I recently attempted a build of the Notes and Volts NaV-1 synth (Board layout and schematic available at http://www.notesandvolts.com/2013/03/nav-1-arduino-synth-build-it.html) and I *think* the ATMEGA328PU that drives it is not wanting to start up.

Upon completion of the board, I triple-checked that all trace cuts were done in the appropriate locations and that my multimeter did not show continuity across them. I verified that all the ground points had continuity, then powered up the board without any chips in the sockets, and verified that all the places that should be getting 5v were getting 5v. I plugged all the chips in, powered the board back up, and found that I was unable to upload the firmware to the ATMEGA using the ISP headers, so I removed the chip and stuck it in the chassis of an UNO to upload the firmware, then popped it out and put it back in the synth mainboard. After that, I powered up the synth, and the LCD turns on, with two rows of boxes, but no splash screen, and there's a repetitive noise coming out of the audio out connection. So, I hooked up the oscilloscope, and probed the pins of both microcontrollers where signals should be coming out, but only got ~2mV noise.

Here's a video of it failing to start:

This is BY FAR the most complicated thing I've ever built, so I'm really beyond my skill level here... I don't know what to check next. I have triple checked that all the components, jumper wires, and trace cuts are correctly located, that the upload was successful to the ATMEGA chip... there is a pin that's slightly bent on the optocoupler, but as far as I can tell it's still working, and even if it wasn't it shouldn't affect the rest of the board.

How should I proceed in diagnosing this so I can get it fixed?

Thanks.

 

Ha ha! One step closer to victory... I found the tiniest bridge ever between two of the rows right by the ISP headers, scratched it out and now the board comes up, shows the Notes & Volts splash screen, formats the eeprom, and all the buttons work as expected.

But no sound comes out. That may be because of my keyboard, which I don't think I can put into omni mode. Either way, now my oscilloscope will be a much more useful tool in figuring out where the sound is getting stopped... I'm hoping I didn't fry the soundgin chip, because there's not too many of 'em left anywhere.

 

Ok, so if I alter the firmware to output some test notes over and over it will do so dutifully, so the soundgin chip is still good. I also probed the sound output pin and got the correct carrier square wave, so we're in business there.



Then, I probed the output of the optocoupler, and pressed some keys on the MIDI keyboard, and sure enough, the MIDI commands show up too.



So, I altered the firmware to listen on channel 1 instead of OMNI and fired it back up... made sure my keyboard was on channel 1... and nothing. So I'm thinking that means the problem has to be in the firmware, and it for some reason isn't actually registering the note commands I'm sending it and forwarding them to the soundgin.

 

If that image is supposed to be MIDI signals, shouldn't they all be logic levels? That could be the area where the problem lies.

 

You should take a hobby knife and run it between the traces when using stripboard to check for any kind of bridges solves a lot of problems.

 

@Alec_t : I probed pins 4 and 5 of my keyboard directly, and got that same rounded off shape instead of nice clean pulses, that might mean that my keyboard's MIDI out is faulty. Ugh... if that's the case then I'm SOL until I get some other device with MIDI out to do some testing with. Is MIDI supposed to be a high signal that's brought low to send a signal, btw? I thought it should be low and go high to send a signal.

@ISB123 : Thanks, will do.

 

Is MIDI supposed to be a high signal that's brought low to send a signal, btw? I thought it should be low and go high to send a signal.

This article describes the MIDI format. Note there is a high/low inversion because of the opto-isolator action.

 

• Logic 1 → High → no current flow → Opto-isolator LED off → MIDI receiver sees High, logic '1' (data bits, stop bit or idle)

• Logic 0 → Low → current loop flow → Opto-isolator LED on → MIDI receiver sees Low, logic '0' (data bits, start bit)
  
  That bit seems to suggest that this is wired backwards, and I am indeed looking for low that goes high whenever I press a key, no?
  
  The layout calls for the pins to be wired as they are, but reversing them would invert the logic.

 

Oh, wait, nm it idles on 1. Ok, so my contoller is b0rken...

 

Ok, so I used the test midi sketch from here:

https://www.arduino.cc/en/Tutorial/Midi

And it also makes a similar waveshape on the scope as the keyboard does. So maybe my keyboard isn't broken after all. Inputting that wave shape into the synth still does nothing, though.

 

Can you probe the pin where the data goes into the atmega?

 

Just did... looks like it goes in on pin 2 of the atmega, which is the RX pin, so that checks out... and it looks identical to probing it at the resistor next to the optoisolator. Same waveshape and amplitude.

 

One bit should be 32us long, so you might try making the timebase faster and see if there is anyhting that looks more like a proper signal. Also, is that resistor really 270 ohms?

 

My multimeter reads 269.6 ohms, and increased time base to 50us looks similar.

 

Any ideas why my scope would show the MIDI signals like that? Is my probe just calibrated wrong? I went through the scope's self-calibration, and the internal signal generator's square wave already looks square down to about 50ns, after which the wave ramps up to the right, and no amount of turning the adjuster in the probe will affect it one way or the other.

 

Can you check what is coming right from the MIDI cable?

 

What opto-isolator are you using? It needs to be a fairly high-speed type. The waveform shows signs of a slow one.
Can you post a schematic of the MIDI input section of your synth?

 

I did... coming in from the MIDI cable looks identical, and if I program my UNO to think it's a MIDI device it looks the same on the scope. I'd say my probe wasn't calibrated right, but it reads the test squarewave on the scope just fine.

 

It's a 6N138, and here's the full schematic:

 

How long is the midi cable? Can you check at your keyboard what is going out?