Digital board woes

[LogiK]

Hello all,

I recently completed the construction of a Atmega32, 22-input module. It's been a really good educational experience for me, but it has a few bugs to work out. Now, I'm no electronics guru (just a quick learner; I make a living in software) and I fully expected a puff of smoke when I applied power for the first time. Fortunately, that didn't happen. But, aside from the relatively dim back-light, there is still no apparent activity on the LCD after entering "manual configuration" mode by bridging pins for the keyboard to ground (I haven't installed the pushbuttons yet.) Since this is a pretty big circuit and it uses a microcontroller I'm not familiar with, I'm not sure how to proceed.
Some things to note:
- Currently I've detached the analog board and USB board from the digital board. Looking at the schematics, I didn't see any problem doing this, and I hope that's an accurate observation. In fact, with the USB board attached, my USB port would not provide the 5v required (it was around 0.75v). This is another problem entirely, which I'm sure you'll see me post about in the near future.
- According to PonyProg, the Atmega32 programming was successful. I've tried doing this twice now, and both times seemed to work.
- I am getting 5v between Gnd and Vcc. The "rails" I've set up around the circuit for those seem to have continuity all the way throughout.
- I'm using a Hantronix HDM16216L-5-L30S LCD (assuming Mouser is done upgrading their site: http://www.mouser.com/search/ProductDet ... 16L-5-L30S)

So, being relatively green when it comes to electronics (and I'm sure the veterans out there will think this a very nebulous question to ask), is there a general process one can go through to narrow down where the problem is on a circuit this size, or at least a good resource someone can recommend where I can read up on this sort of subject? I know there is no substitute for experience, but I've never seen as good of an excuse to really get into electronics as a meshing of music and (DIY) technology.

Thanks for any help!

edit:
I'm suspecting something is fishy with the LCD module. Here are some voltage readings against ground (pin 1):
pin 1: 0v
pin 2: 5v
pin 3: 4.41v (adjusted the trimpot's resistance up a bit to see if it would do anything. Sadly, it did not.)
pin 4: 4.5v - 5v
pin 5-10: 0v (Notable was the fact that I was unable to get anything on pin 6. Should be giving voltage readings at all?)
pin 11: 0v - 5v (all of these 0v-5v readings jump around a lot. I assume it's signaling high/low for binary character data and my meter can't keep up)
pin 12: 0v - 5v
pin 13: 0v - 5v
pin 14: 0v - 5v
pin 15: 4.29v
pin 16: 0.73v
Are these readings normal?

[Synthex]

Have you solved your problem ?

You could plug the LED (Midi activity) to see if the μP works well.

[LogiK]

Thank you for the suggestion, Synthex. I'm still trying to work the bugs out, and I will definitely give the LED a try. A question about that, though: I looked at your recent PCB layout and it appears you're connecting the LED to pin 27 of the Atmega. I wasn't aware that that pin was even being used (along with pins 22-26). Is this some sort of "undocumented" feature that isn't apparent in the schematics, but already implemented in the firmware/microcontroller? Or am I just misunderstanding something? In any case, I was thinking about how I could add such an LED to the final product, so this is a good feature to have.

[Synthex]

See in topic : http://www.megadrum.info/forums/viewtopic.php?f=3&t=58
Changes in 20071124 version

[dmitri]

I'm suspecting something is fishy with the LCD module. Here are some voltage readings against ground (pin 1):
pin 1: 0v
pin 2: 5v
pin 3: 4.41v (adjusted the trimpot's resistance up a bit to see if it would do anything. Sadly, it did not.)

You should definitely be able to get between 0 and +5 on pin 3 depending on the trimpot position.

pin 4: 4.5v - 5v
pin 5-10: 0v (Notable was the fact that I was unable to get anything on pin 6. Should be giving voltage readings at all?)

pins 5,7-10 must read zero. Pin 6 should be closer to +5 since it's only pulled to ground when something is written to the LCD

pin 11: 0v - 5v (all of these 0v-5v readings jump around a lot. I assume it's signaling high/low for binary character data and my meter can't keep up)
pin 12: 0v - 5v
pin 13: 0v - 5v
pin 14: 0v - 5v
pin 15: 4.29v
pin 16: 0.73v
Are these readings normal?

Check pin 6 for a shortcut to ground. Check you trimpot. As Synthex already suggested, try connecting a MIDI LED to check for MIDI activity.

[LogiK]

See in topic : http://www.megadrum.info/forums/viewtopic.php?f=3&t=58
Changes in 20071124 version

Ah... Not quite "undocumented". Thanks!

You should definitely be able to get between 0 and +5 on pin 3 depending on the trimpot position.

Yep, I do. Sorry about the poor wording... I was trying say that regardless of the trimpot setting, I get nothing on the LCD (in case it might have been a contrast issue).

pins 5,7-10 must read zero. Pin 6 should be closer to +5 since it's only pulled to ground when something is written to the LCD

Hmm... Pin 6 might be the problem then.

Check pin 6 for a shortcut to ground. Check you trimpot. As Synthex already suggested, try connecting a MIDI LED to check for MIDI activity.

I went ahead and added the LED. The good news is that it does light up, though it never turns off, either. Then again, I expect no MIDI activity going out (the analog board is disconnected), so this doesn't worry me too much. Is the intent for the LED to be always on until activity, or always off until activity?

I took some resistance readings between LCD pin 6 and the ground pin on the USB connector. With the Atmega in the socket and using the 20k ohms range on my meter, I read around 5.47k ohms. Without it in the socket, I get around 16.2k ohms. Depending on what the "correct" circuit's readings are, this only says to me that either there is indeed an improper connection, or there is simply some way to get from pin 6 to ground through the LCD/Atmega.

Thanks again, Synthex and Dmitri for your suggestions. Perhaps we've narrowed it down to pin 6 misbehaving. This at least makes solving the problem appear less daunting to me.

[LogiK]

I finally managed to get the digital board to display the VU screen. I left space to quickly connect to a 20MHz crystal in case I wanted to "overclock", as it were. Looking over the circuit with fresh eyes, I noticed I screwed up my soldering and connected the 16MHz and 22MHz crystals in parallel, which probably didn't make the Atmega very happy.
I also tried re-flashing using AVRDude in Linux this time. It worked well, to my surprise. It's good to know that electronics development isn't completely anchored to Windows. For someone new to the hobby, it certainly can seem this way at times.

Now to get the USB board to flash correctly!