John Bostwick
TVWBB Wizard
Post a picture of both the front and back of the board to help with troubleshooting. Also, make sure the atmega328 chip is facing the correct way. The half-moon should be facing the edge of the board.
AVR fuses ERROR
- Thread starter Steve_M
- Start date
PatrickOttCA
New member
Here are the pics. I tried to reflow every solder point, I tried a different RasPi (ModelB), I even tried a different ATMEGA chip I had on my Arduino Uno. No love. (And yes, I need to pick up some flux cleaner!!!)
Last edited:
Bryan Mayland
TVWBB Hall of Fame
Check your board's voltages everywhere. It gets this AVR fuses ERROR if it can't communicate with the ATmega at all, even with through the programming interface. There's a diagram in the wiki
https://github.com/CapnBry/HeaterMeter/wiki/Basic-Troubleshooting
https://github.com/CapnBry/HeaterMeter/wiki/Basic-Troubleshooting
PatrickOttCA
New member
So far, the ground had issues at pins 6 and 7 of the atmega... Could not get continuity between pin 23 and the rest of the ground points. No matter what I did. I shorted pin 23 with the neighbouring capacitor, which restored ground points as they should be, but still don't have any luck. Off to the voltage carrying pins...
PatrickOttCA
New member
Any chance I could program that Atmega in my Arduino Uno to see if it helps?
Bryan Mayland
TVWBB Hall of Fame
You can't program it directly, because the ATmega doesn't have a bootloader (and that's the normal way Arduinos program). It would be complicated but you could flash it from an Uno using the Arduino-as-Programmer method.
However I have just been going over the voltage testing image and it looks like it is misleading in the area you circled. It should read 0V, but should have 1k resistance to ground. It is ok to have it shorted to ground but the button won't work if it is.
You can also verify connectivity between the Pi header and MOS (MOSI/Pin 17), MIS (MISO/Pin 18), SCK (Pin 13), and G25 (Pin 1 RESET). These are the pins used by the programming interface, apart from power/gnd.
However I have just been going over the voltage testing image and it looks like it is misleading in the area you circled. It should read 0V, but should have 1k resistance to ground. It is ok to have it shorted to ground but the button won't work if it is.
You can also verify connectivity between the Pi header and MOS (MOSI/Pin 17), MIS (MISO/Pin 18), SCK (Pin 13), and G25 (Pin 1 RESET). These are the pins used by the programming interface, apart from power/gnd.
PatrickOttCA
New member
Is SCK really pin 13 or is it 19? Assuming it's 19, the pins from the Atmega carry through the header pins on the Pi. Grrrrr!!
Last edited:
John Bostwick
TVWBB Wizard
Schematic has it as 19
PatrickOttCA
New member
Agreed! I'll try and find an Atmega chip in town and try it, but I don't have very high hopes considering the 2 chips I have here do the same.
RalphTrimble
TVWBB Diamond Member
Did you ever get around to testing the voltage around the board?
PatrickOttCA
New member
Yeah, everything checks out. I got a hold of a Raspberry Pi A+ and gave this one a shot too and still having the same problem. Sigh.
PatrickOttCA
New member
Actually scratch that. VCC and AVCC are not registering on the Atmega now...
Update: I looks like IC4 (MCP1700) is not giving out 3.3V. I see 5V between GND and input, but nothing on the output.
Last edited:
RalphTrimble
TVWBB Diamond Member
On Pin 23 of the ATMega I can't seem to see where it gets a direct gnd connection, I see 1K to gnd on that pin (though the diagram does show it as white/gnd). I think Bryan said that already? So, don't jumper that one directly to gnd, test continuity to the top leg of the 1k resistor, the bottom leg of the switch to pin 23, I dont think that's your problem (but remove the gnd jumper if you have one in place, try measure resistance to gnd, is it 1K? that would be good!)
So, on to your latest post about missing voltages... I would go down the line on your power and test each step... 12V(ish) input, 12V in the OKI 5v out, 5v in the 3.3v reg 3.3v out.... My guess is you're gonna find your 3.3v wonky, and probably have a short somewhere on the 3.3v line or possibly a wrong component somewhere or something installed backward.
So, on to your latest post about missing voltages... I would go down the line on your power and test each step... 12V(ish) input, 12V in the OKI 5v out, 5v in the 3.3v reg 3.3v out.... My guess is you're gonna find your 3.3v wonky, and probably have a short somewhere on the 3.3v line or possibly a wrong component somewhere or something installed backward.
RalphTrimble
TVWBB Diamond Member
OK, so it looks like I was right.... You could either have a dead 3.3v reg OR you could have a short to gnd on a 3.3v pin somewhere which would pull the 3.3v to gnd and kill the regulator... If the short (or wrong component) is not causing a connection directly to gnd the 3.3v might run low and maybe the regulator didnt die from it... So you need to focus on the ORANGE pins in your voltage diagram and see where things went wrong.... Prob best to do the testing with the rPi disconnected, I would prob pull the ATmega too.....
EDIT: Make sure you didn't put a MOSFET in the place were the 3.3v regulator is supposed to be, that's kinda easy to do.....
EDIT: Make sure you didn't put a MOSFET in the place were the 3.3v regulator is supposed to be, that's kinda easy to do.....
PatrickOttCA
New member
I have a short between GND and the 3.3V traces... I have removed the MCP1700 and one of the BS170 (the one directly connected to the MCP1700) and I still have continuity between 3.3 and GND. Short (no pun intended) of starting to remove headers and sockets I don't know what to do anymore.
Dave Smith
TVWBB Super Fan
Not sure but when I look at the pic, I see a solder bridge on Atmega, left side 6th pin from bottom......might just be me.
PatrickOttCA
New member
No, what you see is flux. I bought a cleaner pen and there is definitely no bridge there. I have found my problem. As far as I can tell, I have a bad board. The pad for Pin 20 of the atmega is shorted to ground. I cut all the traces around it and it is still shorted. So I lifted the atmega pin by bending it so it sticks out of the socket, and stole 3.3V from pin 7. Now everything works. A bit ghetto but heh...
Has there even been reports of bad boards?
John Bostwick
TVWBB Wizard
Not going to say its not possible, but highly unlikely to receive a bad board, I would look for stray wire clippings under components. I had a recent board that I did not realize while soldering the power connector I pushed the solder into the hole for the GND, to fast and had made a bridge to the +12 side of the power jack. Far as I knew, it looked fine, from the bottom solder joints. But, soon as I plugged in the power adapter, I had one fried 12v adapter. I did not realize though the adapter had fried so I tried another one and then I realized what had happened. and made the discovery of the solder. Solder wicked the power jack and all was good.
I was atleast able to scavenge the parts from both adapters and fix one of them.
I was atleast able to scavenge the parts from both adapters and fix one of them.
RalphTrimble
TVWBB Diamond Member
It's not impossible the board has an issue but I really haven't heard anyone reporting this kind of thing. It seems as though you've isolated the problem area, so it should be easy enough to physically look at the traces and see if there is a problem somewhere...
I will take a look at pin 20 and the traces in that area to see what we are dealing with and see if I can spot anything that might be a likely problem area.
I will take a look at pin 20 and the traces in that area to see what we are dealing with and see if I can spot anything that might be a likely problem area.
PatrickOttCA
New member
I have removed the socket for the Atmega, so I'm fairly sure there is no hidden solder bridge. I was baffled when it was still shorted even after cutting all the traces leading to it...