New Heatermeter build dead


 
J

JJones

New member
Hello everyone, I've just built a 4.3.3 and I cant seem to get it working on the Pi's 5v.(I have not tried the 12v supply yet)
When I hook up the Pi3, I can get to the webpage but I see a No Communication error for the linkmeter. It fails with the following info in the system log
Thu Sep 29 05:28:02 2016 user.warn : No response from HeaterMeter, running avrupdate
Thu Sep 29 05:28:02 2016 user.err : avrupdate failed
Fri Dec 30 03:05:07 2016 daemon.err uhttpd[590]: wc: /proc/net/nf_conntrack: No such file or directory
Fri Dec 30 03:05:07 2016 daemon.err uhttpd[590]: sysctl: error: 'net.nf_conntrack_max' is an unknown key

I did some voltage readings and continuity testing and I seem to have a short between MOS pin on the pi header and Ground.
I've checked over and over again and I cant seem to find the cause of the short.
Am I misreading this/is it expected? or should I continue to search for the cause of this?
 
The "No response from HeaterMeter, running avrupdate" would indicate that it's not seeing the ATMEGA chip on the HeaterMeter board.

Post some high res pics of the front and back of the HeaterMeter board.
 
Don't mind the jumpers from the cap to the mosfet...I had some trouble with a trace coming up when removing a capacitor that I placed in the wrong location.
yNQYZgT.jpg

CnURPjF.jpg
 
Last edited:
Do you get 5v and 3.3v at the 3.3.v regulator? Also, are you sure you have the MCP1700 in there, and not a BS170?

Y48NDMc.jpg
 
I measured and confirmed that I do have 5 and 3.3 at those locations.
So far I can tell that MOS is 0v and should be 3.3v (like I mentioned earlier, there seems to be a short to ground somewhere.)
Also, AREF on the Atmega is 0v. According to the voltage diagram on the wiki, this should also read 3.3v
GdA7p8s.jpg
 
JJones said:
I did some voltage readings and continuity testing and I seem to have a short between MOS pin on the pi header and Ground.
I've checked over and over again and I cant seem to find the cause of the short.
Am I misreading this/is it expected? or should I continue to search for the cause of this?
Click to expand...
It should definitely not be expected, but I am looking really closely at your photos and I do not see any reason for there to be a short between MOSI and GND. The trace for that runs from the "MO" pin at the atmega up to the MOS pin on the Pi header, then straight over between the Pi header pins over to the "MO" pin on the LCD riser. All on the component side of the board (the bottom) and I don't see anywhere it is close enough to a GND terminal except at the Pi header.

You might want to check popping out the atmega chip and see if it still shows as a short. If it does then that means that the issue is somewhere along that trace I mentioned and I can't imagine how to diagnose that.

EDIT: Is there actually a continuity short of MOS to GND, or are you just saying it shows 0V? If it is just reading 0V then that could be because the pin is in high-Z state and just doesn't have any charge on the wire. The AVCC is ok at 0V, it stays there until the first ADC reading is taken after the chip boots for the first time.
 
Last edited:
Hi Bryan,
Yes it is a short. I also checked without the chip last night and there is still continuity to GND. I pulled the plastic housing off the LCD riser to check under that but it looks fine.
I'm wondering if its under the chip socket or inside the pi header. Either way, I'm not looking forward to de-soldering either of the components...
 
Last edited:
Found it!!
Thankfully, I didn't need to de-solder anything. just picked at it with a thumbtack until broke the connection.
FRVsr38.jpg
 
Wow that's a bit crazy. So the MO pin there was somehow connected to ground? To the ground plane the encircles the pin solder area? Don't think I've ever seen that before. So is it up and communicating now?
 
Yeah it seems like it was connected between 6:00 to 9:00 if you look at it like a clock. I dug it out until I was comfortable with it and there's no short anymore. After that it booted fine so I switched over to 12v and threw it all in the case. Seems good now but I haven't hooked up any probes yet to test out everything else.
Thanks for your help Bryan and Steve!!
 
Last edited:
Bryan, to clarify, it looks like the trace was connected to the solder pad on that pin hole from the factory.
Not a big deal to me since it was an easy enough fix that I didn't have to scrap the board and start over.
 
Last edited:
There is something else going on now. After 5-10 minutes of being powered on, the backlight starts flickering. Up until that point it works fine and continues to look as though there are no other issues according to the web interface.
Eventually the screen goes blank and then displays random characters. Any thoughts?
 
The only times I have seen blinking on a HM is when I had a weak power supply, perhaps your power supply peters out after running for a bit? I would try another PS if you have one.
 
I found that I had a slowly dropping voltage on the Mos pin on the button board. Ironically it's the same pin that gave me trouble on the base board for my previous problem. Voltage there started at 2.85 slowly dropped down to about 2.6 and that's when the LCD started to act weird. Took it all apart, cleaned all of the flux as good as I could and now it seems to work. Voltage everywhere seems stable and it's stayed running for about an hour.
 
You must log in or register to reply here.

 

Back
Top