• WARNING: Tube/Valve amplifiers use potentially LETHAL HIGH VOLTAGES.
    Building, troubleshooting and testing of these amplifiers should only be
    performed by someone who is thoroughly familiar with
    the safety precautions around high voltages.

Big Tube based voltage regulator

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.

Attachments

  • HSPY-600.jpg
    HSPY-600.jpg
    60.9 KB · Views: 171
  • Big_Current_600V.jpg
    Big_Current_600V.jpg
    30.8 KB · Views: 171
Last edited:
Actually, the concept of a small hi-volt capable shunt comparator controlling the grid of the series-regulating (not “regulation”) pass valve is an elegant, simple idea. Here's the diagram and the relevant math to support it should be fairly obvious.

GoatGuy
 

Attachments

  • simple-regulator-BigTriode-180321a.jpg
    simple-regulator-BigTriode-180321a.jpg
    46.6 KB · Views: 181
Perhaps a little late to the party,

But what is the minimum load? if you need to regulate 1A however the minimum load is guaranteed not to dip under 500mA you can safely pass about 400mA through a resistor parallel to the output tube. The regulator is going to compensate for any output ripple due to the resistor anyway.

For driving multiple tubes, you could look at driving the output tubes with a EL84 or sorts, a 6AU6 is not gonna cut it.

Another approach would be to CCS load a pentode to increase its loop gain well beyond what is possible with resistor loading. This circuit is normally not used in audio because of stability problems (Two CCS fighting eachother)

Attached you will find a schematic of what i built some time ago.
 

Attachments

  • forum-post.png
    forum-post.png
    66.9 KB · Views: 164
Well passing some current through resistors has one major advantage : cost. you need less pass tubes and less current to heat them. Whilst sacrificing little in the way of performance

Multiple grids and stray inductance need to be driven, an 6AU6 with a high plate load is a wimpy driver.

The steve bench supply is really good, however i think that you can manage with two Four 6080(inexpensive) the extra expense for a floating screen supply is not needed at all if you have enough loop gain. CCS loading the driver tube can have merit because it allows the driver tube to operate at high current over its entire range, which increases the loop gain of a single pentode by nearly an order of magnitude

Edit try to find (tektronix circuit concepts -power supply) its a great piece about vacuum tube regulated power supplies.
 
Last edited:
R = V / I
P = V * I or I^2 * R or V^2 / R

(basic stuff, really)

It will of course need to dissipate as much power as the equivalent valve, minus the heater current and that heat still has to go somewhere. It also needs to be rated for the voltage, but that's not likely to be a problem.
 
The pass tube will take the additional current above 400mA in this case. So if the load draws 500mA, then 400mA would go through the resistor, but the tube is regulating by passing only 100mA. Of course, if the total current drawn falls below the 400mA then you lose regulation and the output voltage goes up.
 
While I don't see that happening, the second pass tube eliminates that possibly, and is a valuable piece of window dressing for this build. So why take the chance?

With that logic, there's no reason to “take the chance”.

Interestingly, one can ALSO do it with a plate-series resistor, too; Instead of dropping (say) 100 V thru the tube, drop only 50 V. In theory you've now got 2× the amperage handling THRU the tube. And if it supports such a high current, then none of the over-voltage drawbacks.

Instead, you get an undervoltage condition when current exceeds 2× the nominal setpoint. Semi-active methods come with their drawbacks.

You could (obvious perhaps) even double this up… in the same way, with something MORE regulating. A plate-series MOSFET regulator to pre-filter the raw B+. It too has merit. You can at least be sure that you'll not have a resistor-effected undervoltage drop out condition.

GoatGuy
 
Lastly, as a (PS:)), note that the ripple reduction of the reg tube diminishes by the fraction of current it is handling. So, if the valve carries only 20% of the current, then it can only regulate at 20% of its 100% nominal efficiency. Which then begs the question, “what's the point?”
 
When the output current exceeds the current flowing through the resistor, the regulator is going to compensate for any ripple caused by the resistor, the amount of attenuation depends on the frequency dependant internal resistance of the supply and the value of the resistor used to bypass. Any ripple on the output will be amplified many times by the error amplifier, and fed 180 degrees out of phase into the grid of the cathode follower tubes. Which theoretically should cancel out the majority of the ripple. Think of it this way: only a portion of the current going through the resistor to the output is AC, if the tube can vary the current by the same amount it will be able to compensate.
 
The tube regulation is most limited by its current handling ability here.

Instead of using the resistor bypass, use a Mosfet current source bypass, for something below minimum current. Then no ripple current is passed around the V regulator tube at all. The V regulator tube still controls the output voltage 100%.

One could go even further and make the CCS bypass into a tracking one that supplies a fixed fraction of the load current, even for a varying load. This will require some careful Fdbk design, with two interacting Fdbk loops.
 
Last edited:
OK, now that we've veered all the way to Bangalore in order to get from Paris to Nice, one could observe that following this line of thinking, and very soon the regulating tube is chucked out in favor of MOSFET to begin with.

But that isn't the OP's design criteria. He's going for Overkill Steampunk, and needs the Big Bottle as part of his props. The resistor-in-parallel idea is pretty nice, in that it'll easily handle losing one insufficient-handling tube. And there really isn't too much in downside. Especially if the pass-resistor is fine tuned to be about 50% of the nominal load.

Although it may be “cheating with nefarious intent”, I love adding bells and whistles that actually don't do anything, when going all steampunk. Seriously. You save yourself a world of hurt and excruciating wallet-ache by just lashing things in, powering up their heaters, and otherwise not taxing them. VR tubes are GREAT at this. You almost can't have too many. They're so colorful.

I rather think the same about using mercury vapor rectifiers, except that when they have an envelope failure, there is that darn cloud of toxic gas to deal with. But there are MILSPEC units. And the bluegreen-to-white is outstanding. In any case, I'd never, ever condone cheating to such a degree that the actual amplification is bypassed whole in favor of a shadow sand amplifier. No. Tubes can sound SO nice.

The trick instead is balance between one's functionally pointless steampunk fetishes and one's perfectly rational desire for valve amplification. The (truly!) super-glue is a bit of solid state off the audio pathway. As needed. For outstanding improvements at the mundane level, like solid power supply output.

Just saying.
We're now in Bangalore.
Let's get back to Nice.

GoatGuy
 
Thanks GoatGuy! As for VR tubes, there are 7, granted they all do something.

As for mercury vapor rectifiers, well it's hard to light those up without actually using them. I'm also using 4 JAN milspec units. Above all though, there is an extraction blower that will pass all cooling air through sulfur pellets and activated carbon to sequester any mercury should the unlikely event of envelope failure occur.
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.