Dummies guide to building a 120watt 140 volt power supply to drive IN-9 Nixie tubes.

[spec]

Here is another approach for the CC tube PSU which takes care of all the tolereancing problems by using stabilized supply rails. It embodies the principle of the Les buck strike supply (LBSS) to conserve power. It is also reasonably low power and low cost and requires no circuit building. With the specified PSUs, 416 tubes could be supported.

​

DATA SHEETS & SOURCES
(1) PSU 1 to 3
**broken link removed**
**broken link removed**
(2) PSU 4 (mains powered)
**broken link removed**

NOTES
(1) For this design it is fundamental that PSU 1 to 3 are isolated types.
(2) PSU 1 to 3 are 48V types that can be adjusted by +- 10 %. The approach above uses the PSUs adjusted to their maximum of 52.8V
(3) It may be possible to locate smaller units for PSU 1 to 3
(4) PSU 1 to 3 will operate from 100V and 230V with no change of the tube supply rails while still maintaining essentially the same efficiency.
(4) PSU 4 is a simple low cost, non isolated, 52.8V to 5V board-level buck switch mode PSU for the control electronics, about £2 on Ébay. On reflection, it would probably be better to use a separate mains powered PSU to generate the 5V for the control electronics.
(5) DA1 to DA100 should be low leakage types. Their forward drop is not important, within reason. They do not need to withstand high surge currents as in the case for rectifier diodes.
(6) The worst case power dissipation of the transistors is only 250mW.
(7) The worst case power dissipation of the resistors is also 250mW

ERRATA
(1) For every 5 tubes fitted, place 1 high frequency decoupling capacitor between the 106V supply line and 0V.
(2) For every 10 tubes fitted, place 1 high frequency decoupling capacitor between the 158V supply line and 0V.

 

[fireant007]

oops

 

[fireant007]

Here are two more suggestions to get the right voltage. 1 Connect a 12 volt transformer (Such as this one.) in series with the 55 volts (With windings in parallel.) so that you would have 67 volts AC.
2 Sit the high voltage supply on top of the raw low voltage supply to add that voltage to the high voltage supply.

Les.

2 ... not quite sure what you are getting at but I think you mean as spec has done above with the 3 PSU approach ?

 

[fireant007]

Hi spec,
Fire would probably need just a 5 volt low voltage supply so a 9 volts transformer would be about right. So if he used a 9 - 0 - 9 transformer with a bridge fed from the ends and the negative output of the bridge to ground he could feed the 5 volt regulator from the centre tap (About 12 volts peak) The positive of the bridge would give about 24 volts so if this was added ( HV supply - to this + 24 v) to the 77 volts (55 x 1.414) he would have an HT supply of about 100 volts. This would mean that if the 115 volt mains input was down to 100 volts he would still have 87 volts. I don't think you need to allow as much as 20 volts across the transistor and its emitter resistor. He would not be driving the base with any more than 5 volts so there would only be about 4.4 volts across the emitter resistor so the transistor should still work with about 8 volts on its collector. I think he should try it with the existing power supply and see how low he can turn down the voltage from his variac before it stops working.

Les.

I can't quite visualise this but it sounds good.
Once the tubes are lit you can turn the voltage down quiet a lot (if that is what you are saying).
As far as striking goes, well .. they don't spark up.

What would you like me to do with the vairac ... exactly ? (now now ... don't be rude )

 

[fireant007]

Here is another approach for the CC tube PSU which takes care of all the tolereancing problems by using stabilized supply rails. It embodies the principle of the Les buck strike supply (LBSS) to conserve power. It is also reasonably low power and low cost and requires no circuit building. With the specified PSUs, 416 tubes could be supported.

View attachment 98865
​

DATA SHEETS & SOURCES
(1) PSU 1 to 3
**broken link removed**
**broken link removed**
(2) PSU 4 (mains powered)
**broken link removed**

NOTES
(1) For this design it is fundamental that PSU 1 to 3 are isolated types.
(2) PSU 1 to 3 are 48V types that can be adjusted by +- 10 %. The approach above uses the PSUs adjusted to their maximum of 52.8V
(3) It may be possible to locate smaller units for PSU 1 to 3
(4) PSU 1 to 3 will operate from 100V and 230V with no change of the tube supply rails while still maintaining essentially the same efficiency.
(4) PSU 4 is a simple low cost, non isolated, 52.8V to 5V board-level buck switch mode PSU for the control electronics, about £2 on Ébay. On reflection, it would probably be better to use a separate mains powered PSU to generate the 5V for the control electronics.
(5) DA1 to DA100 should be low leakage types. Their forward drop is not important, within reason. They do not need to withstand high surge currents as in the case for rectifier diodes.
(6) The worst case power dissipation of the transistors is only 250mW.
(7) The worst case power dissipation of the resistors is also 250mW

ERRATA
(1) For every 5 tubes fitted, place 1 high frequency decoupling capacitor between the 106V supply line and 0V.
(2) For every 10 tubes fitted, place 1 high frequency decoupling capacitor between the 158V supply line and 0V.

