The simplistic Salas low voltage shunt regulator

[Salas]

Ikoflexer your neighbor had seen several occasions of oscillations in his many experiments when looking for more bandwidth etc. used Cdom caps on the driver bjts many times. I agree.

The base is conservative V1, well tested in the field with wide V-I settings gamut and uses, always robust till now in many DIY style constructions. All the Simplistic Riaa guys use it, its in a jwb hitec preamp, its on T-Amp, OPA627 preamp, GB DC B1, you guess.

The perfected versions are V1.xx Ikoflexer does a very good job on expanding the performance envelope of the basic idea.

[AndrewT]

the maximum current that a trimmer track can pass is determined by the maximum power the whole track can dissipate.

Originally posted by anatech
That's only at maximum resistance! You must de-rate the power by the fraction of the travel in degrees.

that's exactly what my statement says.
The maximum current of the whole track must never be exceeded no matter where the wiper is located or how the pot is wired.

To your second point.

This is a critical error in the use of these parts..................... You can not use a trimmer as a rheostat!

In a Vbe multiplier, the wiper is shorted to one end. Now it's a variable resistor (rheostat).
The resistor string across the Vbe multiplier can pass anywhere from 100uA to 10mA. All this resistor string current (except for the bit tapped off to feed the NPN base) passes through the wiper.
I have never heard anyone say a multi-turn pot is not suitable here. In fact many builders specifically recommend it.

[Ciu]

To shunt-regs Builders

What kind of pre-regulator is before shunt reg ?

_ Just rectifier and caps ?
_ gyrator/cap-multiplier ?
_ Lm317 or similar ?

Thanks
R.C.

[ikoflexer]

Originally posted by Ciu
To shunt-regs Builders

What kind of pre-regulator is before shunt reg ?

_ Just rectifier and caps ?
_ gyrator/cap-multiplier ?
_ Lm317 or similar ?

Thanks
R.C.

The one thing I would not use before the shunt reg is an lm317 or similar. Personally, I've just used simple CRC or CLC filtering. But some other guys did use more pre-regulation stuff, maybe they'll chip in with comments.

[anatech]

Hi Andrew,

that's exactly what my statement says.

I was just covering the case where the trimmer is used as a rheostat and only a part of the element was in circuit. I felt it should be made more clear for members who are starting out.

Of course, we could always state the blazingly obvious. Always read the product use and ratings information from the manufacturer! This tends to quash all arguments.

The resistor string across the Vbe multiplier can pass anywhere from 100uA to 10mA. All this resistor string current (except for the bit tapped off to feed the NPN base) passes through the wiper.

Absolutely correct!

And that is why a multi-turn control should never be used in these circuits. For one, there is no requirement for them unless you do not restrict the adjustment range. In cases like that, the designer is simply being lazy. The moving contact in a multi-turn control is much smaller than a single 270 ° type of control. That is an important difference!

Let's look at circuits that were engineered properly, which happens to be those designed before the 80s (for example). Bias circuits always used a heavier control, single turn. Multi-turn controls were found only in circuits as a potential divider type where no current to speak of flows through the wiper circuit. These were never designed to pass current through the wiper. Just because a few people started misusing them does not mean everyone else should.

Now, if you look at the very old bias circuits where the control is used as a rheostat, you will see that they are normally wire-wound devices. That's until the cheaper units were manufactured using carbon controls.

Lastly, there is an undeserved reputation of being low quality, or "cheap", attached to single turn controls. The idea is that multi-turn controls are a somewhat higher quality. Well folks, nothing could be further from the truth! Each part is designed for a different job, and choosing an expensive multi-turn does not mean it will do the job at hand. It's up to the circuit designer to understand the parts they use - for real. Going off audio myths will not steer you towards a reliable design. Just like watching a Lemming leap over a cliff doesn't mean the rest should. But just like some humans (Sheeple), they do mostly. Yes, I do know that the Lemming thing is not accurate, but it serves to illustrate a point.

I did once have a number of 10 turn controls that were designed to carry low current. They were panel mount only as they were wire-wound. One final remark on this topic. I worked in a calibration lab where we dealt with process control and test and measurement equipment. The biggest trouble came from those multi-turn controls. The single turn types lasted a longer time and normally had their range restricted to an area that did cover calibration needs. Yes, properly designed. Don't be lazy people!

-Chris

[Salas]

Originally posted by Ciu
To shunt-regs Builders

What kind of pre-regulator is before shunt reg ?

