Another quick update.

Replacing the 10's hours tube worked fine. So I do indeed have two bad 
tubes. I've extracted the other bad tube (1's minutes) and will replace it 
this evening.

Once I got that going -- after tearing out what is left of my hair trying 
to get all 13 tube wires into the PCB -- I noticed that the decimal point 
on the 10's seconds tube was not lighting. Closer inspection revealed that 
not only was it not only not lighting -- it was missing all together. Looks 
like the connecting tab is there, but no cathode. No glow. Owner is not 
concerned.

<https://lh3.googleusercontent.com/-knlOkmTOlNQ/WC4DNOCnFDI/AAAAAAAAAEk/t0Nf0QLDVOAgyyJgft9B8aCPj6WIvE-kgCLcB/s1600/missing%2Bcomma.jpg>




Terry

On Wednesday, November 16, 2016 at 2:58:29 PM UTC-6, Terry S wrote:
>
> Excellent suggestions. Here is an update.
>
> I do have the schematic -- however I do not have permission from the 
> designer to publish it. I have ohm'd out the entire circuit, anode and 
> cathode sides.
>
> I did check the timing -- as you said, it was in the millisecond range, 
> and no tubes are being skipped.
>
> I finally did test the tubes with my sign tester. The working tubes lit up 
> with a nice orange glow. The two non-working tubes lit just slightly, with 
> a purplish glow. This told me there was something different about those 
> tubes.
>
> I extracted one of the non-working tubes from the board, and unlike a 
> known good tube, I could not get the individual cathodes to light with the 
> tester. So I am now convinced the clock has two bad or gassy tubes.
>
> As of yet, I've been unable to install a new tube -- having trouble 
> cleaning the solder out of the PCB holes. Not my clock, so I'm working very 
> hard to not damage the PCB.
>
> Thanks for the help -- I will post updates as I make progress.
>
> Terry
>
>
> On Wednesday, November 16, 2016 at 12:00:11 AM UTC-6, gregebert wrote:
>>
>> OK, all of the obvious culprits have been ruled out so now we have to 
>> consider weird things.
>>
>> Wish you had a schematic, but it's not mandatory.
>>
>> 1. With your scope, do you see similar on/off times at the base of each 
>> NPN 'predriver' ? It should be in the milisecond range. Trying to rule-out 
>> the possibility of the control logic 'skipping-over' the tubes that aren't 
>> lit.
>>
>> 2. This might be risky, but you could try momentarily connecting the 
>> anode of the bad tubes to the anode of a good tube. You may want to play it 
>> safe and use a resistor of a few K-ohms instead of a short. If all 4 tubes 
>> are off, there's a low-impedance path somewhere.
>>
>> 3. Connect a 10K resistor to +170V supply. This will not damage the PNP, 
>> and assuming it takes 150V to illuminate the tube, the 10K resistor will 
>> limit the current to a safe value of 2mA (1 mA per tube). It wont be a 
>> strong glow, but you will see it if the tubes are alive.
>>
>> Connect a voltmeter to the other side of the resistor, and GND.
>>
>> Next, momentarily touch the resistor (voltmeter side, not +170V side!) 
>>  to the dead tube's anode. You should see the 2 dead tubes glow, and see 
>> around +150V. If you see significantly less voltage, or zero, there is a 
>> short or low-impedance path somewhere. It could be the voltage divider, or 
>> a bad tube. You may need to unsolder one or both anodes to confirm
>>
>> If the voltmeter still reads ~170 when you touch the tube, and there's no 
>> glow, there could be an open somewhere. You'll need to buzz-out the 
>> connections between the tube's anode pin, anode resistor, and PNP. If that 
>> still fails, then I see no other culprit besides the tubes.
>>
>>

-- 
You received this message because you are subscribed to the Google Groups 
"neonixie-l" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to [email protected].
To post to this group, send an email to [email protected].
To view this discussion on the web, visit 
https://groups.google.com/d/msgid/neonixie-l/fd94d3e7-130e-49e7-8915-f9ce0744b3a9%40googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

Reply via email to