Help with a weird problem on Ice Tube Clock Moderators: adafruit_support_bill, adafruit

[Frank_tt]

I just built my ice tube clock and it could be that I messed something up. I am one of those idiots that put in C8 and C9 in across from each other but I fixed it. My problem is when I plug in the unit, the numbers show up dim for a couple seconds and flash as they are supposed to, but then the segments start dropping off line and after a few more seconds the display is dark. It is as if something what drifting off voltage or something of that nature. If I let it sit for a while then when I plug back in the display starts off good again and drifts off after about 4-5 seconds.

It beeps on start up and shows numbers, Voltage from Gnd to stripe end of D3 is 13.5

Any suggestions?

Curiously Pin 28 on the Atmel part came with no tip on the it. It had enough to make it into the socket, just a FYI.

Thanks!!

[adafruit_support_bill]

If you post photos of the front and back of the boards we'll see if we can spot any problems.

Voltage from Gnd to stripe end of D3 is 13.5

That is at the very low end of normal.

Curiously Pin 28 on the Atmel part came with no tip on the it. It had enough to make it into the socket, just a FYI.

This pin is not used in normal operation of the clock. Some people use it for adding sensors or other hacks.

[Patrick #13]

I am having the exact same problem. Upon plugging in, everything is lit and functions properly. Within 30 seconds to 2 minutes, illuminated segments of the ice tube start flickering and quit working. Soon the entire thing is non-functioning as far as the ice tube goes. I've looked over all my soldering with a magnifying glass very closely. Everything is soldered rock solid. All my voltages are correct.

**I thought the problem may have been that I used "Oatey #5" soldering flux. It's old and the label is unreadable. I assumed the problem may be that the flux is conductive. Anyway, I used both ethanol and isopropanol (99%) to clean all the flux off. I cleaned it 6-7 times because every time went back and plugged the ice clock in, it worked initially. I had the hair-brained idea to dip the VFD driver end into a cup of isopropanol after the ice-tube faded and voila!...instantly all the segments of the ice-tube illuminate properly. This works like...well, clockwork (sorry for the pun). I removed the VFD chip and reseated it and the problem still persists. I'm assuming the problem is flux that seeped through the holes under the VDF from the back side is messing things up. I'll work at cleaning it again under this assumption.

Do any of you have better suggestions?

[Patrick #13]

I cleaned it better using an artists paintbrush with nylon bristles due to their fineness. The space under the VFD unit was very small as is the space under the JP1 (where the ice-tube plugs into the board). I cleaned under the VFD (for like the 6th time) and still had the flickering. I really jammed bristles of the brush as far as I could get them under the JP1. I blew out as much of the alcohol as I could manage from underneath and so far, I've got success. The clock is not flickering, but it's missing a single segment still. It's noticeably better by a long-shot.

My alternate plan, having researched "Oatey #5", is to use either paint thinner or acetone to dissolve the residual flux if the problem comes back. The main ingredient of that kind of flux is petrolatum. The problem with acetone is it dissolves plastic, so you'd want to test a drop on JP1 before going full-bore on the acetone use. I considered using brake parts cleaner because it comes in a spray can with the straw that attaches to the nozzle in order to reach under JP1.

[phild13]

You already figured this out but for those who don't know:
The Oatey #5" soldering flux contains zinc chloride which is corrosive (an acid flux) and is meant for soldering copper pipes together. This flux is not suitable for situations where all residue cannot be cleaned away, such as in electronic work. If all the flux is not removed, it will completely ruin the board by getting under the boards coating corroding and eating away the copper traces which will cause both shorts and opens as things corrode. This is also true for those "water soluble" type electronic fluxes. They remain corrosive at room temperature and even though they can be rinsed with distilled water, they tend to leave a residue on the board and under components that remains corrosive and should be scrubbed off with alcohol and a toothbrush.

Using the electronic grade isopropanol alcohol (the 99% kind, not the drugstore kind) and an old toothbrush to scrub the board will help a lot.

Strong solvents like Acetone, MEK, etc. will likely dissolve the boards protective coating and will also likely dissolve some plastic parts. Acetone will also dissolve the fiberglass board itself.

I would suggest that if you have a hot air rework station or a heat gun you can probably unsolder the VFD chip socket without damage and clean the board and the bottom of the socket.

For most people, using a quality resin core solder or a quality "no clean" type solder is best for electronic work.

[Patrick #13]

I'm still having the problem and I think it's the VDF chip. I soaked that end of the board in mineral spirits overnight to remove any remaining flux. Still the problem persists. But I'm curious, why does it work perfectly when the VDF chip is immersed in alcohol if alcohol is not conductive? I like your advice to resolder the chip holder. I now have some proper solder for electronics too. I wonder if the original poster used a similar flux or just suffering from a different problem.

[Frank_tt]

I dont think this is my problem as I bought brand new electrical solder from Fry's for this project.I will take a picture in the near day or so of the front and back but I went through all the steps to make sure I didn't miss anything. I looked at the entire board under a microscope but didn't see anything. I will try to resolder everything just to make sure I don't have a cold solder somewhere.

any other voltages to check etc?

thanks.

[adafruit_support_bill]

@Frank_tt - Based on the symptoms, I would suspect that your Q3 is installed backwards. There have been some variations in the packaging for this device recently and it can be a little confusing to figure out which side is which. Installing it backwards won't hurt the device. Just de-solder and flip it around.

@Patrick #13 - Acid-core can make a mess of a circuit board. But it sounds like you have cleaned it pretty thoroughly. If you post photos of the front & back we'll see if we can spot any problem areas.

[Frank_tt]

Sorry it has been crazy, I finally can get the pics up.

I haven't done anything yet on the board but please take a look.

[Frank_tt]

back side

[adafruit_support_bill]

According to the photo, your Q3 looks correct, but we have had a few that had the markings on the wrong side. To test for a reversed Q3:

1.) Make sure you have a fresh batter installed (A reversed Q3 can drain the battery prematurely)
2.) With the clock unplugged, the Q3 lead closest to the edge of the board should read 0 volts (You can use the big tab on the voltage regulator as your ground reference)
4.) If you get a voltage reading significantly higher than that, then Q3 is reversed.

[Frank_tt]

OK, checked it, it shows 0 volts. Just FYI, the pin opposite on Q3 is showing 2.75.

Any ideas?

Also I noticed when it starts flashing the numbers for the first few seconds (while I can read them) it is counting every 2 seconds..... flash(08)....flash(10).....etc.

[jarchie]

My problem is when I plug in the unit, the numbers show up dim for a couple seconds and flash as they are supposed to, but then the segments start dropping off line and after a few more seconds the display is dark.

How exactly do the digits fail? Another user posted a video of what might be the same problem. Is that what's happening in your case?

Also I noticed when it starts flashing the numbers for the first few seconds (while I can read them) it is counting every 2 seconds..... flash(08)....flash(10).....etc.

With the Adafruit firmware, the clock will normally beep and flash the time whenever the clock powers up or wakes from sleep. The flashing should go away when the time is (re)set. What you're seeing here is probably normal.

[Frank_tt]

Mine does something very similar, the first couple of flashes you can see the entire set of digits but mine drops off segments much faster than in the YouTube video. Take that video and hit fast forward. After about 3-4 flashes , half the segments are gone, by 6 or so most are gone, couple more and it's dark.

[adafruit_support_bill]

Let's try replacing the tube then. Contact support@adafruit.com with a link to this thread and we can send you a new tube and side-board.