Feedback on Arduino/wave shield mashup?

[ericzundel]

I was wondering if there was any kind and generous soul out there that could give me some design review/feedback/criticism on the schematic for my new project. I am trying to stuff a atmega328 (programmed with the Arduino environment), the Adafruit Wave shield circuit and a 10x6 matrix LED driver into a single PCB. I'm looking for bugs, readability, Eagle tips, what have you. My goals are:

- To make it fit into a confied shape (I have 4" diameter circle to work in, so I think I'm in good shape)
- Send it out to a pcb prototyping house (never done this before)
- Use all SMT parts (I've not done much of this before, other than soldering parts on proto boards, but I have a hot air soldering iron on order)
- Get it all done by the end of August.

Anyway, I've ordered the parts and finished the schematic and am playing with laying out the board. I toyed with the idea of eliminating the 75AHC125 buffer and going to 3.3V on the MPU and everywhere else I could, but the deadline is approaching - I was hoping to use the already debugged circuits from the wave shield and the led matrix I built on my bread board.

* The Eagle design check flags the op-amp circuit as having 2 outputs on the same net. The - outputs of the both op amps feeds into the - input without any kind of component in between. I am not up on op-amps - is this safe to ignore?

Stuff I changed (besides swapping parts for their SMT counterparts)
- used microsd card holder instead of full sized SD/MMC
- added a cap to debounce the reset switch
- added the 74ahc125d chip to the waveshield schematic
- using the USB-TTL-3.3V cable to upload instead of an onboard usb to serial converter.

[adafruit]

sounds great! maybe put on a better amplifier? LM386 are cheap and theres lots of example schematics

[ericzundel]

I have an old lm386 circiut wired up on a perf board. I tapped into the output of the dac on the waveshield and it sounds like total banned. I assume that' because I'm not protecting the amplifier circuit from noise very well. Hmm, maybe I'll take more risk it in my next project!

Speaking of noise, I'm planning to make 2 big ground planes on the bottom of the board, like Lady Ada did. I assume that's one for the audio circuit and one for the digital circuit and only trace going between them. I looked at the board file in Eagle and it seemed like the spare top layer of copper was not connected to ground... (why?)

[adafruit]

there are two grounds its a little strange but i split them from the ground pins on either side of an arduino. there was noise otherwise. you'll need to buffer/filter -before- an amp. it took many board revisions and even then i think it was a little bit of luck

[SiliconFarmer]

You dropped the R7 1.5K resistor on the output of the DAC. That's a key element of the 11KHz low pass filter. You won't like the results without that resistor. Maybe that was the issue with your LM386 circuit.

Splitting ground planes is... non-trivial. Try to follow the analog layout as closely as possible to what Ladyada got to work well. If you don't know precisely where the two ground planes are supposed to be connected, don't guess, keep looking.

Good luck with your first SMT project. Sounds cool!

[ericzundel]

Thanks for the feedback. I found another minor problem today when working on the board. I used the wrong MPU! In case this board doesn't work out, I am formulating a backup plan that includes a perf board and a larger plunger.

[ericzundel]

I Googled and found this article on the analog/digital ground plane subject http://www.hottconsultants.com/techtips/split-gnd-plane.html which looks like an approach worth trying. If it doesn't pan out, I'm going to align headers so I can strap a waveshield kit on top.

[adafruit]

its a Black Art...takes a few revisions but a good skill to learn!

[SiliconFarmer]

With a deadline at the end of the month, setting up headers for a Waveshield is a very wise move

[ericzundel]

I just thought I'd follow up. I followed the instructions to use a single ground plane but put a break between sections that used different types of signaling. I isolated USB signals and the audio signals from the rest. For some reason, the USB circuit didn't seem to work at all, but I'm not sure the ground plane was the problem. I wired just about every connection incorrectly and had to cut traces and solder jumpers onto the chip... In the mean time, the FTDI TTL cable worked out as a backup.

The sound part of the circuit turned out well - when I play music the sound is on par with the Waveshield 1.1 that I have. Here's a picture of the board front and back that shows the ground plane. I'll publish the rest of the project after its official unveiling.

[adafruit]

rad! you're almost done

[ericzundel]

The electric plunger is done! Here's a video of the completed project. The audio sounds good, but not when recorded on my camera with the A/C running full blast in my friend's condo, so I dubbed in the actual wav file on the audio track:

http://www.youtube.com/watch?v=53q1tcMzT84

If someone is really interested in the final schematics, source code and stuff, I uploaded everything to github:

http://github.com/ericzundel/electric-p ... ree/master

[adafruit]

wow congrats. if only i understood -why-

[ericzundel]

What?!?! C'mon - who doesn't need an Electric Plunger?

It was a gag for a co-worker who was moving on to a new project. Our unofficial project mascot was a plunger.