First Build - Display Problems

[Joe Gennari]

So this is my first HM build and my first complete through-hole soldering project. I'll walk through what happened and what problems I had during the process, and I have now. As you'll soon see this is a 4.3 build. I had relatively few problems assembling the components. The first snag I hit was soldering the remaining BS170 in place of the MCP1700-33 on the base board. I re-ordered a few BS170 and a solder pump, but not after trying to remove the component using only heat. The rest of the base board was uneventful, but I had a few problems on the LCD board. I soldered the LCD pins with the longer side on the wrong side of the board. I removed the old pins and corrected the problem. But then I soldered the LCD to the wrong side of the breakout board.

I connected a RPi Zero and only USB 5v to the Pi and the board booted first time with no problems. The firmware was flashed and the board showed the pit temperature. I didn't have a wifi adapter so I couldn't change the pit probe setup or check any other settings. Other than the pit probe the HM was operating as expected. I left the HM on the bench overnight with 12v/1a and when I came back in the morning the screen was blank. The RPi was dead (no status lights). I pulled the RPi and the HM rebooted and seemed to be working with the following exception: half the screen appears dim when moving through the menu. Occasionally weird characters appear on the display.

I've tried reflowing most (all?) the joints I could. I've tested all the voltage that were called out or make sense and can't find any discrepancies. I've included hi-res pictures of the board and a video of the bootup sequence and problems.

So at this point my questions are:

1. Why did the RPi die?
2. Why is the screen acting the way it does?
3. Is there a voltage diagram similar to the ones for the 4.2?

Video:

Images:
http://imgur.com/a/8QGGL

 

[John Bostwick]

You may need to redo the atmega programming.

 

[Joe Gennari]

You may need to redo the atmega programming.

John, thanks for the tip. Due to the release of the Pi 3 image I was able to test again today. Unfortunately reprogramming the atmega didn't do the trick. I also swapped the atmega with a known good one and still no dice. Half the screen appears dim and LinkMeter claims it can't communicate via serial.

It does, however, flash the .hex file with no problem.

One problem I do see is that step 7 in the new 4.3 build guides of the base boards include 4 100k resistors. The picture doesn't show installing them however, so I must have missed it. I'll need to order some extras and try again.

Any idea what type of problem leaving that 100k resistor out would cause? It's the one nearest the thermocouple in the picture below. Any other ideas on what could be wrong?

 

[Bryan Mayland]

One problem I do see is that step 7 in the new 4.3 build guides of the base boards include 4 100k resistors. The picture doesn't show installing them however, so I must have missed it. I'll need to order some extras and try again.

Any idea what type of problem leaving that 100k resistor out would cause? It's the one nearest the thermocouple in the picture below. Any other ideas on what could be wrong?

Oh you've caught my terrible secret. I went through the whole process building and photographing the boards at every step and left out that 100k resistor in step 7 so it is missing in 7 and 8 but visible in step 9. I have it on my todo list to do another board start to finish and fix the two photos. The 100k resistor should have been in the kit through, if not (or you threw it away because it was extra) let me know and I'll mail you another one. Without it, the pit probe won't work because it is disconnected from the microcontroller.

As far as the LCD, that's something I've never seen before. We don't have access to sub-character resolution so there's definitely something going wrong in hardware there (no bad data lines or anything). Maybe there's some sort of short somewhere that's browning it out? Does any part of the HeaterMeter feel warm in any way? Without a blower running and a Pi attached, everything should be cool as a cucumber. Turn the LCD brightness to 0% from the menu and see if that makes a difference. After that, and this may sound strange but if you don't feel any part getting warm, I sometimes probe about with my face/lips for overheated parts which are more sensitive than my fingers.

I can use my morning tomorrow to do a 4.3 voltage diagram like the 4.2 version and a grab a couple FLIR images to see if there's anything that gets above ambient temps (apart from the LCD from the backlight).

EDIT: I looked at the images you posted but I don't see anything that stand out immediately.

 

[Joe Gennari]

or you threw it away because it was extra

This is most likely the case. If it's easy enough for you to send one that would be great. Since you're just across the state I'll probably get it sooner than Mouser.

After that, and this may sound strange but if you don't feel any part getting warm, I sometimes probe about with my face/lips for overheated parts which are more sensitive than my fingers.

Not strange at all, I'll try that when I get home tonight.

I can use my morning tomorrow to do a 4.3 voltage diagram like the 4.2 version

Perfect, that will be immensely helpful.

 

[Bryan Mayland]

100k resistor is in today's mail!

Here's the base board. This is a thermocouple PCB with a default configuration and with no probes plugged in. All the blower voltages will probably be 0V for you because the thermocouple is disconnected from the circuit, so what I would do is fold a piece of wire or a paperclip into the 100k ohm resistor holes for testing (not the 10k, the 100k!) so it connects it to the microcontroller. Then everything should fire up and let you probe around (the default setpoint is 225F and the thermocouple will read ~21F so it will fully turn the fan circuit on).


Will do the LCD next. EDIT: Done! The "1.4*" is the contrast pot adjustment, so that should be in the 1.4V ballpark but the same voltage at the two points.