My design: Digitally Configurable Audio Routing Patchbay

[aarontcramer]

Hello All,

I am new to the forum, and would like to share my latest design: The Digitally Configurable Audio Routing Patchbay (DCARP).

This device is a 32-input, 32-output audio routing tool designed for professional and home recording studio applications. It allows users to uniquely route 32 balanced inputs to any of 32 balanced outputs, assignable through a convenient 13 sensor capacitive touch-sensing keypad (no patch cables required). Additionally, users can save and recall custom patch configurations, even after power has cycled.

I have been using this device with my modest 16 track M-Audio/M-Powered ProTools setup as a replacement for my 48 point HOSA patchbay. So far I'm loving it! Interconnecting my 16 channel board with outboard gear is quick and easy, and there is no audible effect on the sound.

I have more information on this design at http://www.aaroncramerengineering.com/S ... oject.html

Also, there is a simple demonstration video on YouTube here: http://www.youtube.com/watch?v=Zxm61rff9MY

I am interested in hearing feedback about this design. Is this something you would find useful in your studio? If yes, why? If no, why? What do you like or dislike about the interface? Thanks!

And pictures...

[bantam]

you should post at www.prodigy-pro.com/diy/index.php?board=2.0 those guys would love it

[klangtone]

cool project, Aaron.

I'm wondering if the performance might be improved by getting rid of the diff line receivers and balanced line drivers, and doubling the crosspoint switches to route diff pairs throughout the unit. I'm just thinking that the less the signal has to go through, the better. And since the line recievers/drivers add up to about $144, there's actually a small cost win by doing it this way (4 crosspoints are only $121).

However, I do see your point about buffering the signals to allow a source to drive multiple outputs without loading effects. That's definitely a cool feature and something not easily done with a standard patchbay. So I suppose it's a toss-up.

Did you ever make any THD+N, frequency response, idle channel noise, etc. measurements of the box?

Roy

[aarontcramer]

cool project, Aaron.

I'm wondering if the performance might be improved by getting rid of the diff line receivers and balanced line drivers, and doubling the crosspoint switches to route diff pairs throughout the unit. I'm just thinking that the less the signal has to go through, the better. And since the line recievers/drivers add up to about $144, there's actually a small cost win by doing it this way (4 crosspoints are only $121).

However, I do see your point about buffering the signals to allow a source to drive multiple outputs without loading effects. That's definitely a cool feature and something not easily done with a standard patchbay. So I suppose it's a toss-up.

Did you ever make any THD+N, frequency response, idle channel noise, etc. measurements of the box?

Roy

Hi Roy. Thank you for the response.

When I started the design, I considered exactly what you mentioned... increasing the number of crosspoints and routing the differential signal pairs. The tricky part about that is to achieve true 32x32, non-blocking routing of single ended signals, 4 crosspoints are required. This is necessary because each crosspoint on their own is a 16x16 switch matrix. Using an array of 4 16x16 crosspoints will allow inputs 00-15 AND 16-31 to reach outputs 00-15 AND 16-31. A wired-OR configuration is used so that any input can be routed to any output.

If I were to route the differential signal pairs, I would need 64x64 routing which would increase the number of crosspoints to 16!!!! Ouch. Regardless, it would be awesome to check out the performance without the differential line receivers and balanced line drivers.

Unfortunately I ran out of time in my last quarter of school to do any THD, SNR, or frequency response measurements. Maybe when I gain access to some good test equipment, I will check those things out.

Thanks for the feedback!

[Snarl 12/8]

To test the overall fidelity, I routed a pre recorded audio track out of ProTools, through the DCARP, and back to a new track in ProTools. I then listened while switching back and forth between the original track and the routed audio track. I was unable to hear a difference between the two!

What if you recorded the same stuff into protools some through the bay and some not, (y cables) then flip phase on one set of tracks and see how much disappears? Wouldn't that be some sort of gauge of how much the bay was altering the signals?

[aarontcramer]

To test the overall fidelity, I routed a pre recorded audio track out of ProTools, through the DCARP, and back to a new track in ProTools. I then listened while switching back and forth between the original track and the routed audio track. I was unable to hear a difference between the two!

What if you recorded the same stuff into protools some through the bay and some not, (y cables) then flip phase on one set of tracks and see how much disappears? Wouldn't that be some sort of gauge of how much the bay was altering the signals?

That would give a good qualitative look at the amount of phase shift caused by the circuitry.

[TapeOpLarry]

Cool device/project. Can you build me a 1000 point bay? Just kidding. I find that the visual of a patchbay is great for helping me make decisions - I'll think compressor and look at the options and mull it over. I can see how a recallable bay would be pretty cool for some studios though. Has anyone built a commercial version?

[Andy Peters]

I agree, pretty cool little project, esp. for a senior design project.

One comment, though (I realize the limitations of a school project): to make it more manufacturable, I would have made a PCB for the I/O panel. Use vertical (not right-angle) PCB-mount connectors. Route the signals from the connectors to your ribbon cable connector. Note that in this ribbon cable, you don't need ground -- just the hot and cold from each connector. Use twisted-pair ribbon cable, too.

Basically, this makes your assembly a whole lot simpler -- it's easy to solder-up a field of connectors and run the ribbon cables than it is to use discrete wires from solder-cup 1/4" guys.

Oh, yeah, if you're springing for a PCB, you could have put the microcontroller and the power supply on that board, too, instead of trying to hack in an eval board.

-a

[aarontcramer]

I agree, pretty cool little project, esp. for a senior design project.

One comment, though (I realize the limitations of a school project): to make it more manufacturable, I would have made a PCB for the I/O panel. Use vertical (not right-angle) PCB-mount connectors. Route the signals from the connectors to your ribbon cable connector. Note that in this ribbon cable, you don't need ground -- just the hot and cold from each connector. Use twisted-pair ribbon cable, too.

Basically, this makes your assembly a whole lot simpler -- it's easy to solder-up a field of connectors and run the ribbon cables than it is to use discrete wires from solder-cup 1/4" guys.

Oh, yeah, if you're springing for a PCB, you could have put the microcontroller and the power supply on that board, too, instead of trying to hack in an eval board.

-a

I did not realize they made twisted pair ribbon cable. Where can I find that??

I was limited to a $500 budget to do this prototype, so I had to cut a few corners on the assembly methods. I definately would have done another PCB for the I/O with ribbon cable to the main PCB, as you mentioned.

Thanks!

[ashcat_lt]

That's really cool. I would definitely use it, if I could afford it.

It needs USB and a nice cross platform GUI.

[aarontcramer]

That's really cool. I would definitely use it, if I could afford it.

It needs USB and a nice cross platform GUI.

Yeah! USB and a GUI...that's next on the list.