Ice Tube Clock boost problems

[drjeseuss]

I built the Ice Tube Clock from scratch, etching my own board from the design files, and flashing a new/blank atmega168v with the iv.hex file. My boost voltage is only reading 12v. I've checked all connections and parts. The supply voltage is 8.89v in and 5.00v out. The board beeps on powerup. I've used PICs many times before, but this is my first project with an AVR. When I flashed the chip, I used avrdude with -U flash:w:iv.hex to write. Was I also supposed to do -U eeprom:w:iveep.hex? Was I supposed to change any fuse bits? If so, what should they be set to? I'm used to PIC where all this is done from a single .hex file (which may still be the case here). I'm worried that the low voltage may be due to the chip running at a slower speed than expected, causing the boost circuit to run slow and low... maybe this is not the issue. Suggestions?

[jarchie]

Adafruit support won't answer this question because this isn't an Adafruit kit.

My boost voltage is only reading 12v.

With the tube inserted, I believe this voltage is in the normal range. With the tube removed, I would suspect your microcontroller clock is running too slow.

Was I also supposed to do -U eeprom:w:iveep.hex?

I believe so.

Was I supposed to change any fuse bits?

Yes. See the Makefile on GitHub, but I believe the fuse settings should be

Code: Select all

avrdude [...] -u -U lfuse:w:0xE2:m -u -U hfuse:w:0xC6:m -u -U efuse:w:0x01:m


If you have an ATMEGA328P-PU lying around, the installation instructions are a bit clearer for my unofficial firmware.

Good luck!

[drjeseuss]

Adafruit support won't answer this question because this isn't an Adafruit kit.

Ouch, that's harsh. I appreciate your response. As said, I know little of the AVR world, but browsing the makefile as you suggested gave a lot of answers. It looks like I still need to program the iveep.hex and set the fuses. I'll give that a go and see what happens. Once I get things working I'll likely play with a few of the other available firmwares out there, but not until I can get a 'default' setup going. Thanks again for the pointers.

[jarchie]

Ouch, that's harsh. I appreciate your response.

Sorry. I figured if I "yelled" at you, others wouldn't. And since I am an Adafruit customer, I suspect they'll let this thread slide. But their policy is here.

Since I happened to have a terminal window open, I went ahead and issued the installation command on the Adafruit Makefile. You've probably already figured this out, but just in case:

Code: Select all

% make -n full
avrdude -p atmega328p -P /dev/ttyACM0 -c usbtiny -u -U lfuse:w:0xE2:m -u -U hfuse:w:0xc6:m
avrdude -p atmega328p -P /dev/ttyACM0 -c usbtiny -B 5 -U flash:w:iv.hex
avrdude -p atmega328p -P /dev/ttyACM0 -c usbtiny -B 5 -U eeprom:w:iveep.hex


[adafruit_support_bill]

Ouch, that's harsh

Harsh? We publish the fully documented design, source code, board files, step-by-step assembly instructions and troubleshooting guides. Plus all our support threads are open for all to see. That is already more than you get from a lot of places - even if you do buy their kits.

You are welcome to discuss your clock projects here. There are plenty of knowledgeable folks like jarchie that know the Ice Tube quite well. But Adafruit engineers need to focus on supporting our customers.

[jarchie]

Sorry. I figured if I "yelled" at you, others wouldn't.

Oh well. So much for that strategy.

You are welcome to discuss your clock projects here.

I figured Adafruit would be willing to allow the use of their servers and bandwidth for this discussion, even though neither of those resources is free. I am rather impressed with them as a company.

So feel free to ask any other questions and hopefully someone can help!

[drjeseuss]

I suppose I should point out... I also am an Adafruit customer, just not on this product. As testament to their kits and products I've never had to reach out for help. I've learned a lot, and on this build I wanted the challenge of doing it all from scratch (minus the design). That said, I'm fine with the knowledge of the other forum members only.

I've managed to set the fuse bits and flashed (and reflashed) the flash and eeprom memory. I still get my post beep, but still only getting 12v out of the boost circuit. With confidence the chip is proper I'm going back to the board to check for abnormalities. As said before, this board was home-etched, and being two-sided I had a challenge of vias. I'm rechecking any top to bottom connections, and looking for any cold joints (though on prior inspection they all looked good). If you have any suggestions of test points and expected results there to isolate the problem to a particular part that might help. I'm also going to hook up a logic analyzer to check frequency pulsing the boost circuit. Hopefully I'll find the issue on next look. I've been unable for a few days as my fine-tip pencil ironburnt out, and my backup iron is more like a sledge hammer on a board this small, so I had to refrain from redoing any of the joints until I could get a new pencil iron today. We'll see what happens. I'll post my success (or failure) in case it might help others down the road.

[jarchie]

I've managed to set the fuse bits and flashed (and reflashed) the flash and eeprom memory.