Mein gott ein himmel,
Now there is a power supply.

(like the line from that movie clip. One man crazy 3 (or in this case) 2 very sane spectators)

Kind of bulky though.
416 tubes !
I am not doing the Brighton pier lights this year !

But I do like the idea !

Loving the heat efficiency.

I am going to try the doubler approach first before going too mad with hardware.
So... off to the shop to get diodes and caps.

 

[spec]

2 ... not quite sure what you are getting at but I think you mean as spec has done above with the 3 PSU approach ?

Hy fire,

Les is overcoming the the shortfall in voltage from the 55V +55V transformer, particularly when you have only 100V mains feeding the 115V primary wingdings.

His reasoning is:
(1) You will need another power source for the microcontroller and opamps, either transformer based or PSU module
(2) That being the case, go for a transformer approach and use a transformer that can supply the raw voltage for a 5V stabilized line but also has a winding of a high enough voltage to augment the main transformer voltage.
(3) There are two approaches.
(3.1) Rectify and smooth the secondary of the second transformer and add that voltage to the rectified and smoothed voltage from the main transformer.
(3.2) Simply join, in series, a spare secondary winding on the second transformer to the coils on the main transformer.

Hope I have interpreted Les' cunning plan correctly.

spec

 

[fireant007]

Yes .. I understand the low voltage striking problem.
Wonder how many brown outs go down to 100Vac anyway ?

He is right about another power source and a small transformer makes sense.

But,

I "could" just put a couple of secondary windings on the existing transformer.
But I am loath to cut up the existing windings !
Wonder if that would work ?

I have no idea if the primary is the outer or inner winding.
My guess is they wind the secondary over the primary.

REALLY don't want to take a knife to it.

Well.
Went shopping today and have just got back.
My local electronic's shop is kind of rubbish.
Didn't have what I wanted.

So.
Ordering time again from farnell

 

[spec]

I "could" just put a couple of secondary windings on the existing transformer.
But I am loath to cut up the existing windings !
Wonder if that would work ?

I have no idea if the primary is the outer or inner winding.
My guess is they wind the secondary over the primary.

REALLY don't want to take a knife to it.

I thought something was bothering you about putting extra windings on the main toroidial. transformer.

You don not have to touch the existing wires or construction of the transformer. Simply wind the additional secondaries over the existing coils... dead simple.

spec

 

[fireant007]

No .. not just on the outside .. I thought you wanted me to extend the existing secondary.

 

[spec]

No .. not just on the outside .. I thought you wanted me to extend the existing secondary.

The additional coil/coils are completely independent and not connected to any of the existiong coils on the transformer.

spec

 

[fireant007]

Oh.. well then.. off I go.
I am sure I have a few old transformers with sufficient enamel wire in them to rape and pillage from.
Even have kapton tape so am good to go .....

You might have said !

 

[spec]

Oh.. well then.. off I go.
I am sure I have a few old transformers with sufficient enamel wire in them to rape and pillage from.
Even have kapton tape so am good to go .....

You might have said !

You have to remove the mounting plate in the center of the transformer and and wrap the new windings over the top of the existing windings looping through the hole in the center of the doughnut shape.

Kapton tape- excellent (the yellow stuff)

I did say.. several times.

spec

 

[fireant007]

You mentioned something about raping a virgin transformer .... least that's what I read
Finding suitable windings as we speak. Bound to be plenty up in the shed.

 

[spec]

We had a communication breakdown somewhere.

Good luck with the new secondary- like I said it should be quite easy.

Best wrap some kapton tape over the existing windings first.

Once you have done the four turns test, as previously described, to establish the transformer turns per volt, I can then calculate the best number of turns for the new secondary. You can use any wire that will fit for the test. Ordinary flex will do.

spec

 

[fireant007]

I looked back at your post.
You did say "over the top" ... I just assumed I would be cutting the outer layer of insulation of to do it.
But ... magnets are amazing .. it just goes right through that plastic. Crazy I know.

I did an 8 turn test just to be difficult.
Looking at the trannie .. you can see which way they wound the secondaries so I just went in the same direction.
Don't want any crazy magnetics.
So.
Using - from the micrometer - 1.2mm diameter enameled wire I just, ehem.. borrowed from a transformer that was sitting around being lazy and not producing current like a good transformer should.
Got 1.638 V from 8 turns = 0.20475 V/Turn.

