Pics of my DAC with motorized step attenuator

[Prune]

Front
Remote control receiver; buttons are made from brass tacs; the small faceplate is plexiglass painted black; neons are power lights, whereas the purple LED indicates lock.



Rear
Four inputs, two are the usual RCA, one BNC, and one balanced AES3 with impedance matching and voltage attenuation network; balanced and unbalanced outputs.



Inside
Round metal can houses standby/logic/motor supply; regulated DAC power supply and DAC itself in separate sub-enclosures, all star grounded; MOSFET H-bridge drives motor, which drives attenuator (inside DAC sub-enclosure); IR-sensors on toothed cardboard prevent motor stopping at in-between positions and microswitches prevent motor trying to push attenuator past limits; motor input is from either knob or remote control; knob doesn't rotate except a bit when the user tries to turn it, activating switches; volume indicator is the little arrow behind the knob, attached by a cylinder coaxial to the knob axis and hooked up to the motor by the rope, which is tentioned by the elastic; input/output selection is done by latching relays; all signal cables are 99.9% silver in teflon tubes; the DAC itself is JWB's CS8420/CS43122 dac rev. B design, except that an attenuator is placed directly at the DAC chip's voltage output, shunting between the balanced lines, and the analog section is replaced by a simple self-biasing JFET buffer (the DC offset adjustment trimmers are visible on the sub-enclosure).

[analog_sa]

This is a reaally incredible effort. Using relays would have been so much simpler. Did the step attenuator sound so much better than relays?

[Prune]

This is a reaally incredible effort.

Not really.

Error correction is just that, error CORRECTION. After the error correction the data is 100% correct.

No. Simple information theory. Adding error correction codes is just making the data redundant, so that some amount of damage can be repaired. But it is always possible that damage to both main data and error correcting codes results in the damaged codes matching the damaged data, and the error will be undetectable. Trivial counterexample to your statement: replace the datastream with one from another player, with it's own error correcting codes. Of course, the chance that circuit errors and cosmic rays will produce such a consistent alternate stream is almost nil, but it's not absolute zero. For short blocks of data, it's not all that unlikely.

[skyone]

Hi ,i need service manual for CD5000
Could someone help me .
i want to make some tweaks
-better op -amp
-new clock
-new capacitors (ELNA audiophile grade)
-2 toroid transformers
-copper shielding
-new power cord (some audioquest ,i think).

Thx very much
All the best

[analog_sa]

Simple information theory

I am very much aware of the wrongness of that statement. It looked so wrong and stubborn, i just couldn't help using it

[Prune]

Aha, then you are but a .

[janneman]

Originally posted by Prune
[I][snip]Inside
Round metal can houses standby/logic/motor supply; regulated DAC power supply and DAC itself in separate sub-enclosures, all star grounded; MOSFET H-bridge drives motor, which drives attenuator (inside DAC sub-enclosure); IR-sensors on toothed cardboard prevent motor stopping at in-between positions and microswitches prevent motor trying to push attenuator past limits; motor input is from either knob or remote control; knob doesn't rotate except a bit when the user tries to turn it, activating switches; [snip]

Prune,

Looks impressive, question, is that a stepper motor or did you use a (how shall I call it) continuous rotation type? I have been looking at using stepper motors with the same number of steps as the stepped attenuator, I guess that would also work. Haven't tried it though.

Jan Didden

[Prune]

It is continuous; why else would have I used the IR sensor? Good luck finding cheap stepped motor that provides enough torque (mine was $5 with the gear box).

See the slots in the cardboard? They are aligned so that the sensors are blocked unless in an in-between position. When the user is holding the remote button down or triggering the knob, and lets go, if it is at an in-between position, logic keeps the motor going until the IR sensor is blocked again. This is helped even more by the flexible motor-to-attenuator connector (piece of flexible pipe), as that acts as a damper and allows the click-in-place action of the attenuator to center itself even if the motor is not exactly aligned. Also, microswitches are activated by a groove in the wheel, cutting out the logic signal to the driving H-bridge for a given direction if the attenuator is at the last position in that direction.

[Prune]

I regret not having used a microcontroller. The discrete logic for the whole thing, from standby control, power on muting, remote control, relay control, CS8420-invalid-mode-bug-fixing-reset-circuit, was a pain in the @$$, with a dozen ICs and a bunch of transistors and sh!t.

[jwb]

Hey Prune, I never saw these photos and now they are broken. Can you repost them?