Idea for a better volume control...

[Lisandro_P]

I've been doing some thinking about the "shunt resistor" volume control that's seen a lot arround here. A lot of designs use resistors connected to ground (via a relay or transistors), but only use ONE at the time. This works, but offers 10 to 16 steps at most...

Now, how came no one ever tried to use multiple resistors to ground? As they parallel, the volume shunt would be smaller... this would allow a huge number of steps (128 for 7 resistors), and could be driven more or less directly digitally (perhaps NOT gates would be needed). Now, i haven't done the maths on this one, but what do you think?

 

[kiwi_abroad]

Hi Lisandro,

Someone has tried it, and probably more than one. That is how Nelson Pass does it in his Aleph P pre-amps. Have a look at the service manuals on his web site for the schematics. Unfortunely, you pretty much have to use a uP to control it. As it stands, you would have a linear volume control, not a logerithimic one, which would make control at one end of the volume controls range very touchy. It would be better to have more steps than you need, and use the uP to select which ones are needed.

Cheers, Adrian

 

[Lisandro_P]

Someone has tried it, and probably more than one. That is how Nelson Pass does it in his Aleph P pre-amps. Have a look at the service manuals on his web site for the schematics. Unfortunely, you pretty much have to use a uP to control it. As it stands, you would have a linear volume control, not a logerithimic one, which would make control at one end of the volume controls range very touchy. It would be better to have more steps than you need, and use the uP to select which ones are needed.

Then i'm smart! Wooohoooo!

I'll see if this approach works... actually, i belevie that if the shunt resistors are logarithmic (each one doubling the previous one, like in the Pass preamps - thanks for the tip!) the combination should be log as well. I'll have to draw some curves on Excel or something to see if it happens that way.

But, if it turns out to be any good, with the transistor driven version, a couple quality resistors and two or three basic digital IC would make an excellent volume control. My idea is to drive it using a rotary encoder (a regular small DC motor interfaced), i hate noise in pots. I'll keep you briefed!

 

[Lisandro_P]

Been playing with this idea, and i think you can make it work with a simple binary counter.

Just for kicks, i made a spreadsheet to see how such a circuit would work. 10k series resitance, and 8 shunt resistors (starting with 100k and halving it each step down; 100k, 50k, 25k, 12k, 6k, 3k, 1.5k, and 0.7k). Set the counter so the MSB corresponds to the 0.7k resistor, and voila!
It works well; the step size kinda fluctuates a bit (between 0.03 and 0.07dB), but stays fairly logaritmic - and with 256 steps it should be more than enough for everyone. Besides, it eliminates the need for a uC and it could be driven from something as simple as pushbuttons.

I plan to use transistor-driven shunt resistors as shown here. Next step is to build it in protoboard to see how well it performs.

 

[AKSA]

Lisandro,

I don't know if this is useful, but some 18 months ago I experimented with signal transistors as shunt elements.

I found that if the collector was connected to the signal output, and used as a mute, that when switched off you could clearly hear distortion.

Now your application places the transistor collector and emitter at the ground point in saturation, with take-off from the midpoint of two resistors, a la ladder configuration. This will likely reduce distortion to a mere fraction of the ratio of the divider resistors, which will reduce distortive effects by a factor of ten or twenty, so it may be OK.

However, a word of caution; the distortion was not nice........

Cheers,

Hugh

 

[sajti]

Lisandro,

I don't know if this is useful, but some 18 months ago I experimented with signal transistors as shunt elements.

I found that if the collector was connected to the signal output, and used as a mute, that when switched off you could clearly hear distortion.

Now your application places the transistor collector and emitter at the ground point in saturation, with take-off from the midpoint of two resistors, a la ladder configuration. This will likely reduce distortion to a mere fraction of the ratio of the divider resistors, which will reduce distortive effects by a factor of ten or twenty, so it may be OK.

However, a word of caution; the distortion was not nice........

Cheers,

Hugh

Hugh,

did You ever try to use jFET as a switch? Maybe it can gives lower distortion.
Another possibility: LED+Photoresistor. As I know the photoresistor has no distortion, and if change the current across the LED we can use it to change the gain. I made some limiter for professional audio application with this solution....

sajti

 

[sajti]

I've been doing some thinking about the "shunt resistor" volume control that's seen a lot arround here. A lot of designs use resistors connected to ground (via a relay or transistors), but only use ONE at the time. This works, but offers 10 to 16 steps at most...

Now, how came no one ever tried to use multiple resistors to ground? As they parallel, the volume shunt would be smaller... this would allow a huge number of steps (128 for 7 resistors), and could be driven more or less directly digitally (perhaps NOT gates would be needed). Now, i haven't done the maths on this one, but what do you think?

Relay would be be much better, I guess. But You need to use lot of them

sajti

 

[jcx]

Lisandro,

I don't know if this is useful, but some 18 months ago I experimented with signal transistors as shunt elements.

I found that if the collector was connected to the signal output, and used as a mute, that when switched off you could clearly hear distortion.

...

However, a word of caution; the distortion was not nice........

Cheers,

Hugh

Did you really turn the mute Q off? All the time?

