across two of the diodes in series and reverse polarity on several places of the board and I'm getting 90Kohm across them with the tester set to 20Mohm, 70Kohm with it set to 2Mohm and 42Kohm with the tester set to 200Kohm. Below this testing values I get above threshold resistance.
The inputs report different resistances when setting the meter to higher resistance measurement too (some 10Kohm over the BAT85s).

Might there be something wrong with my BAT85s?
Just to be clear if I have done something wrong, the diodes should have the end marked with a black stripe connected to each other for the ones near the 40 pin input and facing away between the muxes (as I have them now) or should they be reversed?

I'm guessing that maybe with 4851s instead of 4051s it could solve my problems, but the damn thing is that they're not available here!

mackanov wrote:Might there be something wrong with my BAT85s?
Just to be clear if I have done something wrong, the diodes should have the end marked with a black stripe connected to each other for the ones near the 40 pin input and facing away between the muxes (as I have them now) or should they be reversed?

Doesn't sound right. If you mean near the pin 40 of the inputs sockets, which is ground, then all diodes connected to the ground must have the black stripe facing away from ground. The analogue board-brd components.pdf in viewtopic.php?f=3&t=266 shows how the diodes must be placed.
To conclude, you must be sure that the diodes are placed correctly and they are in fact BAT85.

I'm guessing that maybe with 4851s instead of 4051s it could solve my problems, but the damn thing is that they're not available here!

4851 doesn't need BAT85s at all.

I guess I wasn't clear. I meant that between the input connector (the 40 pin connector) and the 4051s they are black stripe to black stripe and between the 4051s they are reversed, like this:

In this case it looks right. Anyway, viewtopic.php?f=3&t=266 is the ultimate reference for components placement.
Regardless, the resistance between the inputs is inconsistent so something is not right there.

Regardless, the resistance between the inputs is inconsistent so something is not right there.

Agreed, but where?

The diodes say BAT85 on them, how should I test them in case they are counterfeits? Should I pull two of them from the board and check resistance across two of them with reversed polarity?

mackanov wrote:The diodes say BAT85 on them, how should I test them in case they are counterfeits?e

Is there a reason to believe they are counterfeits?

Should I pull two of them from the board and check resistance across two of them with reversed polarity?

I don't think it is possible to test that the diodes are in fact BAT85 although it is possible to presume they are schottky if the forward voltage drop is ~0.3V rather than ~0.7V for regular diodes. But again, non-silicon diodes may have lower voltage drop.