As long as avrdude did not spit out any errors, I would be quite surprised if the problem is with the ATMEGA168V programming. If you have any doubts, I'm willing to mail you a known-good chip.

I still get my post beep, but still only getting 12v out of the boost circuit.

On my clock, I set the brightness to 1 (the default brightness for the Adafruit firmware), and the boost voltage was 13.3v without the display attached. A reading of 12 volts is probably okay. I would swear that the voltage was higher when I first assembled the kit, but I trust my multimeter more than my memory. My apologies for leading you astray.

I'll post my success (or failure) in case it might help others down the road.

I'm sure it will help others down the road, so thank you for that. Less importantly, I get quite an ego boost when I think I've helped someone with a project.

[drjeseuss]

Several touchups of questionable joints, but still the same results. Here's what I've got...

@7805 - 12.4v in, 4.9v out
@Q2/L1 Junction (bridged pins of Q2) - 12.4v
@D3 - 12.4v
@IC1-12(Boost) - 0.00v It appears the boost pin is staying low which would simply pass my 12.4v across L1 to C5/C6 and out to +UB... unboosted.

I'm now (again) questioning whether the atmega168 is doing it's job. Aside from the short beep on startup, any other life signs I could look for? Any other thoughts why the Boost pin would remain low instead of cycling? Is this possibly "intended" behaviour such as in power fail/battery mode? I currently do not have the coin cell or VFD chip installed. Everything else is in place.

[drjeseuss]

On my clock, I set the brightness to 1 (the default brightness for the Adafruit firmware), and the boost voltage was 13.3v without the display attached. A reading of 12 volts is probably okay. I would swear that the voltage was higher when I first assembled the kit, but I trust my multimeter more than my memory. My apologies for leading you astray.

This is good to know! I may not have a problem after all then. I still need to solder up the tube so I'll do that in the next day or so, then pop in the VFD chip and see what happens. I greatly appreciate your offer to send a chip. Hopefully I'll be able to pass on the offer if all goes well, if not I may take you up on it. If so, I'll gladly ship you a replacement in exchange.

As for the ego it's well deserved... You've been a big help so far.

[jarchie]

@7805 - 12.4v in, 4.9v out
@Q2/L1 Junction (bridged pins of Q2) - 12.4v
@D3 - 12.4v

I'm not sure what output you're referring to on the 7805, but the other values seem okay.

@IC1-12(Boost) - 0.00v It appears the boost pin is staying low which would simply pass my 12.4v across L1 to C5/C6 and out to +UB... unboosted.

I believe you may be mistaken here. The N-Channel HEXFET will block current when the ATMEGA's boost output is low. At default brightness, boost should spend ~12% of the time pushed high and 88% of the time pulled low.

I suspect that if the HEXFET were always "on", the output of L1 would be perpetually pulled low and the boosted voltage would be zero. But the fuse would probably trip under the load. Since you're getting boost voltage, I suspect everything is functioning normally.

To be certain, maybe try checking the voltage from the 9 volt power adapter? I'll wager it's less than 12 volts and that boost is just fine. Maybe also check the ATMEGA's output to Q3. If that's pulled low, Q3 is "on", the clock is awake, and your multimeter is probably just reading the ATMEGA's PWM cycling as zero volts. If pushed high, the clock is probably sleeping.

[drjeseuss]

You were right again on all counts... The duty cycle of BOOST wasn't reading properly on my multimeter (no surprise). Once connected to my logic analyzer, I verified the cycling of BOOST. In the guide, it suggestes that when testing boost output, the reading should not be below 40v. This is misleading, mine was 12.4. I decided to go ahead and populate the VFD driver IC and tube. On powerup, success! Kind of... It fired up dimly as you noted it would, but I have a few lines (VFD2 and VFD18) that aren't active. Since this is a homemade board vias and top-trace connections were difficult under the PLCC socket. I'm seeing if I can do some corrective surgery on it, if not I'll be removong, fixing, and replacing the PLCC socket. As said before, thanks for all the help with this. I'll update again as I make progress, and hopefully eventually a fully functional photo.

[jarchie]

On powerup, success! Kind of...

Bravo!

board vias and top-trace connections were difficult under the PLCC socket. I'm seeing if I can do some corrective surgery on it, if not I'll be removong, fixing, and replacing the PLCC socket.

Sounds like you've got everything under control.

For connecting broken traces, I've had good luck with silver pens, although I haven't tried the Adafruit one. Since the bottom of the board doesn't show, making the connections with wire is always an option. I did that for the extended battery hack and am happy with the results.

In the guide, it suggestes that when testing boost output, the reading should not be below 40v. This is misleading, mine was 12.4.

Now that you mention it, I vaguely remember being concerned at that step... and I had an official Adafruit kit!

As said before, thanks for all the help with this. I'll update again as I make progress, and hopefully eventually a fully functional photo.

Awesome! I would enjoy seeing how everything turns out.