Nixie Clock Trouble

Thread Starter

SMB84

Joined Jan 30, 2010
8
Electronic Gurus - I really hope someone can help me here. I decided a little while ago to build my own nixie clock. For those that are unfamiliar, please take a quick look (this one is not mine): http://www.youtube.com/watch?v=KpYnzbQG7H0.

Before I describe my problem, let me explain how my circuit is supposed to work. To save pins on the microprocessor, I will only power one bulb at a given time. The processor will cycle which bulb is lit so quickly that it should appear that all the bulbs are lit at once (many inexpensive modern LED clocks use this type of circuit). For this to work, I have wired all the cathodes on the bulbs together and left the anode leads independent. The bulb receiving power to its anode should be the bulb that illuminates. Grounding a specific cathode lead would select the digit.

The problem is that whichever bulb I power, there is always a second bulb partially lit (showing the same digit). I'm not really sure how this is possible. The other bulb is grounded on the cathode but is not hooked up to any positive power on the anode. My best guess is that there is some sort of electrical interference causing the bulb to illuminate.

I posted a video to YouTube to show what I'm talking about: http://www.youtube.com/watch?v=gDAx2vqAWz0

I would really appreciate any help you guys could provide.

Thanks,
SMB84
 

hgmjr

Joined Jan 28, 2005
9,027
The open connections leave your nixies subject to unintended illumination. Have you tried hooking all of your nixie cathodes to the drive lines? Do you still have the problem when you do?

hgmjr
 

Thread Starter

SMB84

Joined Jan 30, 2010
8
hgmjr, I could do that but it would defeat the purpose of this type of circuit. Doing so would cause whatever digit I am sending to show on all bulbs.

If it were 12:00:00, I would send a "1" to the first bulb, then a "2" to the second bulb, etc. If they were all powered then all bulbs would show 1, then all 2, etc.

SMB84
 

hgmjr

Joined Jan 28, 2005
9,027
It sounds like more than one of your nixies is being powered by your high voltage source. Normally only one of the nixies would see the 170V.

hgmjr
 
I've been working on a nixie clock too (mine is a wristwatch format). I'm planning to use the same strobe technique you are. I'm not as far along as you, but I think you might need to put a pulldown resistor on each nixie anode. Given your other circuit values, I'd try something in the 100K~270K range.

Please post your results too, whether the pulldown or something else fixes your problem, because I'm going to have to do the same thing!
 

eblc1388

Joined Nov 28, 2008
1,542
Why the problem only manifest itself on the second leftmost digit?

As shown in your youtube video, lights up one digit. Then use a DVM to measure all NIXIE tubes anode voltage to see if the second one has anode voltage even if not being driven. if so, the PNP MPSA92 might be leaking. Also clean the board thoroughly on that nixie connection.

Also you can try disconnect the anode control signal of that "problem" digit from the BASIC Stamp2 pin to check if there are narrow pulses on that STAMP pin that switches on the anode drive of the second digit.
 
Last edited:

Thread Starter

SMB84

Joined Jan 30, 2010
8
I should clarify a few things I left out.

First, I have not actually completed building the circuit yet; I am simply sending ~170V directly to an anode and grounding the appropriate cathodes for a specific digit (cathodes are all wired together for 'like digits'). I am doing this manually to bypass the decoder chip and the whole "switching" circuitry with the transistors.

Spacewrench, This morning I built some of the missing circuitry thinking that the problem may have been related to not have a pulldown restor that you mentioned as I was not using the 10k resistors in series with the 170V shown in my diagram as R14, R5, R10, R25, R32, and R37. Doing so did help, but I had to add additional resistors to completely eliminate the problem. I didn't have the correct value handy though so I dropped the power a little too low causing the bulb to pulsate a little (not sure why that would happen) but no other bulb lit.

I retested my power supply circuitry and discovered that I am putting out ~178V when drawing current. My plan is to change the value of R33 (bottom-right, power schematic) until 170V is obtained. I am hoping this will fix the problem.


eblc1388, the problem does not always manifest in the left most digit. It can happen in any of the digits actually. In fact, switching the power on/off quickly can cause the "leak" to jump to another bulb. The problem bulb does read ~17V but I wasn't even using the circuitry mentioned earlier. The anodes were physically left floating.
 

hgmjr

Joined Jan 28, 2005
9,027
One thing that I can see from your video is that the phantom digit never happens on the nixie tubes that you have not put any leads on yet. I am specifically referring to 10-minutes and the minutes tubes. That tells me that the problem is pickup. Nixie tube filaments need only microamps of current to cause them to glow lightly. I think you need to go ahead and finish construction of the circuit. Then you will most likely not see the problem. As long as you have floating leads everywhere you are probably going to get squirrely behavior. Not to mention that you run the risk of getting a nice little shock from the 170V.

hgmjr
 
...may have been related to not have a pulldown restor that you mentioned as I was not using the 10k resistors in series with the 170V shown in my diagram as R14...
Actually, I was going to ask about those 10ks...my plan was to just use PNPs to pull up to V+ directly. AIUI, on a neon bulb, you have to get the anode-cathode ΔV above a certain value (the strike voltage) to get the gas to ionize, but the voltage necessary to maintain the arc is somewhat lower. The 10K resistor functions as a ballast: when no current is flowing, it doesn't drop any voltage, so you can get to the strike voltage. When current starts to flow, the 10k dissipates some power and reduces the anode-cathode voltage. However, this causes a couple of problems: first, you waste power in the resistor; and second, you don't have direct control over the bulb current (I think bulb brightness is proportional to current, or maybe power).

Anyway, I had planned to have my micro control the boost converter via a PWM output, and monitor both voltage and current. To turn on a digit, I'd increase voltage until I saw current jump up, then reduce voltage to get to the target current (so I would have an active lamp ballast, instead of just a resistor). I suppose it would also be possible to PWM the digit cathodes instead of adjusting V+. If you did it quickly enough, you could turn them off, then on again before the ionization had completely faded, so you wouldn't have to apply the full strike voltage.
 

THE_RB

Joined Feb 11, 2008
5,438
It's a "carry-over" effect caused by capacitances etc and system timing.

Try adding a "dead time" of a few percent where the bulbs are all OFF between multiplexing.

So instead of multiplexing; 1-2-3-4-

Do this; 1-deadtime-2-deadtime-3-deadtime-4-deadtime-

Hope that makes sense. :)
 

Thread Starter

SMB84

Joined Jan 30, 2010
8
THE_RB,
As I mentioned in a previous post the circuitry is not 100% complete yet. The microprocessor isn't even being used yet so there is no multiplexing. Instead, I have been manually feeding the high voltage to the nixie bulb.

Spacewrench,
I'm not too focused on efficiency since the clock will be plugged in to an outlet. Had this been a battery powered design I would have reconsidered.

I did go ahead and change some resistor values. I was able to get the voltage down to ~171.5. Then, using a 47k resistor rather than the 10k I have no more phantom lights.

It looks like my problem is fixed. I'm still not 100% sure why it was happening in the first place but its working now and I'm happy. Thanks!
 
Top