Why can't this thing just work... (frustrated)

[Michel Kraaij]

Why is this thing not booting normally.... i'm getting really frustrated and i'm about to throw this entire project in the bin.
It worked fine for a while, but often it just doesn't want to boot. Sometimes a blanc screen. Sometimes a few leds lit, sometimes the squares. But almost never a normal boot. I've already replaced the RPi 3b with an RPi ZW and verified any mistakes on my soldering part. But i can't seem to find the issue.

Please, anyone?




And 5 minutes later it boots up again....



And crashing again during the session....

 

[Bryan Mayland]

I don't have any insight apart from saying that it should not do that. The only HeaterMeter I've seen that has behaved in such an unpredictable manner was due to using plumber's soldering flux during assembly, which left an invisible destructive film on the board that was eating copper and conducting between things that should not be connected. It also was burning out Pi's and 5V regulators, the servo output was wrong, it would show weird temperatures sometimes, etc.

 

[Michel Kraaij]

No soldering flux used here. The weird part is that it (sometimes) does function. This morning, after a dozen failed startups, it suddenly booted. The only weird thing that i could detect is that probe 1 wasn't showing up. This probe had my custom ABC settings. After erasing this settings the probe did showed up. The other thing i noticed is that my set temp was at 900+ degrees C. My last bake off was a nice roast, but definitely not that roasted. After correcting that it worked without problems. As a matter of fact, it's running as we speak it's running for 8 hours straight now (https://heatermeter.michelkraaij.nl/).

I really don't get it...

 

[JKalchik]

Michael (or Mike, if you prefer....) Can you post a pic of the solder you used? Acid core solder is used for plumbing and is decidedly detrimental to circuit boards. Electronics use flux core. I know you said you didn't use flux, but that should be in the core of the solder you did use.

Maybe post a pic of the solder side of the board as well.

 

[Michel Kraaij]

I think i used the top one, but could have been the bottom one. And yes, there actually is some flux in it, but to my knowledge this is regular electronic solder. No additional flux was added.

And i assume that if something actually would have shorted out, it wouldn't be functioning right now...



PS. It's Michel. The French version of Michael ;-)

 

[Larry D.]

Intermittent problems are the worst. You checked your soldering; did you check the other soldered connections on the board? Cold soldered connections can create intermittent problems, electronic "noise". https://handycrowd.medium.com/easil...-joint-on-printed-circuit-boards-50c53cc974a1

 

[JKalchik]

It's Michel. The French version of Michael ;-)

Oops.... my mistake, apparently I need to learn to read.....

Given the very fine apparent diameter, I'd suspect it's the right stuff, but I don't see on the label if it's acid or rosin (proper term..) core.

 

[Bryan Mayland]

Yeah the pipe soldering stuff tends to be bigger, you're right. I can't read the label on the top one but I'm going to assume it is ok. Just wanted to make sure you didn't slather everything in the plumbers stuff, because that's a real recipe for disaster and causes all sorts of mystery problems.

I'm really at a loss though what it could be. I know you've been struggling with this for a while now, so if you want to PM me your address I'll solder you up a replacement myself and send it out to you in Wednesday's mail. Take a while to get to you, but hopefully it can turn your experience around.

 

[JKalchik]

Michel, post up a high res pic of the soldered side of your board. Something may leap out at the nerds who hang out around these parts.

 

[Gary V]

Michel,
Have you tried a different power source and a new sd card. Those item could possibly cause intermittent problems.

 

[Michel Kraaij]

Yeah the pipe soldering stuff tends to be bigger, you're right. I can't read the label on the top one but I'm going to assume it is ok. Just wanted to make sure you didn't slather everything in the plumbers stuff, because that's a real recipe for disaster and causes all sorts of mystery problems.

Both tubes are rosin core solder for electronics. It's definitely not the pipe soldering stuff. Well, i can try this solder for the pipes, but i'm pretty sure the end result will be an uncontrolled waterfall in the bathroom

I'm really at a loss though what it could be. I know you've been struggling with this for a while now, so if you want to PM me your address I'll solder you up a replacement myself and send it out to you in Wednesday's mail. Take a while to get to you, but hopefully it can turn your experience around.

