Finished my build, testing/troubleshooting, then...


 
J

Jason G

New member
I posted some of this in another thread, but it was a bit off topic, so I'm starting my own thread/cry for help. I started my HM v4.1 build this past week. Hardware assembly the first night. Software installation and network setup the second. All in all, it went pretty smoothly (for an admitted novice), the only issues i had were:

A) The servo in my damper/fan setup wasn't working (but the fan was)
2) My leftover ET-732 probes weren't registering
III) LED1 (red) wasn't working

I set out to address the servo issue first. Got some good advice here, and I ended up adjusting the screw on the servo a little more and re-seating the signal wire in the dampers cat5 port. Success! took some time to set my servo range values, and I was also dealing with a bit of vibration/chatter from the servo, so i kept adjusting the screw until i found the sweet spot.

Then I moved on to the probes not registering. I read here about the connector size, so i used tiny rubber bands we use for our toddler's hair as a spacer. I also read that some of the 732 probes have trouble reading at room temp, which turned out the case for mine. Since i was using a steam kettle to boil water to test this, I decided to connect everything and test basic operation. Probes were a bit touchy, but i chalked that up to the size issue, and everything else except Red LED1 was working fine. With the set temp was set to 200, and water temp at about 190 the servo was doing it's thing trying to bring the temp up to 200 without overshooting. If i jacked the set temp up to 300, the servo would open wide and the fan would come on, which was consistent with the servo/fan settings I had configured.

I was feeling pretty good about everything at this point. So i set out to address the LED. I heated the solder connection and removed the LED. Then put the HM back on power and in the configurator, set all LEDs to turn on when the fan was on. I then put the fan at 5% and green/yellow came on. I then touched the red led leads to the yellow connection, to make sure the red LED would light up. It did. So i re-soldered LED1. No luck, but this is where it took a turn for the worse. This time, when I powered the HM up, it didn't work at all. Instead of "no pit probe" at startup, it showed have 16 blocks on the top line of the display, and the red pwr light on the Pi. Nothing else. I find it hard to believe that the LED connection could destroy everything, but what do I know?

It could have been totally coincidental, or a problem somewhere else on the board that occurred handling the part while trying to deal with the LED, but I have no clue where. I rechecked my connection points for anything obvious, but I don't see anything. ARGH! Anybody have any clues as to what may have happened? I've included a few photos of my HM (minus the Pi) in the hopes that someone might spot something that my novice eyes have not. I'm already considering ordering new parts, so if you think I did a crap job soldering or i burnt something out, go ahead and say it. I can take it, but i would like any tips on what parts you think are likely good to re-use from this build in the next one.

2rqeweq.jpg

2e2nlf5.jpg

uls8g.jpg


PS - Next time you have a HM setup and working, but and LED is out. Let it be, man. Let. It. Be.
 
Steve_M said:
Maybe try a different SD card in the RasPi and see it it will re-flash itself back to life.
Click to expand...

Thanks... good idea. Questions:

Why would I need a new SD card? Couldn't I just wiper the current card, and start the re-write process over?

What about what's current uploaded to the HM? Do i have to erase whatever's on the device currently, before trying to re-flash?
 
When you say that you checked the solder joints, what do you mean? I ran into a similar problem when I made my first heatermeter. I ended up reheating each solder junction with a tiny bit more solder and that solved my problem. I just had a cold joint somewhere and that fixed it.
 
Tom Kole said:
When you say that you checked the solder joints, what do you mean? I ran into a similar problem when I made my first heatermeter. I ended up reheating each solder junction with a tiny bit more solder and that solved my problem. I just had a cold joint somewhere and that fixed it.
Click to expand...

I checked for joints that in my view "looked questionable" and I hit them with a bit more solder. It was probably only 20% of the joints, so I did not do this to all of them, but will try that. Thanks.
 
Jason G said:
I checked for joints that in my view "looked questionable" and I hit them with a bit more solder. It was probably only 20% of the joints, so I did not do this to all of them, but will try that. Thanks.
Click to expand...


Sometimes it can be helpful to take a close up digital picture and then you can zoom in on the computer to inspect the solder joints (look for bad connections and solder bridges). It's also good to inspect the TOP side of the board to see if solder made it though to solder both layers of traces. If you see all gold and no silver on the solder joints on the top side of the board re-flow fresh solder to the bottom side gently moving the component leg with the soldering iron to encourage solder to flow to the other side of the board. Pay close attention to the to IC chips (processor and shift register for the display). It's a good idea to remove the ATMega from the socket if you are going to re-solder there...
 
