Newb issues - AVR Firmware flash on WebUI says success, but still not working


 
C

C Gautcher

New member
Hi Everyone,

Let me first say that I've been following the project for a while, but I had yet to dive in until this long weekend. I bought a 4.1.4 kit back in August, and it sat out of sight for a little while.

I am a complete soldering newb - this is the first time. I didn't take the Physics 201 class in college that covered this circuitry stuff, so reading a lot of these posts are a bit overwhelming. I am really comfortable on the software side, but the hardware/electrical side, not so much. Following the wiki for 4.1 Assembly, I took my time, a really long time, and everything seemed to turn out ok.

The yellow LED on the HeaterMeter stays lit, and the LCD is just the black squares, and I do not see the "No pit probe" message. From what I've read, the black squares means that the AVR firmware is not flashed, is that correct? If I go to the web UI and flash the AVR firmware, it suggests that it was successful, but the symptoms remain the same.

Here is the output from the AVR firmware flash that makes me think that it was successful:

Code: 
Stopping LinkMeter OK

LinkMeter platform is BCM2708
AVR fuses ffd705 OK

                          0c30822fd7c575f7221bedeb404ea093  /lib/firmware/hm.hex
hmdude: compiled on Oct  3 2014 at 09:58:46
Using port: /dev/spidev0.0
Loading ihex file: "/lib/firmware/hm.hex" (23712 bytes)

    0% |                                                  |     0 (0.0s)
    0% |                                                  |     0 (0.0s)
    5% |##                                                |  1186 (0.1s)
   10% |#####                                             |  2372 (0.2s)
   15% |#######                                           |  3558 (0.3s)
   20% |##########                                        |  4744 (0.3s)
   25% |############                                      |  5928 (0.4s)
   30% |###############                                   |  7114 (0.5s)
   35% |#################                                 |  8300 (0.6s)
   40% |####################                              |  9486 (0.7s)
   45% |######################                            | 10672 (0.8s)
   50% |#########################                         | 11856 (0.8s)
   55% |###########################                       | 13042 (0.9s)
   60% |##############################                    | 14228 (1.0s)
   65% |################################                  | 15414 (1.1s)
   70% |###################################               | 16600 (1.2s)
   75% |#####################################             | 17784 (1.3s)
   80% |########################################          | 18970 (1.3s)
   85% |##########################################        | 20156 (1.4s)
   90% |#############################################     | 21342 (1.5s)
   95% |###############################################   | 22528 (1.6s)
  100% |##################################################| 23712 (1.7s)
Update successful
Starting LinkMeter OK


Regarding the yellow LED staying lit on the heatermeter board, from what I've read, that is probably a bad thing. Would that indicate some sort of hardware failure on the HM board?

I feel like I'm pretty close to getting this working, but I was hoping someone else had experienced a similar situation and could shed some light. Any suggestions?

Thanks
 
Does the web interface show temperatures? If it does then the AVR flash was successful and the cause of the problem is hardware related.
 
I adjusted the contrast back and forth, and I only get the black squares.

The web interface does not show any temperatures.

I'm guessing these messages on the Configuration tab of the web UI is not a good thing?

https://www.dropbox.com/s/u2ce9z2ll785ftl/hm-config-no-communication.png - note, I tried to use the IMG tags, but the pic didn't show up in the post.

Regarding the yellow LED staying lit, it seems that it does not do that any longer. Strange that it would be intermittent like that. Now it seems to either not have any LED's lit, or if it does (this thing can't make up its mind), the yellow one will stay lit for about 3-5 seconds, and then it will not come back on.
 
Sounds like you do have some issue with the HM hardware, and the AVR hasn't been flashed. As John suggested, some hi res photo's might help someone point out your problem. The one time I had this problem the socket for the ATMega wasn't soldered well, I removed the ATMega and resoldered the socket, moving the legs around a little with the soldering iron to allow the solder to flow through the holes to the traces on the other side of the board.....

It may not be completely your fault this happened, I recall a while back reading the OSH Park boards have a shelf life, so letting the board sit a long time may have made it harder to solder. The board I had a hard time with was a single board purchased from pre-printed stock, the boards I had an easier time with (soldering) were fresh prints (of three boards) that I had made to order... So maybe the older boards are a bit harder to get the solder to flow onto. With this in mind I think if you are going to solder up an older board applying some flux to the board first might be a good idea.
 
Last edited:
The ATmega is flashing properly, it reads fuses and goes through the whole update process no problem. That said, there are likely multiple solder issues, as it seems both the serial lines that go from the ATmega pin 2 and 3 aren't getting to the RX/TX pins on the RPi header (or are shorted to other lines).

The LCD and LED are both run from the shift register on the SPI bus. The SPI bus is working because the ATmega flashes, but the shift register doesn't sound like it is so start there in looking for faults.

The boards do have a shelf life but it is measured in years thanks to the ENIG (gold) coating. If the pads are dull and tarnished looking then you might have a harder time soldering it. However the OSH Park stock turns over on the order of weeks so this shouldn't be an issue. A much more common issue is not enough heat causing the solder to not flow and make a good connection, because the bonus of longevity from the ENIG comes with the downside of marginally more difficulty in soldering. Solder should at least come close to filling the pad, and should be concave in shape, not look like a ball.
 
Thank you, everyone, for your helpful insight.

Here are the front and back pics, hopefully they are clear enough to make out any glaring errors. Please go easy on me guys; I know it's not pretty. I'm hoping that the next go-round is much nicer looking.


https://www.dropbox.com/s/wwgggchsw12vk1h/hm-back-12-01.jpg?dl=0 - Back

https://www.dropbox.com/s/n4be1af0vc8e62y/hm-front-12-01.jpg?dl=0 - Front


Regarding the soldering iron, I am using a weller 40-watt iron. I used the teeny-tiny conical pointy tip, because the other chisel type tips seemed way to big for working on the HM board.

If the current board is FUBAR (in the opinion of the experts here), I don't mind starting over with a new board. I'm guessing all of the really small components could be reordered, and the bigger stuff I can try to get off the current board. I really hope it doesn't come to that though. What do you think?
 
I think it doesn't look all that bad. I do see that the capacitor nearest the ATmega isn't connected on one of the legs but that shouldn't cause this.

I'd check pins 1, 2, and 3 on the ATmega with a multimeter. Looking from the side with the Boxy Robot, that is the pin to the bottom right of of the 4.1.4 silkscreen. They should all be 3.3V. Pin 3 should bounce a little every second indicating transmit of data.
 
You must log in or register to reply here.

 

Back
Top