First nixie clock: do I need to switch anode?

[Justin Scott]

I want to build a nixie clock similar to this one:

https://www.dos4ever.com/TiT/TiT.html

I will supply HV to the anode of each tube (through a current-limiting resistor), and switch the cathodes using BJTs. The BJT's bases will be connected to a microcontroller's output through resistors.

I have read a few people online saying that in order to prevent ghosting, you must "blank" (turn all cathodes off) the tube when you change digits. This would be fine, except the same folks say you should never blank the tubes by turning all the cathodes off; instead, you have to turn off the anode. I don't really want to add a switch for the anode. Is this really necessary?

Also, can I just switch my digit transistors on and leave them on, or do I need to pulse them?

Thank you!

[gregebert]

Switching the anodes is what multiplexing does, the advantage being that you can share the cathode logic across several tubes. But with savings, there is also a hidden cost: You must run the anode current higher for multiplexed operation versus direct-drive. If the tube is specifically designed to support higher peak-current for multiplexing, then there's minimal risk; I recall some Burroughs tubes state in the datasheet not to use multiplexing.

I've done a number of clock designs, all of them direct-drive. Cost is a secondary concern; maximizing the life of the almost- irreplaceable nixie tube is the overriding goal. With direct-drive, you dont need to switch the anode. However, I have some designs that use anode current-regulators which is basically a switch that is not fully-on.

Ghosting only occurs with a muliplexed display, so if you are concerned about it, be sure that your design has programmable blanking-time, refresh-rate, and on-time. You will have to experiment to get the best results.

[Paul Andrews]

That is a one tube clock. You don't need to switch the anode.

[Paul Andrews]

I had seen that article before, but its significance to several of my interests hadn't really sunk in. Recently I have become more and more obsessed with power supplies and time bases. My father-in-law was an accomplished engineer. After retirement he sold various electronic gadgets, one of which was a small mains-powered photographic timer. I was amazed when he showed me inside one to see no chips and no transformer. So when I my interest in this stuff started, I was intrigued that I couldn't find any power supplies like the one he had used. Yet here it is. Furthermore, the guy is using the mains frequency as a time-base, something I am very interested in doing, since my current project is also mains-powered.

His article is very thorough, and this just got added to my to-do list of future projects, which is ever-increasing. I also want to use Osmond to design some simple PCBs on the Mac - I am interested in producing PCBs that are themselves a work of art (as far as I can make them), and I think this would be a great project for that. The more interesting things I can squeeze into one project, the more likely I am to work on it.

[Dekatron42]

One thing about the type of power supply that relies on a capacitor in the manner that the clock design in the link from www.dos4ever.com is that the capacitor will degrade over time for every voltage spike on the mains, this is as designed by the capacitor manufacturers but it also means that the capacitance will become lower over time as the self-healing properties are not 100% so the voltage you get from this power supply will get lower over time. There was a big problem with this a few years ago as the capacitors that were manufactured over a time degraded faster than designed so many home appliances went dead long before they were intended to (I worked with smart electricity meters that were also affected by this problem and it took a long time to find out that it was these capacitors that were the problem). Finally they found out that there were manufacturing problems of these polypropylene capacitors that meant that they degraded faster than designed. The costs for repairing the electricity meters were in part covered by the capacitor manufacturers but I guess that it was a loss to everyone in the end. So you should count on the capacitor giving up at some point, but modern capacitors nowadays take some of this into account as the manufacturers realized that they had to change the manufacturing process to make better and more reliable capacitors (but since cost has always been the driving point for these types of components they might not be so much better anyway).

Do some googling on for instance "smart meter capacitive power supply problems" and you'll see some discussions and documents on this.

Choose a specially designed capacitor for this use and your power supply will live a lot longer and don't just put any capacitor in there!

/Martin

[Paul Andrews]

Interesting. What about a ceramic capacitor? For example, there is this 400V AC X2 cap on digikey, though quite large at almost 25mm diameter.

[gregebert]

I dont recommend line-operated designs unless you've done some previous design work at  high-ish voltages. There are a lot of not-obvious things that go wrong (line noise, transients, component failure, inadequate isolation) with very bad consequences.

If you have done a lot of past designs, I can share some tips on how to make a line-operated design safer and more reliable. The last thing we want is for someone's nixie clock to cause a fire, or worse.

I've had a few "learning experiences" with line-operated circuits that ended-up with sparks, smoke, and/or small explosions despite careful forethought. There's always something you overlook, and sooner or later it will get you.

[David Pye]

I agree.

I can't help thinking that for a first clock a 9/12vdc powered design using a known HV module seems to offer the safest option for people without line voltage experience.  

David

--
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 neonixie-l+...@googlegroups.com.
To post to this group, send email to neoni...@googlegroups.com.
To view this discussion on the web, visit https://groups.google.com/d/msgid/neonixie-l/990de368-8092-4214-b133-8bbe0dcdf849%40googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

[Justin Scott]

Thank you! I completely understand now. My design will be direct-drive, so I won't need to switch the anode. Will I need to switch the cathode though, or can I just leave it on continuously?