RalphTrimble said:
Sometimes it can be helpful to take a close up digital picture and then you can zoom in on the computer to inspect the solder joints (look for bad connections and solder bridges). It's also good to inspect the TOP side of the board to see if solder made it though to solder both layers of traces. If you see all gold and no silver on the solder joints on the top side of the board re-flow fresh solder to the bottom side gently moving the component leg with the soldering iron to encourage solder to flow to the other side of the board. Pay close attention to the to IC chips (processor and shift register for the display). It's a good idea to remove the ATMega from the socket if you are going to re-solder there...
Click to expand...

Thanks, Ralph. I don't know if you can see the 2nd pic, but I did notice during the build that there were a few connections that didn't flow through the board, but I kept on moving. When everything worked on the initial boot up, I thought we were good there, but I should go back and work on those. Thanks for the advice.
 
I think I'm going to start over. Last night while I was re-soldering, one of my capacitors fell out, and I had a hell of a time putting it back.

So I ordered parts, only re-using the LCD, atmega, and Pi. I have to re-use one of the resistors too as mouser didn't have enough to ship right away, but no biggie.

Figure I can tinker with what I've got until parts show up. Oh we'll. thanks for all your suggestions. I will be much more careful/deliberate my second time around.
 
this was my first project about 3 years ago with the Heatermeter and Linksys router. I think it took me about 4 or 6 months of redoing the board over and over, lol. Now, the heatermeter is so much easier to do, but I still make a mistake once in a while when I build one. The picture is of all the wires connecting the components on the heatermeter, no printed circuit boards, lol.
2u5t9ih.jpg
 
OK, so I'm back, but no real news. I saw the LED1 issue, and it looks like i also have one of those boards, but i don't know if that's related to my now "nothing works" issue.

so i built a new HM4.1 using a brand new board, .032 solder, and all new parts except for the rpi, atmega, sd card, and led1 + one resistor which were on backorder. re-formated the rewrote my sd card, booted up using the rpi micro usb, and i'm getting the EXACT SAME symptom, LCD comes on, but instead of "No Pit Probe" i'm getting 16 blocks of nothing...

[][][][][][][][][][][][][][][][]!!!!!

what are the odds that i screwed the exact same thing up again, or is it more likely given that i had it working, then trying to solve my original LED issue i caused the problem? i didn't replace the atmega or rpi... could i have fried either of those unknowingly?

I did try to troubleshoot that theory, and was able to connect via ethernet to the heatermeter's interface. that took some time too as the Pi's IP had changed. It wasn't the default, nor did it match the IP i had saved during the original setup. I had to discover the IP by connecting it to my router and discovering the address in my router's config tool.

I didn't test any probes, but the ambient temp thermistor was not registering. I was however able to control the backlight % on the device via the configuration tool.

I've got some new atmegas on order as that's the cheapest route right now, but I'm admittedly a moron, and could use some sound advice.
 
it does not sound like the 328 is bad, if you are able to get to the HM on the net. I would double make sure you have everything in correctly, for instance the Diodes have to be in the correct way with the line on them facing the same as the board has them. Check all the soldering to make sure you have no bridges and use a multimeter to verify continuity.
 
If you re-used the ATMEGA, try re-flashing it from the web interface. Not a huge chance of it being your issue, but it can't hurt.
 
Andy Snider said:
If you re-used the ATMEGA, try re-flashing it from the web interface. Not a huge chance of it being your issue, but it can't hurt.
Click to expand...

I did. A few times. No impact. It's so weird... I just can't shake the feeling that I did something when I was trying to fix the LED on my first board because I had everything working for hours beforehand. it's crazy that the whole thing was rebuilt with new parts (except for 5) and the symptom is 100 % exactly the same. That's why I'm trying to vet out everything on those 5 parts before doing anything else.

ALSO, I hooked up the pi to a monitor on bootup and here's what it showed. No idea what any of this means, but i thought I'd share in case somebody does:

2exni1j.jpg

28qxbo2.jpg


Does this look normal to everyone?
 
