If you pulled up the board layout, that airwire is actually the PE connections for the DIN jacks. They are just the stress joints and I didn't see a need to tie them to ground, so I didn't.

I did go back and forth on using a ground plane versus traces on the analog section though. Since it was basically point to point in each side of the tranny, I didn't see much value in a plane there. Doing it that way also helped keep the analog wires further away from the digital section.

I don't know if you ever figured this out, but Eagle allows you to explicitly define different grounds. Use add, then look under supply1 or supply2, and select for example GNDA for everything on the analog side. What this does is just to allow you to complete the board without having an airwire hanging, but it's the right thing to do.

I think I rambled about it but didn't mention the current solution. I basically kept them separate. The digital side has both ground and power planes and the analog section has separate ground traces. After I thought about it, there was really no other way to do it because the transformer is doing both noise rejection and isolation and you can't do that with sharing a plane on both sides of the tranny. Given that, it also made no sense to connect digital and analog grounds together. Just in case, I used plastic connectors for the analog side, should the case ever be grounded to the digital count (e.g. if wanting to attach the voltage regulator to the side of the case to help with heat).

So, in short, you can think of the digital and analog section as entirely separate - they just happen to be in the same box.

Did I miss it, or where did you write what you did about the separate ground planes?

Wow!

I finally finished the thing! Works quite well, though I'm still working on adding a few features. The DIP switches aren't even being used right now, and the LED doesn't blink how I'd like. I also need to get the side panels cut still. However, it currently functions in the way I needed it and does a damn good job of it.

The rest of the photos are on our Facebook page (https://www.facebook.com/VictimCache) and I also have a GitHub repository for this here:

https://github.com/m00dawg/stageboy

The project isn't "complete" but it is functional at this stage so be mindful if you want to trudge on ahead and build it.

EDCOR transformers finally came in WHEE! I've also been thinking about the LED and button situation. I'm not pondering using DIP switches to set the mode and doing away with the button altogether. Then I'd only have 1, maybe 2 LEDs (BPM and, optionally power). The DIP switches would not require reading from EEPROM and there won't be any external switch to switch the mode (say if you accidentally kicked the button during a show or something).

I've been using the above pictured solution for a while now and I have to say, the extra iron is well worth it in terms of noise rejection performance. It's very good.

dope! super sick man!

Still waiting on those EDCORs but I have been playing with the on I do have and I've grown more impressed at the noise rejection abilities of the thing. It doesn't look like much, but here is basically an ugly prototype before the prototype:

The board at the top is my ShieldBoy and the yellow protoboard is the noise rejection circuit. That's basically all there would be to the StageBox. The aluminum enclosure came in along with some samples of the audio jacks I will likely use. Waiting to have the board fabbed until I can measure the EDCORs just to make sure. Still am pondering the LED situation as well.

I think a different type/rating on the resistors may help? The mixer I setup while testing the EDCOR had the same problem, though I couldn't really hear a noticeable problem when using my GB test song.

The EDCOR I'm looking at is perhaps as big as the whole PISSBox I plan on using the 1455K1201 from HammondMFG (http://www.hammondmfg.com/1455.htm), which should fit both the PC and PCW line of transformers from EDCOR. Specifically this one is the one I'll likely go with, though I may try a 2:1 or 4:1 so see if that fixes the speaker funkiness on the GB. Since I'm having to send both channels to the tranny, I needed something a bigger to keep that bass response up. For the price, the EDCOR I tested (the 600:600:600) did an amazing job so I think it should be good to go.

I should probably setup some sort of test to verify the balanced circuit works well for longer cable runs. Don't have great means to do that, but I guess I can connect all the XLR cabling I have and make the longest cable possible and measure that high frequency rolloff with an unbalanced and balanced output and compare.

I may also change the design a bit for how the LEDs work on the ArduinoBoy side. I'm trying to avoid using extra boards which means all my I/O has to be on the bottom board in the enclosure and there isn't enough room for 6 LEDs. I'm thinking 2 tri-colors or perhaps 3 standard colors and doing sort of a binary count kinda thing to determine the modes. Smaller detail though (I plan on having headers on the board itself for folks wanting to do something else).

Nicely done!  The PISSbox "mixer" does indeed have some odd behavior, will try to address this in future revisions.  Just how big is this transformer you're using?

As another update, I got the EDCOR tranny. It didn't work quite as I thought it would, but in the process, I think I found a better way - hook the L and R signals as the outer taps of the input side and use the center-tap as ground. This creates a difference signal. As with the PISSBox, play all the content on one channel and, BAM, noise reduction. It has the added benefit of not needing a passive mixer which seemed to be causing some harmonics from what I could tell. The EDCOR transformers are much larger than the one used on the PISS but, as a result, they sound pretty good - only a very marginal loss in bass from what I could tell. The output side is also balanced so two birds, meet one stone.

The caveat is that the speaker volume on the GB is noticeably attenuated. I rarely use it when hooked up to a mixer or my sound-card so no big, but I think that indicates an impedance mismatch in the GB and input side of the transformer. I was using a 600:600 Ohm transformer. But since the output seems unaffected, so long as it's not a risk to the GB itself, it's a non-issue. I'm just not sure if it is indeed a risk or not.

Any thoughts there? Apart from that, the box is coming together nicely and, if this works, it's basically ready to be prototyped.

EDCOR came back and suggested this as a means to accomplish the both goals using a single transformer. In practice, I'm not sure how well it will work, so I am going to buy one and give it a go before changing the above design.

I am cautiously optimistic, however! The only downside is that using a single transformer like this means that it cannot be bypassed like the current design (so as not to color the sound more than the noise rejection). So it'll have to sound good, but the spec sheet indicates 20Hz-20kHz and I've read good things about EDCOR so I think it's worth a shot.

As an update, EDCOR makes a combination summing and balancing transformer so I'm checking with them to see if inverting one of the inputs is possible. If so, it means those two tranny's turn into one. EDCOR also has much better specs and for way cheaper. The downside is availability is sometimes an issue, but is probably worth it.

Well here is what I came up with. For now, I am calling it the GigBoy:

Note that air wire is intentional as there are two ground planes. I did it this way because both sides are grounded to the GameBoy and I haven't figured out how best to do that to avoid a ground loop.

Still a work in progress as things need to be tested, measured, etc. I need to make a simple control surface board for the LEDs and button for instance. I am fairly happy with it so far. I do wish I could have figured out a way to house a 9V battery in the thing, but a bigger enclosure would be needed and the cost increase didn't seem worth it. Similarly, an active circuit for the balanced out would have likely provided better sound but the power solution just didn't seem worth it.

On that note, though, I suspect the transformer will still attenuate some bass (still need to test that), so for cases where that cannot be compensated for and balanced output isn't needed, I added jumpers to switch it to an unbalanced out and avoid the output transformer entirely.

Feedback certainly welcome! I haven't decided on which open hardware license to use, but I plan on using one with it and, like my other designs, will make the board layouts and schematics available.