With the base-emitter junction reverse biased (making sure the signal dosen’t exceed the bias on peaks), and low Z at the base you have only a few pF of reverse biased nonlinear diode capacitance – unless you had such large series R that the diode C is actually affecting the signal there should be little distortion

For absolute minimum mute distortion you can ac bootstrap the base drive to the signal - no ac signal across the bc junction, no ac current thrught the nonlinear cap or modulation of the leakage current to flow in the series R


Bjt don’t make the best switches but a few are designed to be used “upside down” relying on the reverse Beta of the transistor which is typically < 10, this can give saturation V if only a few mV – but since the reverse biased emitter is serving as “collector” you only have 5-6 breakdown V

 

[dmfraser]

Volume Control

If you don't like pot noise, either buy a better pot or look into solid state volume controls. I use the Maxim-Dallas DS1802. Besides having CPU control, you also have the ability to have Up/Down push button control.

TI has a volume control IC that runs from a ±15V supply, the PGA2310. Find it at www.ti.com

 

[sam9]

IC volume control sounds attractive, especially, 4 or more channels need to be controlled simultaneously. However, I worry about distortion and noise comparred to a got pot. There was a post not long ago that refered to some company's whitepaper that discussed this with respect to remote controls. They had some new scheme to overcome noise and distortion limits, but what struck me most was a pot (or motorized) pot was the baseline to match. They rejected the usual ICs, and FET switches on sonic grounds. It seemed to say the main objection to pots was that they were old fashioned. I don't recal if there was a discussion of relays, but I think that was dismissed (reasonablY when you consider number needed to control multiple channels.

I would be interested to know if there is anything on the web expesssly comparring sonic qualities of good pots vs. IC volume cobtrollers.

 

[Lisandro_P]

My main issue was how to control the shunt elements in parallel without a uC - i was kinda surprised when i found out i could get away simply with a counter, as i expected to have to thinker a bit more. I still have to figure how to drive a display for it with discrete digital part though

Now, what's left is the shunting element itself. Relays are cool, but at 8 relays per channel it simply becomes too expensive.
Hugh, i'm interested in your experiments. Do you recall how you wired the BJTs?

 

[Kermit]

As far as controller goes I recently made a relay volume control using two bcd counters to drive the relays and two 7 segment led displays.
Using 2xbcd and not true 8-bit limits the number of steps for any give number of relays, but how many steps do you really need? I’m using 6 relays for 40 steps and that’s enough for me (but I’m using a different arrangement then what you had in mind). Anyways if you want the schematics or pcb layout let me know and Il send them to you. I’m using 74hc192 bcd counters witch is pin compatible with 74hc193 4-bit counter so the same pcb is usable for both 2xbcd and 8-bit.

 

[Lisandro_P]

The 8-bit design is simply because 8-bit counters are cheap and easy to come by, and if i'm using transistors for swicthing the price impact is negligible. If i were using relays i might even set for 6-bits (64 steps).

I would like to see your works' schematic if you don't mind.

 

[Kermit]

As requested, here is the schematic for the controller. I’m having some problems with opening acrobat at the time, so I hope the pdf is viewable as I am unable to check it myself.
(Btw you can get 1, 2 or even 4 pole relays so that’s 8 or 6 relays total for two channels, then its realy not that expensive.)

 

[Lisandro_P]

Nice work. Yeah, 6 or 8 DPDT relays would work. Would contact fatigue be a consideration with them?

 

[Kermit]

Eventually I guess there will be some. I’m thinking about leaving the power for the controller on all the time so that it remembers the last settings, so that I don’t have to “click” my way thru steps 0-15 to find my preferred listening level. I think that would make the relays last longer. Less clicking.
Btw if you haven’t already figured out what 4025 in the schematic is there for, it prevents the counter from jumping from 0 – 256 (99 for bcd) if you press down when the counter is in its 00 state. Remember to include that if you design your own.

 

[Lisandro_P]

Actually i was thinking of that myself today. I think there were some counters that had an overflow/underflow flag pin, otherwise yes, you'll need extra logic to deal with that situation.

BTW, in the design i plan, the power is controlled by the digital part of the circuit, which stays always on; it's then trivial to unhook the display and controls from the volume circuitery (just check for the "power on" signal )

 

[Kermit]

It’s possible that other counter has such a function. I chose the 192/193 series because it has separate clocks for up and down counting. Meaning it’s easy to interface with just pushbuttons or a rotary switch witch is what I’m using at the time.

 

[AKSA]

Lisandro,

I used a BC550 npn, with collector to hot, and emitter to ground. Base was pinned with 1K to ground, and I seem to recall driving the base via a 1K resistor from a 1.5V. I noticed when off the collector had adverse effect on the sound quality.

Send me your email address, Lisandro, and I will send you a circuit using relays for the purpose you describe from Bjorn Kolbrek, a clever Norwegian.

Cheers,

Hugh

 

[EUVL]

Circuit using relay

> Send me your email address, Lisandro, and I will send you a circuit using relays for the purpose you describe from Bjorn Kolbrek, a clever Norwegian.


Dear Hugh,

I am sure there are other people who would be interested as well (me for example). Would you be kind enough to post it ?

If not my email address:

[email protected]


Best regards,
Patrick