Power Scale pot sweep

[K O'Connor]

Hi Guys

There is a compromise in the Power Scale pot sweep, depending on the specific Power Scale circuit, the supply voltage, and the amplifier.

With the older SB-style approach using the mil-spec pot, the original kit offering incorporated a linear pot. It is easy to see on a scope or meter the direct relationship between the pot setting and the output voltage. Due to the logarithmic nature of our hearing, a linear pot is not a good choices, so, we changed to a log pot. Linear is good to use the circuit for a bench supply. The loudness change follows a squared curve, and a log-squared curve, respectively, and the latter ended up providing better low-loudness resolution of control.

With the newer Power Scale kits, there is a compromise that causes a dead spot in the sweep at one end or the other, or at both ends. The dead spot at the top end means no voltage change occurs over that part of the sweep. This is not too critical for most players once they get used to the fact that you will not hear a loudness change until you get to the power level you are actually using - which is lower than they think. But... to the player this means the useful range of the pot is smaller.

A dead spot at the low end means the amp has gone completely quiet or off before the end of the pot sweep. How big this dead spot is is a bit more concerning from an ergonomic view. In the present SVn-D kits, R1,2,3 form a voltage divider with the Power Scale pot in parallel with R2. The divider values are chosen to give a 100x power reduction with the pot set to 12-o'clock, and to go to zero volts at full sweep. Most amps will be silent at some positive voltage out of SV, so getting to a true zero is not necessary, more aesthetic. With R2=47k5 and R3=1k the output can go to zero, but changing to R2=43k2 or 44k2 and R3=2k21 gives  better sweep overall.

Have fun

[K O'Connor]

Hi Guys

The original form of the SV1-D is what I called SD-1, never released as such but I used it in my amps.

In that form, R2=56k2 and R3=6k81. This definitely does not go to zero volts out, but the amp sound would cut off right at the end of the pot sweep (values tuned by ear). When I decided to change the SV kits to this format for general release, the PCBs were labelled with these values, but later shipped with the newer values listed in post-1. Recent boards are labelled for the new values.

Note that a compromise with the top end dead spot is that if you try to eliminate it you may end up dropping voltage across the mosfets at the nominal full power setting of the Power Scale pot. Losing 10V or so in a 4-500V environment is no big deal BUT the mosfets will be running a lot warmer than they should at full power. Mosfet dissipation at full output should be essentially zero.

[RockmanCentralBob]

With Kevin's assistance, I was able to install the Power Scaling kit (RBX and SV1) into my Marshall JVM HJS.
Now I no longer need a Power Soak to blast it at bedroom levels or feed my Two Notes Torpedo Live without overheating the Torpedo.
We converted my Master Volume (1 and 2) controls into Power Scale controls.
He even helped me fix a bad PPIMV design with a better one with an added relay so I can have separate Drive Compensators for Master Volume 1 and Master Volume 2.
Switching is dead silent and instantaneous!
I have one set as a "clean" Power Scale and the other set to overdrive the Output tubes since 100 Watt Marshalls are designed to stay clean.

As for the Power Scaling pot sweep, for my particular configuration, I can start to hear the amp at about 9 o'clock (or a volume of 2).
A volume of 3 is very comfortable for bedroom playing and wouldn't disturb anyone else in the house.
But the TONE is very much THERE!!

Thank you, Kevin, for all your help!!

   

[K O'Connor]

Hi Guys

The comment that "I could hear the amp at about 9-o'clock" suggests that there is a dead spot in the sweep at the quiet end (CCW).

That got me thinking about the actual resistance values of the pots I am using. Measuring a handful of them they are quite varied and the 12-0'clock half-way position is not actually one-tenth of the total end-to-end resistance; rather, it was usually a bit higher. These are "average" pots which tend to have 20% tolerance. Like every other "log" taper pot, it is not actually a logarithmic-valued track but a series of linear sections, which is easier to manufacture. With enough sections there is adequate response for most applications.being log,

I grabbed some other pots that are typically thought to be much better (conductive plastic instead of carbon), but their values were varied and within 20% of their rating.

I grabbed a stepped resistor style pot used in hifi equipment. This is not a true Alps, more likely a clone. In any case, the switch has forty or so positions with a resistor soldered between them. It comes down to the designer doing the math correctly to attain a proper incremental resistance sweep and net value desired. Here, the overall value was off by about 10% but the half-way sweep is right where it should be at one-tenth of the total. These stand a chance of being more consistent from unit-to-unit, but that depends on the tolerance for the resistors used.

The Alps pots are 2-section as they are intended for use in a stereo. This is actually okay for a Power Scale installation where we can effect a single-knob solution for any amp that has a Schmitt splitter AND pre-PI-MV, and for any amp using a concertina.

Over the years I have been adjusting values within the Power Scale kits to try to have the broadest range of control, good low-level resolution and optimal mosfet heat dissipation. With normal pots made by Humans it is nearly impossible to achieve all three at the same time. One solution, is to add a trimpot that would allow the installer to optimise the sweep for the specific amp sample combined with the kit component tolerances. However, I have found that trimpots are best avoided on kits whenever possible,