vr tube / glow tube as filter?

[mrneedle]

Hi all

I have ordered some VR tubes / glow tubes / regulator tubes from ebay. first because they look so darn nice

then I started to read about how they work and what they are supposed to do

so i've decided to actually use them instead of just building the nightlight

while i was reading about various possibilities of usage vr tubes in a power supply I came across the Decware website.

and read this: the actual text

REGULATION

In both instances, where the grids of the output tubes and the input stage are regulated, they are not regulated in the typical way. Instead the regulator tube is wired in series with the load. The tube creates a voltage drop and acts as a filter that seems to block much of the nasty harmonics and noise that seeps in from your AC power. Rather than resistors or chokes which provide a direct conductor for the high voltage, the regulator tube creates a condition where all the high voltage must jump between the cathode and plate through a vacume, so hey folks, look no wires! It simply amazes me what doesn't make the jump. In fact even ripple rejection is many times higher that it would be with the equivalent voltage dropping resistors. Off all the tube amplifiers available on the market today, very few use tube voltage regulation so it's realistic to think this amp will be somewhere in the top of pack sonically. NOTE: If you wire the regulator tubes the normal way as a shunt to ground, they actually inject noise that must be later removed.

"Instead the regulator tube is wired in series with the load. The tube creates a voltage drop and acts as a filter that seems to block much of the nasty harmonics and noise that seeps in from your AC power."

it seems the oa3/0a3/vr75 is connected to the 6v6 and the OC2 seems to be connected to the 12at7/12au7...

i don't fully understand how that would work, and i would love to see a schematic of this... i mean the vr75 isn't able to supply enough voltage to the 6v6, is it? and is it really possible to use them as a "filter"? they aren't called regulator tubes for nothing, are they?

can some of more experienced tubeheads shed some light on this?
(no pun intended)

thanx

 

[sreten]

Hi,

S.D. makes up his inaccurate marketing blurb as he goes along,
and his site is full of half truths and sheer fanciful drivel generally.
A tube cannot act as a filter as described.

Well it is his site and he can say what he likes, caveat emptor.

rgds, sreten.

My favourite is the detailed design and principles exposition
of a phase plug for a Fostex 208 which turned out to be a
a standard 1/2" socket that fitted the drivers polepiece. No
design and no principles, just experimentation and pure guff.

 

[Rundmaus]

+1

 

[DF96]

Regulators tend to be a bit noisy, however you wire them. He seems to suggest that they are noisy when one end is grounded, but not noisy when the other end is attached to the supply rail. This can't be true.

Whether a regulator can 'filter' incoming noise would depend on how quickly the current can change with voltage. If the response is sluggish because the ion avalanche takes time to develop then some filtering action might be possible. I doubt if it would reduce ripple, but it is conceivable that it might remove more incoming noise than it adds of its own. His 'explanation' involves a vacuum, which is nonsense as these do not have a vacuum.

 

[kstagger]

Based on an old Sound Practices schematic I've used a zener after the screen tap of a ultralinear transformer - just to drop the voltage. I know Tubelab has used VR tubes to do the same but (this is from memory) had problems keeping them lit.

 

[Tubelab_com]

I know Tubelab has used VR tubes to do the same but (this is from memory) had problems keeping them lit.

The "filter" stuff that is posted above is as suggested pure BS. Wiring a VR tube in series with a load for the purpose of dropping a fixed voltage IS a perfectly good and valid use from an electrical standpoint, PROVIDED the tubes ratings are obeyed. A zener diode can also be used in this manner.

Filtering "nasties" with one isn't.

A noisy device connected in parallel with the load will inject voltage noise into the load. If its dynamic impedance is low enough it may shunt some of the power supply noise supplied by the "direct conductor for the high voltage" to ground.

A noisy device connected in series with the load will add its noise current to the load, and if its dynamic impedance is lower than the resistor that would ordinarilly be used, offer less impedance to power supply noise, allowing MORE noise to enter the circuit.

Indeed I did experiment with a VR tube in series with the screen grid for a particular purpose, and I had problems with the tube dropping out. Sweep tubes have a low screen grid voltage rating. This rating can be violated on some tubes, and on some tubes it can not. This leaves out triode or UL mode for a large selection of sweep tubes. I pondered the use of a VR tune to lower the screen grid voltage. It seems like a good idea, until you consider the large signal condition where the plate voltage drops to near zero. This voltage is less than the VR tube's voltage and the tube goes out removing screen grid voltage from the output tube.

6CD6 testing

 

[DF96]

Can I explain why I think there is a possibility (to put it no stronger than that) that a VR might act as a weak filter?

Everyone knows that putting a cap in parallel with a VR can lead to relaxation oscillation, but this only happens for larger caps i.e. slower potential oscillation. This implies that the VR cannot respond quickly enough to maintain oscillation with a small cap. If its current response to a change in voltage is sluggish, then it could act as a crude filter for higher frequencies - almost like a choke. It would be easy enough for someone to measure this.

Note that a VR does not usually exhibit negative resistance except when striking. Under normal operation it has a positive slope, but with hysteresis between the striking and extinguishing voltages.

 

[DF96]

Just looked up the 0A2 data sheet. This specifies an average change of 2V for 25mA current change, with possibly up to 6V. Slope resistance is therefore 80-240 ohms. Let's assume 150 ohms, and hope we are within a factor of 2. The sheet also says don't use a parallel cap greater than 0.1uF. 0.1uF and 150ohms gives a frequency of around 10kHz. The 0.1uF probably errs on the side of caution so it is likely that the VR reacts a bit more slowly than that. So we find that a VR might have a rising slope resistance from a few kHz and up.

Personally I would not want to use a filter component which adds significant noise of its own.