So....

(1) put 4 test windings on transformer.

Put 8 on for better resolution.

(2) measure the RMS voltage from the windings.

1.638Vac RMS

(3) calculate RMS Voltage/4 to give turns per volt (TPV).

1.638/8 = 0.20475 TPV

(4) Then number of turns = Required Voltage/ TPV.

30V/0.2475 = 146.52 Turns....

Oh my lordey loardey that's a lot of turning.
Not sure I can count that high.

1. Probably don't need such large wire .. even the wire in the toroid is not THAT big.
2. Good craft project for the misses ? or Kids ? ...

Ok.. why are we doing this again ?
Oh yer .. brown outs at 100Vac.

Hummm....

I think heat is the biggest problem I have here.

So... striker circuit and only sayyyyy ... 100Vdc Max at the anode.

I think Les wanted to know what the open circuit RMS from the secondaries was at a low voltage.
Will go and measure that now.

Oh yes.. my multimeter won't measure that cap ... too big methinks.
Measures small ones ok though.

 

[fireant007]

Toroid Secondary Voltages.

Are.

110 = 56.8
100 = 51.7
90 = 46

As for the how low can you go question before they won't light again.

Pretty hard to answer for all tubes ... but this one won't relight if Vdc goes below 130Vdc

Above that it lights. (mostly).

As I say though.. all tubes have their own personality's so I am sticking to the data sheet.

 

[dr pepper]

I thought that was the herb ladies.

Have a look here too:

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

 

[fireant007]

Not sure what you are pointing too exactly Doc.

I have only ever seen one flyback circuit that had enough oompf to drive a lot of amps. (for vacuum tubes). none of these little tackers will do it.
In the 5 to 6 watt range at best.

This is the fattest one I have ever seen on the net and it uses a custom flyback trannie which I have no idea where to find or make etc etc.

http://www.tube-tester.com/sites/nixie/nixie-clock-cd47/cd47.htm

Edit: from his page about the transformer "which I pulled out of an old PC-monitor and equipped it so, that it works with the switching regulator:"

Errr. ?

 

[spec]

But ... magnets are amazing .. it just goes right through that plastic. Crazy I know.

That is quite true, only magnetic materials affect magnetic fields

Looking at the trannie .. you can see which way they wound the secondaries so I just went in the same direction.
Don't want any crazy magnetics.

The direction of winding is not important.

Using - from the micrometer - 1.2mm diameter enameled wire I just, ehem.. borrowed from a transformer that was sitting around being lazy and not producing current like a good transformer should.

Thick wire is good electrically but hard/large to wind 1.2 mm is around 17AWG. 20AWG (0.81mm) would be OK electrically and would be easier to wind and would not take up so much space.

(1) put 4 test windings on transformer.

Put 8 on for better resolution.

(2) measure the RMS voltage from the windings.

1.638Vac RMS

(3) calculate RMS Voltage/4 to give turns per volt (TPV) [VPT].

1.638/8 = 0.20475 TPV

My mistake should have been Volts per Turn (VPT) but no worries- you got it sorted

That low VPT corresponds to 4.88 turns per volt (TPV), which surprises me for a 120VA transformer. I would have thought more like 2 TPV.

Can you confirm that when you made the measurement you had 230V RMS connected to the primary of the transformer.

(4) Then number of turns = Required Voltage/ TPV. [VPT]

30V/0.2475 = 146.52 Turns

Hmm, that is more than I was expecting, but not impossible with 20AWG wire.

I think heat is the biggest problem I have here.

I have done a load of calculations and you are right, heat dissipation is the big problem, although not insurmountable. One of the main problems is the requirement to operate from 100v RMS mains. Any chance that can that be upped to 115V +-5%?

Oh yes.. my multimeter won't measure that cap ... too big methinks.
Measures small ones ok though.

Don't worry about measuring the the reservoir capacitor as it is new. I thought it might have been an old one out of your spares box.

spec

 

[fireant007]

The direction of winding is not important.

spec

Humm... I guess you are correct in the sense that we are drawing from it not inputting into it .. but I would have thought the current draw on it would effect fields. (no expert).

Can you confirm that when you made the measurement you had 230V RMS connected to the primary of the transformer.

No .. that was with 110Vac on it.

As for operating voltage.
I have no idea what the nominal 110 voltage is 115 or 110 ?

We use 240Vac here .. not 230

I am just assuming 110 is a standard ... google here I come.

edit: quick google later .. it seems that 120v is taking over.
It also says that 230 is taking over instead of 240
Not here it hasn't !