Ok, I scoped the base and the collector at the same time for the same transistor, this time with power applied to the tube (I cut the power to the tube where it drops off) and got a very interesting result, the red line is the base set to 100mv, the blue is the collector set to 10V, it looks like the transistors are not turning off but are on the whole time.

 

Ok, I'm royally confused.
Heres the form for the base of one of the transistors while power is applied to the tube, pasted together with paint. 100mV/div.
For size reasons it's monochromatic and a bmp, but you're gonna wanna look at it, it looks like the transistor latched up or something.

 

I must have miss read a value. I had 3.3k in my head.

To the problem at hand, Are you SURE the nixie is in good shape?

Do a quick check.

Pull the nixie and get a ohm reading from anode to cathode for EACH character.


They should be dang close.

Premature operation non-limited could cause problems with individual characters.

nixie wristwatch: http://www.dwf.nu/viewtopic.php?p=25846

 

I must have miss read a value. I had 3.3k in my head.

To the problem at hand, Are you SURE the nixie is in good shape?

Do a quick check.

Pull the nixie and get a ohm reading from anode to cathode for EACH character.


They should be dang close.

Premature operation non-limited could cause problems with individual characters.

nixie wristwatch: http://www.dwf.nu/viewtopic.php?p=25846

Shouldn't they all read as open circuits?

 

To the problem at hand, Are you SURE the nixie is in good shape?

Do a quick check.

Pull the nixie and get a ohm reading from anode to cathode for EACH character.


They should be dang close.

Premature operation non-limited could cause problems with individual characters.

nixie wristwatch: http://www.dwf.nu/viewtopic.php?p=25846

Dang close to what? ∞? Because that's what they all are, or at least greater than 2000K, which as high as my meter goes.

I'll try replacing the nixie, see what happens.

 

To each other. (infinity-esque)

 

To each other. (infinity-esque)

Yea, they're all over 2000K, beyond that I can't measure.

 

Ron, I see what you're saying about Icbo and agree with you. I also, have experienced liquid flux causing very bad leakage problems.

 

To each other. (infinity-esque)

Also, anode to each cathode should be non-conducting. Conduction doesn't occur below the ionization voltage.

 

Ron, I see what you're saying about Icbo and agree with you. I also, have experienced liquid flux causing very bad leakage problems.

Yeah.

Magnet18, have you checked for contamination (e.g., flux), as I suggested in post #37?

 

Not as of yet, I'll do so when replacing the tube and trying a new one.

I take it the scope traces didn't make any sense to anyone?

 

Not as of yet, I'll do so when replacing the tube and trying a new one.

I take it the scope traces didn't make any sense to anyone?

In post #41, they do not make sense. Your base voltage is only about 440mV, and has a 480mV blip in it. It looks to me like a wiring error, or maybe there is a microscopic short between traces, or something like that. I don't think it is a transistor problem. It might be due to a defective 4017. Are you powering the 4017s with 10V, 12V, or ???.

 

I thought it was interesting that the blips were exactly the same size, but the were set to totally different voltage/div levels
an additional .48 volts on the transistor when it was supposed to be on dropped 4.8 volts from the collector.
I don't think it's the 4017 but I'll check, I can see the numerals cycling through just like they would normally, the one that should be on is brighter, and it rotates through all normal like, they just won't turn all the way off.

I'll check the wiring again, theres no traces, it's on perfboard.

When I pull the nixie I might wire it up to LED's instead and see if that reveals anything. Obviously not using 180V. Or, I could use a giant resistor... I might do that...

 

For many, many years, when I needed more flux than the core solder could provide, I used 'Paste Flux' sold in a shallow tin. Then one day, about a year ago, I was prototyping a little Picaxe circuit in the Bio-Med Electronics shop, at a hospital I contract with. The flux they had was a clear liquid. I didn't give it much thought until I began testing the circuit. Nothing was working properly so I started wringing out the circuit with an Ohmmeter. WOW!!! I couldn't believe the leakage that stuff caused. I brushed the board down with mineral spirits followed by another cleaning with none residue electronics cleaner.

This story dealt with a mere 5V and much lower circuit resistances than this Nixie, so I can imagine what it could do in a circuit like this.

BTW, I stuck the Ohmmeter probes in the jar of this stuff and with about 3/4" between the probes this stuff measured < 7KΩ!!!

 

Here's another shot in the dark, but just in case the base is picking up noise, try placing a .1uF cap between the base and GND.

 

Is 1478 K ohms between the base and the emitter acceptable?

 

Is 1478 K ohms between the base and the emitter acceptable?

Depends on which way you have the meter connected, and what kind of meter you have.
Considering that the base resistor comes from the output of a CD4017, through a 33k resistor, the measurement should be different when you swap the probes, but the circuit is nonlinear, so it is hard to predict the "resistance" from base to GND.

 

Are you decoupling the 4017?

 

Modern DMMs (Ohmmeter mode) have an open circuit voltage below the knee of BJT base-emitter and base-collector junctions. It makes in circuit measurement of other components much more reliable. The 'Diode' test function is used to test silicon junctions.

 

Speaking of other things -
Magnet, your Reset AND gate (on the Hours schematic) sucks very much bad. When Q45 base is high and Q46 base is low, you want Reset to be low. Instead, it is in no-man's land, somewhere between a solid 1 and a solid 0.
If I were you, I would bite the bullet and use a CD4011. See the attached schematic.
You will only need the diode OR gates on the most significant digits of the minutes and seconds counters.