HV power supply for Nixie Tubes

[shadewind]

Hello,

I want to build a nixie tube display with 10 tubes (or something along those lines) and it's going to be powered by 12 volts (or something like that) so I need to boost that to the 170 - 200 V required by nixie tubes.

There are lots of example circuits for this on the web but I really don't want to just copy someone elses work without understandning properly how it's done. So I'd prefer to make my own design.

There seems to be two main approaches here - using a DC-DC boost converter and using a flyback converter.

A boost converter simply needs an inductor so the circuit will probably be rather simple. But I'm having problems finding converters that can output this much voltage, most only go up 28V or along those lines. Is such high voltage output uncommon in boost converters or am I missing something? Any chips you can recommend?

A flyback converter uses a transformer to transform the voltage to a higher voltage if I'm not mistaken. There doesn't seem to be many chips that do this, unfortunately or is this a topology that can be used with most regular switchers? There doesn't seem to be a lot of app notes and similar on this subject. Any pointers you can give me?

[Neilm]

Most flyback application notes assume that you are going from a high voltage to a low one, but there is nothing to stop you from going from 12V to 200V. Be warned however that there are affects of transformer design that cause problems when steping up voltages. I will admit I have only seen these at higher voltages (>5kV).

For referance read http://www.fairchildsemi.com/an/AN/AN-4137.pdf. This can be easily adapted for your application. For more fundimental explainations - a quick look on the web for flyback converters will yield results http://en.wikipedia.org/wiki/Flyback_converter.

One hint I will give, use a simulator like LTspice to design it first and make sure you put a snubber on it. Due to the leakage inductance of the transformer you can get VERY large spikes on the FET when it turns off (after all a lot of current  was just flowing through that inductor). Unsnubbed spikes of 2kV are not uncommon.

Yours

Neil

[deephaven]

I use the MAX773 in my Nixie clock. It's quite a simple design with an external mosfet and a small inductor.

Edit: I just looked it up and it's not recommended for new designs, so you might like to look at the MAX608

[shadewind]

I use the MAX773 in my Nixie clock. It's quite a simple design with an external mosfet and a small inductor.

Edit: I just looked it up and it's not recommended for new designs, so you might like to look at the MAX608

I checked out the data sheet (http://www.farnell.com/datasheets/46701.pdf) and it lists no maximum output voltage. This can hardly be infinite, can it? Anyway, 200V or thereabout should apparently be possible.

[deephaven]

The chip itself doesn't get invloved with the high voltage bit. It drives an external FET so it is that which needs to handle the high voltage. The chip gets a divided down voltage derived from the high voltage output for feedback.

[shadewind]

Right, makes sense. Thanks.

Are there any disadvantages to using a boost converter vs. using a flyback transformer topology?

[Zero999]

Most flyback application notes assume that you are going from a high voltage to a low one, but there is nothing to stop you from going from 12V to 200V. Be warned however that there are affects of transformer design that cause problems when steping up voltages. I will admit I have only seen these at higher voltages (>5kV).

What do you mean?

I thought flyback was the best topology for really high voltages. Old CRT TVs, camera flashes and car ignitions all use a form of flyback transformer to get high voltages.

[shadewind]

A little question about PCB layout:

Should I use a separate ground plane for the HV parts of the circuit? There's going to be logic level stuff on the board as well. Or should I put the HV power supply on a separate board?

[deephaven]

As hero999 said, boost is more appropriate for this application.

You shouldn't need a seperate groundplane as most of the cicuitry is digital. As long as you're careful with any crystal oscillator tracking/wiring you should be ok. Just be sensible wiuth how you route the earthing, i.e. route the high current stuff seperately back to the PSU 0V.

[tecman]

Actually, flyback topography is the most common for high voltage, low current applications, such as your app.  Efficiency is not the highest, but since the power is low, you do not lose much.  In most high voltage flyback converters, a transformer, rather than just an inductor, is used.  A small 120:6 volt transformer usually does a nice job for nixies.

paul
 

[Psi]

Just found this circuit http://www.tayloredge.com/storefront/SmartNixie/PSU/OgiLumen.pdf

It might be of some use.