On Thursday, May 2, 2019 at 10:43:20 AM UTC-4, gregebert wrote:

[Justin Scott]

Thanks; good advice. I am an EE with 12 years of experience and have worked with HV before. Built my own tube amp with 420 VDC inside. Would love to hear your tips.

What I'm actually building is a volume display for a tube amp. When you turn the volume pot, a couple of nixies will display 0-99. The amp in question already has a transformer that delivers 220VAC, so I will be dropping and rectifying that to get my 180V for the nixie anodes.

[Paul Andrews]

I’m going to say the obvious thing here: Surely it should show the volume as 0-11?

--
You received this message because you are subscribed to a topic in the Google Groups "neonixie-l" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/neonixie-l/gIa9ngzJVKk/unsubscribe.
To unsubscribe from this group and all its topics, send an email to neonixie-l+...@googlegroups.com.

To post to this group, send email to neoni...@googlegroups.com.

To view this discussion on the web, visit https://groups.google.com/d/msgid/neonixie-l/64ede73e-aabd-413e-9528-2ffb59db79f4%40googlegroups.com.

[Justin Scott]

Why don’t I just make 10 a little louder and then make 10 the top number and make that a little louder?

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 neonixie-l+...@googlegroups.com.

To post to this group, send email to neoni...@googlegroups.com.

To view this discussion on the web, visit https://groups.google.com/d/msgid/neonixie-l/A010590F-9635-4218-880C-3531B2B18793%40gmail.com.

[Jonathan Peakall]

Because 11 is louder. It is one more than 10! Duh!

To view this discussion on the web, visit https://groups.google.com/d/msgid/neonixie-l/CAM%3DKN7yKdNDZ-UdHj%3DEY8hxopd8baoeTAaK5HQiKFvLqgQfJ8Q%40mail.gmail.com.

[Paul Andrews]

Perhaps I misunderstand, but each digit is a cathode. To turn it on, pull it ground, to turn it off let it float or push it to around 80V.

[Justin Scott]

Right. Or to turn a digit on you could switch the cathode quickly, so there would be lower current on average moving through the tube. I assume this would lengthen tube life, but sounds like it's not necessary.

[Paul Andrews]

You mean PWM it (because they don't latch on)? Yes, this is how you dim them. There is some anecdotal evidence to indicate that this does indeed lengthen tube life. However, if the tube is doped with mercury, the life is in the 100,000s of hours. For undoped tubes it is around 10,000 hours. So let us know what tubes you plan to use. The very common (and very robust) IN-12 are long-life. Some people find the '5' digit on them annoying...

[johnk]

Well, what can we say?

I suspect that he knows Spinal Tap and maybe HE was the interviewer  J

https://www.youtube.com/watch?v=4xgx4k83zzc

 

BTW, Justin, I hope you caught my recent diatribe on Watts and Watts RMS.

 

John Kaesehagen

Australia

 

 

From: neoni...@googlegroups.com [mailto:neoni...@googlegroups.com] On Behalf Of Jonathan Peakall
Sent: Friday, 3 May 2019 23:11
To: neoni...@googlegroups.com
Subject: Re: [neonixie-l] Re: First nixie clock: do I need to switch anode?

 

Because 11 is louder. It is one more than 10! Duh!

On 5/3/2019 6:22 AM, Justin Scott wrote:

Why don’t I just make 10 a little louder and then make 10 the top number and make that a little louder?

 

On Fri, May 3, 2019 at 8:52 AM Paul Andrews <judg...@gmail.com> wrote:

I’m going to say the obvious thing here: Surely it should show the volume as 0-11?

On May 3, 2019, at 8:13 AM, Justin Scott <just...@gmail.com> wrote:

Thanks; good advice. I am an EE with 12 years of experience and have worked with HV before. Built my own tube amp with 420 VDC inside. Would love to hear your tips.

 

What I'm actually building is a volume display for a tube amp. When you turn the volume pot, a couple of nixies will display 0-99. The amp in question already has a transformer that delivers 220VAC, so I will be dropping and rectifying that to get my 180V for the nixie anodes.

 

On Thursday, May 2, 2019 at 5:01:27 PM UTC-4, gregebert wrote:

I dont recommend line-operated designs unless you've done some previous design work at  high-ish voltages. There are a lot of not-obvious things that go wrong (line noise, transients, component failure, inadequate isolation) with very bad consequences.

 

If you have done a lot of past designs, I can share some tips on how to make a line-operated design safer and more reliable. The last thing we want is for someone's nixie clock to cause a fire, or worse.

 

I've had a few "learning experiences" with line-operated circuits that ended-up with sparks, smoke, and/or small explosions despite careful forethought. There's always something you overlook, and sooner or later it will get you.

….clip..