I appreciate the offer. Let me first explore a bit more. I'm a professional software tester. I don't give up that easily. The last few times i powered up the device, it was booting without issues. Could it be that those custom A-B-C settings for probe 1 are interfering in some way?

If nothing works, i'll send you a PM. I'm still a fan. Don't worry.

Michel, post up a high res pic of the soldered side of your board. Something may leap out at the nerds who hang out around these parts.

I hope it's 'high-res' enough. Had some trouble with lighting.



It's no flawless soldering, but i can't find any issues so far (except for hitting the plastic at some point and doing the infamous Q1 fix)

Michel,
Have you tried a different power source and a new sd card. Those item could possibly cause intermittent problems.

I thought of that too and have tried it. But it made no difference. I even tried different wall outlets all around the house.

 

[Bryan Mayland]

Could it be that those custom A-B-C settings for probe 1 are interfering in some way?

Internally, all the probes use "custom" Steinhart–Hart coefficients and only the webui checks them to see if they're pretty close to one of its presets and then selects the preset instead of Custom, so that for sure wouldn't cause any issues on the HeaterMeter itself.

Just verifying, when it gives you the row of boxes on the LCD on boot... does it actually boot and just the LCD isn't working? Or is it actually not booting at all and you never can connect to the webui either?

 

[RalphTrimble]

Michel, I understand your frustration.

In a situation like this I take the "shotgun" approach. Remove the rPi, pull the ATMEGA, then with a clean soldering iron QUICKLY reflow the solder on the entire board. If you soldered heavily the first time then a reheat to reflow is probably all you need, if you soldered lightly the first time add a little solder on the second pass. When you reheat, wiggle the leg of the component a little with the soldering iron, solder melts, leg wiggles, pull iron off quickly. Again, QUICKLY, you dont want to reheat and burn up all your parts. Wiggling the leg a bit helps the solder flow to the traces on the other side of the board.

After you have done this take some isopropyl alcohol (or flux remover if you have it) and clean the board with a soft bristle brush (old toothbrush will do). After letting the board dry thoroughly reinstall the ATMEGA and boot up the HM on its own, see if it loads properly? If so, power cycle a few times just to get a feel if it is now reliable? If reliable then the HM board is good. Now connect the rPi and boot, did things turn bad? Hopefully not! If the HM wont boot it is possible that bad connections caused the program on the ATMEGA to get corrupted, I've seen this more than once. Connecting the rPi will sometimes get the ATMEGA re-flashed, if the ATMEGA refused to operate there are some things you can try wipe it clean so it can be flashed properly. Cross that bridge if you come to it.

Also make sure you don't have any long component legs that can bend and touch other component.

The HM doesn't need the rPi to function (after the first boot when the ATMEGA gets flashed). So best troubleshoot this without the rPi connected. You should get display and temps from the probes, menu and fan will function, everything will function without the rPi except for the web interface. I have built several versions of the HM board, one time I had a similar problem but only when the rPi was connected. Turns out the rPi leaned into the HM board enough to touch some components, I put a piece of paper between them and never had another problem. Troubleshooting without the rPi connected led me to that solution...

Good luck to you...

 

[Dave Smith]

check solder joints circled in yellow

 

[Darren L]

 

[Darren L]

Check the one in blue also.

 

[Larry Naylor]

Any more progress???

 

[Michel Kraaij]

Lately i've been leaving out the HM when i'm cooking. I haven't been able to find the problem. I made another attempt using the HM, but it freaked out on me again.

Here is an example of the HM acting up. Only connected are a thermocouple and a poweradapter.
HeaterMeter acting up

 

[Gary V]

Michel,
Have you tried removing the PIC and making sure it did not have a bent pin when it was inserted into the socket? Watching it boot, it looks like a problem I had one once when the PIC had a pin that was not firming inserted into the socket. Also do you have a spare PIC to try.

 

[Michel Kraaij]

Michel,
Have you tried removing the PIC and making sure it did not have a bent pin when it was inserted into the socket? Watching it boot, it looks like a problem I had one once when the PIC had a pin that was not firming inserted into the socket. Also do you have a spare PIC to try.

Help me out a bit. The PIC?