Voltage stabiliser/reference valves.

[Diabolical Artificer]

I have a few of these in my stash and thought it about time I had a play with them. I have two VR105/30's that I'll use at first. As far as I can see from an STC Application Report you set the valve with an anode resistor to set Ik so that 1)the valve doesn't exceed the max current and 2) the Ra passes sufficiant current for the valve to work. That and the valve has to have a breakdown voltage. What I can't establish is if the valve has a Va absolute max. The VR150/30 datasheet says Va supply 185v, Va breakdown 180v, Va stabilising 162v with Ik @ 40mA. It then goes on to give Va breakdown 166v, Va stabilising 150v measured with a KI of 15mA. So from this I assume, the valve will have a different Va stab for different Ik.

I have a board that has a mains tfmr with a sec of 330v 0v 330v, a choke with two 8u caps configured as a pi filter, an octal base for the rectifier valve( I'll use a 5U4G) and two bases for the VR150/30's. I've worked out that I'll get 460v DC unloaded and that a 22k Ra will set the two valve's to a working I of 20mA approx. But do I need to drop the HT prior to the two VR's?

These two valve's in series will sit across the rail (in parallel) does this mean they won't restrict the output current? I realise that usually the VR's are used to fix another valve's operating condition's, a small pentode say, that in turn control's a series pass valve like an 807. But for the moment I'd like to knock up a si,ple test circuit, to get a feel for how they work, how stable they are etc.

TFL, Andy.

[Boater Sam]

I always thought that the reference was used to control a series pass valve as you intimate rather than sink the excess voltage.
How you achieve current control using the voltage reference I know not.
Many, many years ago I built a 350v stabilised power supply using a 75v reference controlling an 807 pair, it worked well but has gone in the intervening decades.

[turretslug]

Very broadly, these devices first appeared for shunt stabiliser use and were used as in post 1 (think of it in the way that a high-power Zener would be used), with a typical running current of low tens of mA, then more precise devices appeared intended for use as a reference in electronic regulators using "hard" valves, their running currents are typically a few mA, i.e. like a low-power Zener. Having said that, it seems that Blumlein published the classic reference/comparator/pass architecture as far back as the early '30s, but the topology really took off in the post war period with B7g references such as the 85A2 and specialist high current, low impedance pass valves, though many cooking PA valves got used (and definitely cooking....)

The VR105/30 comes into the earlier, shunt-type category, the /30 indicates 30mA maximum running current, I think the minimum running current is characterised as 5mA for stated performance and best stabilising performance is between these figures. The "anode load" resistor is indeed the currentlimiting resistor here, the output being taken from the anode/resistor junction.

Beware of using them in parallel, the tolerance spread is likely to mean that one ends up doing most of the work. Also, they are funny devices altogether, with operating current regions that can have negative resistance characteristics- who knows what electronic mud-wrestling a pair might get into! Series operation is OK (with a few provisos) but parallel use sounds like "here be dragons" territory.

[Diabolical Artificer]

Thanks both. I have a spare 85A2 as well, and came across one as used in a Solartron PSU. It like it's brother's does as you say act as a voltage reference.

I'll have a play and report back.

Andy.

[G8HQP Dave]

Using two VRs in series can require some extra help to get them started if the incoming voltage is not enough to get them both to strike; you need to do tricks with a diode and a couple of resistors so they are effectively in parallel before striking. However, 460V should be enough to get two lots of 180V going so nothing to worry about.

Be aware of the issue of parallel capacitance. Too much and you have an oscillator instead of a regulator.

[Ed_Dinning]

Hi Andy, simplistically you can think of them as a zener.
The series resistor is sized for the load current and a few mA for the tube; as the load current increases the tube current falls and vice versa. This keeps the tube and load voltage sensibly constant .

Ed

[Bazz4CQJ]

I've used them in simple series circuits with valve RF oscillators (e.g. an HRO LO) and had very good results. Not sure of the pros & cons wrt zeners in such applications.

B

[kalee20]

Originally Posted by Ed_Dinning

Hi Andy, simplistically you can think of them as a zener.
The series resistor is sized for the load current and a few mA for the tube; as the load current increases the tube current falls and vice versa.

Yes - apart from one thing!

A 150V Zener passes no current until you apply around 150V across it, then it turns on and passes as much current as you let it have, clamping the voltage at 150 (in practice, the voltage rises slightly as current increases).

Whereas, a 150V stabiliser tube passes no current as you apply 150V, 160V, 170V, typically 180V and then suddenly springs into conduction, snapping back to 150V and clamping at that. And if you reduce the current below 5mA, you might find it 'goes out' and the voltage rises to above 150V, depending on the circuit.

That's why they can oscillate with a capacitor in parallel. A Zener will never do that.

