Nixie led driver help

[cataliz3er]

Hello world! Says the newbee,
I've just started work on a nixie tube clock using Z573M tubes and I'm having some doubts regarding the driving. First of all, size is important, so I'm going for full SMD circuits (plus it's hard to find vintage nixie drivers and they're all TTL which makes interfacing the uC a b**ch).
The driver I am planning to use is a SCT2024. It has 16 channels (not going to use them all) and an integrated shift register, so I'll be cascading 4 of these.
The problem I'm having is probably worthy of a "silly question of the year" award, but here goes:

the nixie ignition voltage is 150V (my PSU pushes 157V);
the nixie cathode current should be between 1.5 and 2.5 mA (so let's take 2);
the voltage drop across the IC should be around 5V.



From what I know, once ignited, nixies have very low resistance, so limiting is required. I,ve calculated the limiting resistor at 76k, which gives me a 152 voltage drop at 2mA.

Can anyone say whether or not this would work, or if there is an easier (and space efficient) way of doing things?

Many thanks!

 

[betwixt]

As far as voltage goes you should be OK and the resistor value looks reasonable but you will have to change the way the SCT2024 interface to the driver transistor. The IC alone will not work, it's output stage will almost certainly break down under the high voltage. In normal operation, each output is either 'off' or a constant current sink to ground. It does not produce any output voltage at all so to use the circuit as you show it, you would have to use a pull-up resistor on the outputs, drive a high voltage transistor and invert the data to the IC.

To be honest, as you need to add your own high voltage interface anyway, it would be far more sensible to use a 10-bit parallel bus and a 4 way multiplexer. It reduces the number of HV drivers from 40 to 14.

Brian.

 

[cataliz3er]

Sorry, I forgot to specify that the mosfet is part of the SCT2024.



Here is a schematic from the datasheet. I'm using the configuration on the left, except my "led" is ahead of the resistor and has a negligible voltage drop, so I'm relying on the resistor to drop it down to the specified Vout.



Does this seem reasonable?

As for the multiplexing, I would like to avoid that in order to increase the lifespan (and brightness) of the tubes.

 

[FvM]

Sorry, I forgot to specify that the mosfet is part of the SCT2024.

Yes, but as betwixt explained, it isn't suited to drive nixie displays directly, it has 17 V maximum rating.
I would use high voltage NPN transistors in common base circuit as level translators, base connected e.g. to 5V, emitter to the driver output through series resistor.

For clarity, you should show how the nixie tube terminals are connected in your schematic.

 

[cataliz3er]

Here are some snapshots of the schematic:


That's what I'm trying to find out. Doesn't the fact that I put a 76k resistor before the driver drop the voltage to an acceptable level?
Or do I also need to put a PMOS on the anode for the case when all cathodes are off?

 

[FvM]

As said, you need high voltage transistors as level translators, e.g. BF822. Emitter resistor about 1k for 3.3V base voltage.

An anode switching transistor would be only used in a multiplex scheme.

 

[cataliz3er]

I think I've got it...let's see:
At any one time, my driver will open only one channel. That particular channel, since it has current flow, will only see 5V since the voltage drops on the resistor. In the mean time, because there is no current flow on the other channels the resistors act like pull-up resistors and so the channels see 157V...for the first and last time...ever...smoke.
Did I get this straight?

 

[FvM]

Yes. The open drain output in off-state won't probably see 157 V but very likely more than 17V and can be destroiyed.

 

[betwixt]

FvM is giving good advice.

Think of the output FET problem like this:
When the digit is lit, the tube drops ~100V with low voltage at the cathode.
When it not lit, the voltage across the tube must be less than ignition voltage and no current is flowing so no voltage is dropped across the resistor. That only leaves one place the voltage can be present! In practice, the output of the SCT2024 would almost certainly break down and all digits would light at once.

Note that as only one digit per nixie tube will be one at a time, it makes no difference whether you place the series resistor in the anode or cathode. In the anode you only need one per tube, in the cathode you need one per digit. I'm assuming you are using a common anode nixie of course.

Brian.

 

[cataliz3er]

Note that as only one digit per nixie tube will be one at a time, it makes no difference whether you place the series resistor in the anode or cathode. In the anode you only need one per tube, in the cathode you need one per digit. I'm assuming you are using a common anode nixie of course.

Yes, it's a common anode nixie. But won't the limiting resistor drop the voltage to something below the ignition/maintaining voltage of the nixie if I put it before the anode? I do admit that it would be a whole lot simpler, PCB-design-wise.

 

[cataliz3er]

Dear Cataliz3er,
Not if you calculate it properly. Next time please take the extra 20 seconds to think things through before asking silly questions.

 

[FvM]

To find the minimal required output driver voltage rating, you have to look at the nixie characteristic, particularly the "extinguishing voltage". The driver must at least stand the difference between supply and extinguishing voltage which is about 50V. Assumed you have a driver that has a rating of 50V but not 170V, there's still a risk to damage the driver if e.g. the tube is defective.

 

[betwixt]

Refering to post #10, in the case of a nixie display there should only be one cathode conducting at a time so you can consider the resistor and tube rather like a neon lamp and resistor in series. It makes no difference which comes first, the voltage is the same and the current is the same.

The situation is quite different in displays where the digit shape is made from segment parts and several may be lit at a time. In that case a single resistor would be sharing the current of all the segments and the voltage drop (= display brightness) would vary according to the number of segments illuminated. In that configuration you would have to use one resistor per segment.

Brian.

 

[zuisti]

A long time ago (in 1974 :-?) I've used such Nixie tubes (type Z560M, made in DDR), with a SN74142 from Texas (a decimal counter and high-voltage driver). The anode voltage was around +150v, without any current limiting resistor. It was a 6-digit (cascaded) frequency counter for max. 20 MHz but - using different SPI-like interfaces - you can make a clock circuitry too.

 

[cataliz3er]

Thanks for the advice everybody!
It really helped a lot
Here's the final product

 

[chuckey]

When I was at college (40 years ago), I too made a frequency counter using ZN. . . dekretron(Sp.?) tubes. I used the TTL drivers, a point with these tubes was, that all the cathode needed to be biased to something like +50 V to stop adjacent digits glowing due to leakage within the tube when a digit was on. From memory I used 100K resistors (60 off!).
Frank