[Zero999]

Actually, flyback topography is the most common for high voltage, low current applications, such as your app.  Efficiency is not the highest, but since the power is low, you do not lose much.  In most high voltage flyback converters, a transformer, rather than just an inductor, is used.  A small 120:6 volt transformer usually does a nice job for nixies.

Yes with flyback the turns ratio can be lower than what you'd normally require for a given output voltage by a factor of 10 or so, a 6:120V transformer will give just over 1kV out when a 6V pulse is applied to the primary, although this only applies to very low current applications such as a spark gap, if you need a regulated HV supply it's best to go on 4 to 6 times the turns ratio.

When the current is suddenly interrupted in the primary, a high voltage pulse is created across the primary because the transformer's inductance tries to keep the current flowing and the voltage is boosted again by the turns ratio of the transformer.

Another way of looking at it is to put the following facts together:

When current flows through a coil it generates a magnetic field,

When a coil is exposed to a change in magnetic field a voltage is induced in the coil, the faster the rate of change of the magnetic field and the more turns on the coil, the higher the induced voltage.

So when the current is applied to the primary a magnetic field is created. When the current is abruptly interrupted, the magnetic field surrounding the coil suddenly disappears, exposing the coil to a rapid change in magnetic field which generates a theoretically infinite voltage. In practice the voltage will be limited by, the speed at which the current is interrupte, the coil's resistance and parasitic capacitance which will cause a damped oscillation.

[Harvs]

There's a few things in here that aren't quite right.  And if anyone is seriously interested in designing switching supplies you cant go past getting some decent literature on the subject.  The first of many books I got is by the legend A. Pressman.  A great book to start with...

http://www.amazon.com/Switching-Power-Supply-Design-3rd/dp/0071482725/ref=pd_sim_b_1

So first, the magnetic component in a flyback converter is NOT a transformer as such.  It is an inductor, however it has two windings that are magnetically coupled.  This is very important to understand, otherwise you would be designing for a forward converter.  It's not that transformer action is not carried out on the flyback action as Hero said, but the moment you start thinking of it as a transformer instead of designing it as an inductor, you're bound to go off the track.

Flyback topologies can be much easier on the power transistor than boost converters.  In a boost converter the switching transistor has the full voltage stress across it of the output rail, so in effect it has both downsides of the high current from the low voltage side when on, and high potential across it when off.  Comparatively the flyback only subjects the switching transistor to Vin + (Vout/winding ratio), ignoring parasitic inductance.  Therefore you can trade off your secondary turns (and therefore inductor size & cost) as to how much voltage stress the transistor is subjected to.

Anyway, they're not terribly complicated things, but do require a level of knowledge of switching supplies to do a reasonable job of.  Far more information than can be imparted in a forum, or a webpage in my belief.  A. Pressman's book is close to 900 pages and still leaves a lot of information out.  I certainly enjoyed the journey of learning switching supplies and have built several car amps with the knowledge.  Just be prepared to burn a few transistors along the way, as a lot of traps just cant be simulated

[FreeThinker]

Elektor april 2011 have a nice article on nixie tubes, you should be able to download a pdf from it's website (or somewhere (cough!)) .

[Alex]

Elektor april 2011 have a nice article on nixie tubes, you should be able to download a pdf from it's website (or somewhere (cough!)).

Unfortunately the HV power supply circuit published there based around the MC34063 is wrong.

[johnboxall]

Hello,

I want to build a nixie tube display with 10 tubes (or something along those lines) and it's going to be powered by 12 volts (or something like that) so I need to boost that to the 170 - 200 V required by nixie tubes.

At the bottom of this page:
http://www.ogilumen.com/nixie-tube-power-supply-p-91.html
there is a .pdf link with schematic for a MAX1771-based supply. We have one running four IN-18 tubes but it is good for up to 24 tubes.