[joebog1]

As I remember, the maximum capacity across the regulator, is very low, and in many circuits there is no need for capacitors at all, perhaps only when used as stabilizers for very sensitive radio front ends to restrict any noise that may be generated.
I have also used things like NE2 neons to establish grid voltages for "not so important" voltage regulators. The disadvantage is like the garden variety gaseous regulators, and that is that they strike at slightly different voltages at every switch on. This is usually only a volt or two, and in many cases, unimportant, BUT if you have lots of gain in the pre regulator/driver it can be significant in output voltage from the main regulator valve.
Joe

[Diabolical Artificer]

Finally got round to trying these out as have been busy fixing my motorbike. On a quick test with no load they sat at around 304v DC. With a change of 100v IE 350v - 450v input, the OP voltage differed by 3v.

I'll try them with a load today and plan to try them for regulating the "screen" voltage on a LTP cascode. Their physical size would rule them out for use on an amplifier, but they look handy for use as part of a valved regulated PSU.

Andy.

[Skywave]

Originally Posted by Dr Wobble

Their physical size would rule them out for use on an amplifier, but they look handy for use as part of a valved regulated PSU.

For which purpose I've successfully used them in many a home-brew design. But a few points need mentioning.

First, with a conventional post-rectifier C-L-C pi-filter, there is the risk that if semiconductor diodes are used and the PSU is switched on with no load, the voltage across the regulator valve can rise to an excessive level - which can be harmful for that valve. For such PSU designs, I usually favour the choke input arrangement, for the usual & well-known reasons.

Second, some regulator valves have an internal jumper fitted between two base pins. That enables the voltage to not appear if the valve is unplugged from its socket: it's a safety / over-voltage protection technique.

Third: I have had two cases of excessive R.F. noise coming from the mixer in a couple of SW radios. The cause in each case was a noisy VR150/30.

Al.

[Bazz4CQJ]

I'm just wondering if there is a "bottom line" here, especially in considering pros and cons of gas regulator valves and zenners?

For a new homebrew project, is the case totally clear for zenners? Are there any downsides there relative to the regulators?

In terms of existing equipment, where gas regulators were fitted as standard, you'd probably be inclined to retain their use, but in the event that for some reason you wanted to change over to a zenner, there is no (technical) reason deterring that?

B

[G6Tanuki]

Like pretty much all 'breakdown' devices, Zeners can be powerful sources of white-noise up to surprisingly-high frequencies; you need some good filtering to keep this away from sensitive places - particularly important if you're using a Zener in a bias-network.

Also, Zeners can present a significantly variable parallel-capacitance as the current they pass varies and their temperature varies (think of them as being like a varicap diode...) - I've had experience of this where adding a Zener stabiliser to a HF local-oscillator actually worsened both the thermal and voltage-stability until I included some solid RF decoupling between the Zener and the VFO circuit itself. Seems the Zener was appearing as a [variable] capacitor in parallel with the oscillator's supply.

The reverse is (largely) true with glow-tube regulators - inappropriately excessive decoupling capacitance paralleled across the glowtube can turn it into a 'neon relaxation oscillator'!

[Bazz4CQJ]

Originally Posted by G6Tanuki

Also, Zeners can present a significantly variable parallel-capacitance as the current they pass varies and their temperature varies (think of them as being like a varicap diode...) - I've had experience of this where adding a Zener stabiliser to a HF local-oscillator actually worsened both the thermal and voltage-stability until I included some solid RF decoupling between the Zener and the VFO circuit itself. Seems the Zener was appearing as a [variable] capacitor in parallel with the oscillator's supply.

That makes sense. Maybe the Bottom Line is that both approaches need to be used with care, or the problems could outweigh the benefits.

B

[trobbins]

I added a shunt regulator to a Selmer organ 300V rail, regulated by series OA2's. The OA2's can be returned at any time and the ss regulator link removed. The 300V rail is used for the note oscillators, so the better regulator should avoid re-tuning over the decades to come. The oscillators are a series feed Hartley that allows vibrato to be applied across all notes by modulating the B+.

http://dalmura.com.au/projects/Selme...Schematic2.gif

[Wendymott]

I used 150V zeners to replace the VR150's in both my Eddystone HF receivers, but these are modified with a M5 threaded case to heatsink to chassis. It may be a moot point about thermal effects causing drift, but with circuits over 50 years old, surely anything that can reduce the overall heat buildup, is better than nothing at all. Incidentally both eddies are much more stable and get less hot as a result of removing these heat generators.

[joebog1]

With respect!! The zeners also look as cold as they run.
In other words VERY dull and boring.


Regards
Joe

[Wendymott]

No argument there Joe....when the case is on... its just another radio, hopefully working better than its original.