e-licktronic Forum

I figure I'd better start a new topic..

So I've spent the last few days trying to troubleshoot my Nava build. The build took me around 3-4 days and was straight forward. Unfortunately on power up I'm getting odd behaviour and the unit isn't functional.

The main issue is with the display - I get top row all blocks and the LEDs across the board exhibit random behaviour upon each power up, inconsistent. I first replaced IC115 (EEPROM) but still had the same behaviour. I then traced top and bottom layouts of just the sequencer section and removed all other ICs in all instruments and only left ICs in the sequencer section. Still the same behaviour. I then tried to remove everything else and leave just the Atmega1284p and used my Arduino Uno to try and see if the chip was OK.

At this point I could not read or get any useful info from the chip or even reburn the boot loader. So I decided to try and factor out the PCB and I built a breadboard to see if I could read the chip from that instead. Success, I could read and refresh the Atmega chip! I popped it back into the Nava but the same symptoms were shown.

I've checked continuity between the main pins on the Atmega and the ICSP header and all are there. Remember, all other IC's are not populated so I'm effectively using the Nava as a breadboard, but I cannot read the Atmega from the Nava board.

I'm at a total loss here, the circuit I built on the breadboard is simple - pin 6, 7, 8, 9 for MOSI (also with 10k to +ve rail), MISO, SCK and RST, pin 10 and 11 +5V and GND. Pin 12 and 13 with the 16Mhz crystal and 22pf caps to ground off each. On the analog side, 0.1uF cap across the +ve and -ve rails, that's it.

Can anyone suggest what I can look at next? All voltages across the board are normal at all the relevant power pins on all ICs. When I had all components populated, I couldn't get into TM2 calibration, nor get any voltage from TP1. I've also tried initialising EEPROM (blindly without display) but this also didn't seem to work.

The issue with my unit seems to be that the Atmega is not functioning properly while plugged into the board. Any help much appreciated! I'd still like to try troubleshoot more with some guidance as to where to look but I may resort to paying someone for repair. Any takers?

Hi druu,

besides checking traces between the display and Atmega, did you also check for continuity between each of the solder points? Could it be that you have a short somewhere?

Different LEDs lit each time when powering up the Nava is nothing to worry about. I experienced this on a lot of units I've build.

Cheers, good luck and despite all your efforts: don't forget to enjoy christmas!

KG

KlangGenerator wrote:Hi druu,

besides checking traces between the display and Atmega, did you also check for continuity between each of the solder points? Could it be that you have a short somewhere?

Different LEDs lit each time when powering up the Nava is nothing to worry about. I experienced this on a lot of units I've build.

Cheers, good luck and despite all your efforts: don't forget to enjoy christmas!

KG

Thanks for the response Klang. Yes, I've tested continuity across the whole sequencer/digital section and all are OK. It really seems there must be a short somewhere at this point. I'm going to merge the top and bottom layer PDF's into the BRD file to make tracing easier with the eye tool. I think I've pinned it down to being a short on the Atmega pins.

Thanks for the kind words, enjoy your festivities as well.

So I found the issue finally, two actually. First was the crystal - I bought this from Tayda and thought they would be generally quite robust. I removed the one installed on the board and put it onto my breadboard and it was consistently failing to read the chip. I put the other crystal back in and swapped them a few times to see how consistent it would be - it wouldn't read the chip at all with the Tayda crystal.

Secondly, the socket for the Atmega wasn't getting reliable continuity from the Atmega to the crystal! I was testing consistently continuity from the socket pins to the crystal but never thought to check from the pins of the chip itself while installed. I jiggled it a few times and found an 'angle' for the chip to sit at where I was then able to read the chip and get a display!

Funny, this also resolved the issue I was having with trying to get 5V across TP1. So anyone out there still having issues - check your Atmega socket and crystal!

Congrats! Now you can finally enjoy christmas!