[z2trillion]

My plans for the attenuator:
256 position encoder to control volume up/down.
20 bit counter that goes up and down based on pulses from encoder and direction of rotation.*
20 bit counter feeds parallel eprom which has 720 binary values in a roughly exponential sequence.
Values from eprom turn relays on/off in a 20 bit r-2r ladder adc.

*I'll hook it up so that the counter can't go from 1111...1111 to 000...000 and vice versa.

Does this look like it will work? One obvious issue could be sending the signal through 20 relays; I'll probably hook all 40 up in series and see if I can hear a difference. Other than that, are than any other issues I am overlooking? Thanks.

 

[morsel]

I can't imagine wanting more than 64 steps. Too fine a resolution becomes a nuisance when large changes in volume are desired. If 256 position encoders exist, they are likely to cost a fortune.

 

[Publius]

Also, wouldn't 20 relays switching at the same time be a bit loud?

 

[ble0t]

I would agree...that many steps is too many steps...

As for the encoder, it's not that bad...only $50 or $60. You cant get encoders with resolution up into the thousands if you really want

 

[z2trillion]

The real reason for 720 positions is that it gives nice happy .1 db steps from -0 db to -719.9 db. In any case, you need at least 16 relays to get attenuation down to -72 db, and large number of steps is relatviely "cheap" in terms of relays you need to add. Theoretically, with 20 relays I could get away with using .01 db step sizes from 0 db to -50 db.

I'm planning on using: these. When you hook them up so that open and close really fast, all i can hear is a faint buzzing.

Like ble0t says, 256 position Grayhill encoders are "relatively" cheap at $50 from digikey.

My main concern was if the r-2r ladder will work as a volume control, or if it will do horrible things to the signal.

 

[ble0t]

I think you mean to -71.9dB

That is a very, very picky amount of volume control...you must have sensitive ears!

I guess I'm still a tad confused...are you trying to avoid using a PIC? I would just use the encoder hooked up to a controller that controls some relay driver ICs. That would be a much cheaper solution...

 

[Edwood]

My Grace m902 has 200 steps, and it's kind of annoying. 100 would've been perfect.

But then again, with the Shure E5's, you can hear music at step 1.

-Ed

 

[z2trillion]

Just a question, on your Grace m902 does it make a little click on each of the 200 steps (detents I think they're called)? With a 256 position encoder, for I would need like 2.5 turns for the range range.

 

[Edwood]

Yes, the rotary encoder used in the m902 has small "stops".

They are very slight, so it's quite easy to spin the knob quickly for rapid volume changes.

-Ed

 

[z2trillion]

bleot: yeah, I meant -71.9 db. I'm trying to use the relays because I have 40 of them I bought on a whim a while ago, and I believe they offer better performance than active switches. (I doubt I could hear the difference though.)

Just one more question though, are these relays suitable for the signal?

 

[ble0t]

They seem OK...don't see anything that would prevent you from using them...

 

[jmedin]

You could use the REF input of a CMOS Multiplying DAC as your Signal IN, and use the DAC out as the "wiper".

Then use your encoder/memory algorithim to control the "log" type funtion of the linear 12 bit MDAC....

Check the app notes at Analog Devices and National for more info.

just a thought.....

 

[Crowbar]

Quote:

Originally Posted by z2trillion /img/forum/go_quote.gif Values from eprom turn relays on/off in a 20 bit r-2r ladder adc.

This is pointless, when you can get 1024 volume positions from 10 relays and exponentially weighted resistors. Also, use latching relays; they are activated by momentary pulses and thus you won't get noise coupled in from the coils.