 

[Justin Scott]

Thanks. I'm hoping to use the IN-4. I like the circular bulb and they're a good size for the front panel of a stereo.

[Paul Andrews]

IN-4 tubes aren’t doped with mercury.

--

You received this message because you are subscribed to a topic in the Google Groups "neonixie-l" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/neonixie-l/gIa9ngzJVKk/unsubscribe.
To unsubscribe from this group and all its topics, send an email to neonixie-l+...@googlegroups.com.
To post to this group, send email to neoni...@googlegroups.com.

To view this discussion on the web, visit https://groups.google.com/d/msgid/neonixie-l/2552f2c0-cca1-4269-b857-ce0826bf62e4%40googlegroups.com.

[David Pye]

One downside of the in4 is a supposed lack of mercury inside, and so a shorter lifespan than some of the other tubes eg in12.

How about a few bargraph Nixies for a graphic equaliser too? :-)

--

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 neonixie-l+...@googlegroups.com.

[Justin Scott]

That would be sweet, but I'm going keep it simple for my 1st go-round.

I want to use these tubes:

https://www.ebay.com/itm/IN-4-IN4-4-nixie-digit-tube-display-indicator-vintage-lamp-USSR-NOS-NEW/323117163758

The listing says "lighting voltage" of 170, "operating / discharge sustaining voltage" of 160, and "power supply voltage" of 200. Which of these do you think is the voltage at the anode? I'm trying to figure out what my HV power supply voltage and anode resistor should be. I thought these tubes typically dropped 150V?

To unsubscribe from this group and stop receiving emails from it, send an email to neoni...@googlegroups.com.

[gregebert]

The exact anode voltage is not critical; it needs to be high enough to ensure the tube is reliably ionized over the life of the tube. 200 volts is a good value.

What matters the most is the current. If the current is too high, the tube's life will be reduced. If the current is too low, the brightness might be less than desired, and your tube may develop cathode poisoning, where portions of some cathodes dont fully illuminate. Generally, cathode poisoning is reversible so dont get overly concerned about it at the moment.

The simplest method to limit anode current is to use an anode resistor. You can calculate the approximate value of the anode resistor from basic circuit theory:

Anode_supply_voltage = (anode_current) * (anode_resistor) - Tube voltage.

If your tube's voltage is 165V, and the anode supply is 200V, and the ideal anode current is 2.75mA, you would need about 12.7K for your anode resistor. After you setup a tube and your supply, measure the voltage across the anode resistor to calculate the current to make sure it's the right value.

From the formula, you will notice that variations the tube's voltage drop and the supply-voltage affect the current. But at higher supply voltages, the variations in these will have smaller impact on tube-current. My first nixie clock has a +340V anode supply, so I just use larger resistors and waste a bit more energy.

Personally, I dont like anode resistors because the anode current will vary as the tube and power supply age, so I use a simple current limiter. I've posted about this a few times in the past so you can do a search for details. It's a matter of choice, and there are valid arguments a regulator is overkill.

[Justin Scott]

Thanks for this helpful info! I just need to find out what tube voltage I should expect for these tubes. Sounds like you think 165, so I'll go with that. I can adjust later, as you say.

--
You received this message because you are subscribed to a topic in the Google Groups "neonixie-l" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/neonixie-l/gIa9ngzJVKk/unsubscribe.
To unsubscribe from this group and all its topics, send an email to neonixie-l+...@googlegroups.com.

To post to this group, send email to neoni...@googlegroups.com.

To view this discussion on the web, visit https://groups.google.com/d/msgid/neonixie-l/6359ee9c-1f60-4448-9d62-74b8e9624409%40googlegroups.com.

[gregebert]

The datasheet I found online for the IN-4 has these values, so your resistor value will need to be increased somewhat. Remember - This is the operating voltage, not the ionization voltage (which is always higher).

operating voltage	Voltage drop	V	150
Cathode current nominal	Cathode current nominal	mA	2.5

[Justin Scott]

Yeah, I've seen that as well, but the ebay listing has different values listed (see my msg above). Oh well, I guess I'll have to buy them and test it out.

--
You received this message because you are subscribed to a topic in the Google Groups "neonixie-l" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/neonixie-l/gIa9ngzJVKk/unsubscribe.
To unsubscribe from this group and all its topics, send an email to neonixie-l+...@googlegroups.com.
To post to this group, send email to neoni...@googlegroups.com.

To view this discussion on the web, visit https://groups.google.com/d/msgid/neonixie-l/63acc247-9f63-4249-995d-c466a4f00aa6%40googlegroups.com.

[【ツ】John Smout]

On 2 May 2019, at 18:41, Paul Andrews <pa...@nixies.us> wrote:
>
> I also want to use Osmond to design some simple PCBs on the Mac - I am interested in producing PCBs that are themselves a work of art (as far as I can make them), and I think this would be a great project for that.

I am a beta tester for Osmond PCB and know it quite well. So if you need help or a second pair of eyes, do let me know.

John S