Last edited:
John Bostwick said:
it does not sound like the 328 is bad, if you are able to get to the HM on the net. I would double make sure you have everything in correctly, for instance the Diodes have to be in the correct way with the line on them facing the same as the board has them. Check all the soldering to make sure you have no bridges and use a multimeter to verify continuity.
Click to expand...

multimeter + continuity = over my head. But so was this project... I better start learning!
 
I've looked over your older board photos and nothing really sticks out as being obviously no good. The fact that you can access the webui and adjust the backlight means the Pi is fine, the ATmega is at least partially OK and contains the right firmware. There probably won't be anything on the Pi's console either because this is something happening between the mega -> shift register -> LCD.

Continuity just means "connected", so if your multimeter has a setting that beeps when you touch the two probes together that's it. If not, you're looking for ~0 ohms resistance between connected parts. If you look at the HeaterMeter board image you want to check that things that have a red or blue line between them have continuity. You should do this with power disconnected mainly because if your hand could slip and make things more broke. I'd check all the pins on IC3 (the 16 pin chip) to where they go.

If that all seems to be ok then check the voltage on the LCD pins RS, E, D4-D7. You should see their voltage changing once every second. Some may stay high (3.3V), most will go back low (0V) once per second. Nothing there? Check voltage on the input to the shift register Pins SH_CP, ST_CP, DS, same story as those LCD pins. Also verify 3.3V at VCC. Still nothing? Check them coming out of the ATmega (pin 6, 11, 13 I think).

Everything good? LCD is good or the contrast is whacked. You can check the contrast by measuring LCD's CNT pin voltage. Should be around ~1V.
 
Bryan Mayland said:
I've looked over your older board photos and nothing really sticks out as being obviously no good. The fact that you can access the webui and adjust the backlight means the Pi is fine, the ATmega is at least partially OK and contains the right firmware. There probably won't be anything on the Pi's console either because this is something happening between the mega -> shift register -> LCD.

Continuity just means "connected", so if your multimeter has a setting that beeps when you touch the two probes together that's it. If not, you're looking for ~0 ohms resistance between connected parts. If you look at the HeaterMeter board image you want to check that things that have a red or blue line between them have continuity. You should do this with power disconnected mainly because if your hand could slip and make things more broke. I'd check all the pins on IC3 (the 16 pin chip) to where they go.

If that all seems to be ok then check the voltage on the LCD pins RS, E, D4-D7. You should see their voltage changing once every second. Some may stay high (3.3V), most will go back low (0V) once per second. Nothing there? Check voltage on the input to the shift register Pins SH_CP, ST_CP, DS, same story as those LCD pins. Also verify 3.3V at VCC. Still nothing? Check them coming out of the ATmega (pin 6, 11, 13 I think).

Everything good? LCD is good or the contrast is whacked. You can check the contrast by measuring LCD's CNT pin voltage. Should be around ~1V.
Click to expand...

Thanks Bryan. So continuity looked good for every pin on the IC3 to whatever it was connected to (and beyond where applicable). I checked the PINS you mentioned at the LCD, all read 0v (including the CNT). Traced thosed back to their corresponding pins on the shift reg, same thing 0v.

tried to trace that back to the atmega, but got a bit confused by the pin numbers so i just tested each pin. Here's what I got:

2vijmo5.png


so it sounds like the problem is in around the Atmega or it's connections, correct?
 
Also, i'm going to check this when I get home, but i couldve sworn that #2 (2nd from the bottom right) was ~0v, not 3.3v

So let's say pin 11 & 13 check out, but pin 6 doesn't, but there's continuity, what does that mean?
 
OK, so I was wrong about pin #2, it's reading 3.3v. It was pin #14 that was lower than stated above at less than 1v... Anybody know what the outputs should be?

AND if the voltage is lower than it should be, but continuity is good, is it a matter of creating a better solder joint? or something else?

Sorry for all the noob questions. I'm a novice when it comes to electronics, and I'm trying to learn as much as I can well reach the finish line on what's ultimately been a great project.
 
So the new atmega I ordered a while back showed up today, and wouldn't you know. Installed the chip, fired it up, and about a min later...

No Pit Probe

BOOM! No idea what was wrong with the other chip, or how it got that way, but at this point I don't care.
 
You must log in or register to reply here.

 

Back
Top