Wave Shield Pumpkin with IR sensor audible Noise Issue

[vorp]

I just got the parts for the pumpkin wave shield kit and everything came together nicely.
Wave files play nicely.
However, my sensor for triggering the thing is a Sharp IR sensor ($12 job: GP2Y0A02YK).
The audio plays great but when I plug in the IR sensor line (I've used a detachable molex cable) I get a very noticeable, very annoying pulsing coming through the speaker (and headphone jack). It's high-pitched, pulsing at about 30hz.
Resets don't pause the pulsing so it must be circuit-based (as opposed to software).
I've double-checked all my connections (I've wired the IR identically to the sonic Ms. Ada uses in the project tutorial) and all I can think is there's something funky with my IR sensor (which _does_ successfully measure distance - it just adds the audible noise too).
I'm very new to electronics so any help or guidance is appreciated.
Thanks!

[adafruit]

check which ground you used
one side of the arduino the one towards the headphone jack is audio ground and the one towards the volume is digital
try one or the other

[vorp]

If you mean the ground on the shield just next to the other ground...
That is, just slightly closer to the headphone jack by one pin, then that doesn't help.
I get the same pulsie whine.
If there's another ground connection for my sensor you're talking about, please clarify.
Thanks for the idea!

[adafruit]

no the two .1" away are the same pin

look again:
one side of the arduino the one towards the headphone jack is audio ground and the one towards the volume is digital ground

[vorp]

Okay. I think I found the ground you're talking about.
Just a couple pins behind the volume wheel, next to digital I/O #13.
Sadly, same problem when I move the IR sensor's ground to there: same dang whine.
I don't want to connect the analog and digital grounds, right?

I guess there's a chance my Sonic will arrive in time (talking art door for a gallery show in NYC this Saturday) and, hopefully, that will be whine-free.

[vorp]

Incidentally, we noticed with a voltmeter that both leads to the speaker connect to ground (analog).
Is this correct?

[karlgg]

Electrically, all the grounds will be 'ground', so that's normal. It's just a matter of keeping the noisy bits far away from the noise-sensitive bits. If that IR sensor is creating a lot of noise, there may not be enough distance anywhere on the board to get away from it.

Maybe you need a filter cap on power to the sensor? Does the spec sheet show anything like that in a sample circuit?

[adafruit]

yes most IR sensors need 10uF across power and ground AT THE SENSOR

[vorp]

Does it matter if it's an electrolytic cap?
I held a 10uF electrolytic cap across the sensor's power and ground (I tried both directions) and it didn't make a spot of difference to the audible pulsing noise.

[adafruit]

tantalum is better. try larger caps too
and dont put it on backwards

[vorp]

Thanks for the suggestion.
On the final day the sonic arrived and it popped right in with no noise.
So I skirted the noise issue with the IR.

Sadly, as an epilogue to the project...
The "art door" project was a success: It played different bizarre songs or sounds depending on how far people were from it.
The artists involved painted an inverted crucifix in custom leopard print on the front with the speaker and sonic hidden thereupon.
And then, while moving it to the Brooklyn gallery space from Boston, strapped to the back of a car, it came loose, flew off on the Mass Pike and was never seen again. It was absorbed back into the matrix apparently.
Lower the Italian flag for a lost Arduino...