_ Just rectifier and caps ?
_ gyrator/cap-multiplier ?
_ Lm317 or similar ?

Thanks
R.C.

I made the shunt so to avoid series regulation chips and clearly beat their sound. So I tried the sound of a few parts Mosfet shunt. The idea of a series regulator chip before the shunt... What is the purpose of making it if to wrap it in a LM317? Make shunt to avoid regulating in series and then feed it with a series chip... A contradiction. Make a better V2, don't make it in to series IC chip again.

P.S. A CRC or CLC is what we mostly used, only to a small benefit.

[jameshillj]

Yes Chris,
It's quite surprising how fixed is the use, and misuse, of those multiturn pots.

Salas,
I also have played around a bit with different diodes, ripple caps and Cmultipliers before the shunt, with quite mixed results.
The search was on to have smaller high quality ripple caps and top diodes, etc in the raw supply with a very "live" and dynamic sound and use the Cmultiplier to reduce ripple and to also bring up the weight, and low bass definition. This was intended for a specific 250mA current load (my headamp).
In the usual supply setups, any change in the raw supply is transferred to the load, but the Salas reg appears to absorb much of the variations of the raw supply - don't know if this applies everywhere, with different loads, etc, etc. A bit of a surprise, this.
So, from the point of view about reducing dependency on the raw supply, this is good news, especially if it also reduces some of that line noise that can be so troublesome - this will be part of the testing procedure, when we eventually get there ....

Okay - Now, the lower shunt current idea .....

On the high voltage thread, there has been some discussion about using far less shunt current than the load current, particularly for class A amp loads. This is quite different to our normal reg method. A few of our more adventurous power burners have built some Salas "heavy duty" shunt regs and have had excelent results, but with high shunt currents. This was also talked about for the F3 awhile back, but shelved as too impractical.
However, it's perhaps worth another look ....

For example, the F3 amp (First Watt F3) consumes 1.7A at 42volts (single rail). With our usual ratio of having 120% times the load current going thru the Shunt arm (= 2A), this requires a Total current of 3.7A (and a 5volt higher rail) = 3.7a x 47v = 175W per channel - a rather high power consumption!
But .... If the Shunt current need only be 20%(Iload), this would reduce the current thru the Shunt arm to about 0.3Amps and a Total Regulator current to a more reasonable 2A (1.7 + 0 .3) at the higher 47 volts (= about 100W/ch).
The question is, does it give all, or some, of the benefits of the shunt reg?

We'll see.
[I would imagine this would apply to any other ClassA amp.]

.... Any comments?

[Salas]

In my T-Amp shunt, I just specified about 25% on top of its peak demand (two channels). We can't practically have the luxury of double the current in a heavy situation like with power amps. We retain most of the benefits for +20% peak setting, plus at Ampere and over settings, the gm of the Mosfets is much better and they compensate subjective loss to our general practice. All in all a sonic blast from just a capacitor filter anyway. If you have headroom so the minimum 5V Vin-Vout drop won't make you lose much power, try it. Of course if someone is thinking of a shunt before ordering the mains trafo, just allow for some extra Volt, no problem. Can even burn it with still beneficial RC filtering if will not use a shunt, is always better to have some spare.

[stormsonic]

@anatech
nobody is saying that single turn is inferior.
I recommended multiturn for only one single reason: with single turn trimmer you won't be able to precisely set desired shunt current.
Agree with you, after desired current is set, measure resistance & replace with fixed resistor to avoid possible problems.

One more question: already built regulator?

[anatech]

Hi stormsonic,
I was addressing a continuing issue regarding multi-turn vs single turn controls.

I recommended multiturn for only one single reason: with single turn trimmer you won't be able to precisely set desired shunt current.

Here is the idea I was looking at. Your statement is not true, unless there are no other resistors in circuit. That would be a very poor design, and fairly risky on top of that.

Generally speaking, a circuit that uses a trimmer is designed so that the movable contact sits about mid-range. You do this by using two fixed resistors, plus a single turn control in that leg. If you want a 10% adjustment range, the trimmer will make up about 10% of the total resistance. You can go ahead and replace the trimmer with a few fixed resistors if you want, but there is no need to do that.

Do you follow now? That is why I commented that people were just being lazy these days. If you do this correctly, the range will be restricted to a range that will not instantly destroy an amplifier by cranking the control up and down. You build in safety.

-Chris