Alternate 32 step attenuator design

[neutron7]

So i ordered some of the group buy 'clone cards and decided i need a nice volume control.

I am looking at the prices of a nice switch to use for an attenuator, and they are kind of pricey. especially if you use the really nice resistors.

anyhow I came up with this as an alternative.

5 opto transistors pick up light or no light from an encoder attached to the volume control shaft instead of a potentiometer or switch. the transistors drive reed relays which alter the level of the signal passed from the input to the amplifier.

the signal only passes through switch(relay) and some resistors. more than a switch and 2 resistors you would get in a proper stepped attenuator, but still quite minimal.

some reed relays are rated for 1 billion operations which should make it reliable.

reed relays are sealed so there will be no problem of dirty contacts.

the control circuit can be placed a long way from the audio circuitry. it would not require much current. with 10mA relays all on for 2 channels would use 100mA + whatever 741s use. (dont have to use 741s, anything could work, maybe just a transistor.)

here is a rough circuit. I have not tested it, i will have to get some parts.



since you only need 5 relays per channel they can be good quality ones.

same with the resistor network. only 10 good quality resistors to buy. per channel.

the resistor network is based on the same idea as an R2R digital to analog converter except the signal level is changed instead of a fixed voltage.

I was worried about how to make it "log taper" and thinking how i would change the resistor values etc but duh, i just put that on the encoder.(the one on the pic is fake log taper i just did it for this image)



the encoder will be printed on transparent plastic printer sheets, and then attached to a plexiglass disk that will have a shaft attached or attached to an old potentiometer with the guts taken out.

[neutron7]

I have been doing some test with simulation software and thinking about this and decided it would be better to use 6 bits instead of 5 because the steps at the bottom of the range are kind of large. also the r2r thing is a bit bogus because the input and output resistance go all over the place.

it might be better to use some deviders and "mix" the outputs with the switches. i will see when testing! (that way might be cheaper because I could use spst relays)

there are some non optical ways to do this i have seen here which use a ADC to control the relays with a pot, but none of the circuits i could find mention how they set up the relays/resistance network.

im a bit of a n00b at this as you can probably guess

[Nuuk]

I've no idea how well that would work but 10/10 for some original thinking