Like so many others, I'm building a source selector/attenuator; in this case the source selector will be a bunch of DPDT relays run by a microcontroller, shift register, and Darlington array.
What's totally unclear to me is what to do with the signals when they're not connected, and I see lots of methods that I'm not sure I understand.
- Some say disconnect signals and ground completely. However the diyaudio grounding page (here) says "...now that we have cleaned up the grounds, it may not necessary to switch the grounds." Sounds good to me, as I'd rather not buy lots of extra relays, and I see many designs that leave the grounds connected.
- One of those is the Twisted Pear darwin, but in that schematic all the signals are always tied to ground through a resistor, whether connected through the relay or not. The diyaudio grounding page does recommend tying signal to ground through a resistor, but only when unselected: "If there is still a problem you can connect (with a relay) a 50 Ohm resistor between the signal and signal reference of all inputs except that which is selected."
- I've seen other places where they recommend connecting signal to ground with a capacitor on unused inputs.
With the idea that the diyaudio grounding page probably was pretty well researched, I'm inclined to tie grounds together and connect unused signals to ground via a 50-ohm cap, but is there something I'm missing before I go through with it?
What's totally unclear to me is what to do with the signals when they're not connected, and I see lots of methods that I'm not sure I understand.
- Some say disconnect signals and ground completely. However the diyaudio grounding page (here) says "...now that we have cleaned up the grounds, it may not necessary to switch the grounds." Sounds good to me, as I'd rather not buy lots of extra relays, and I see many designs that leave the grounds connected.
- One of those is the Twisted Pear darwin, but in that schematic all the signals are always tied to ground through a resistor, whether connected through the relay or not. The diyaudio grounding page does recommend tying signal to ground through a resistor, but only when unselected: "If there is still a problem you can connect (with a relay) a 50 Ohm resistor between the signal and signal reference of all inputs except that which is selected."
- I've seen other places where they recommend connecting signal to ground with a capacitor on unused inputs.
With the idea that the diyaudio grounding page probably was pretty well researched, I'm inclined to tie grounds together and connect unused signals to ground via a 50-ohm cap, but is there something I'm missing before I go through with it?
both my selectors, the PAS2 tube preamp, and the RA88a op amp disco mixer, tie all the grounds together and leave them tied permanently. The disco mixer had a hum improvement when I changed the ground tie point to the board near the op amp inputs, rather than to the case at the RCA jack panel.
The PAS2 shorts all the inputs not used with a rotary selector switch. this was fine in 1960 with tube equipment, which runs fine powered up into a short. In 2010, with CD and MP3 players having potentially overcurrent sensitive output devices, shorting the unused inputs is not so smart. Isolating the low level (magnetic phono cartridge and dynamic mike) inputs with a grounded bulkhead from the high level inputs (FM, CD, MP3) may eliminate crosstalk if using the turntable with mag phono cartridge, and a high level input is left turned on.
commercial mixers do not have bulkheads, but have the low level dynamic mike inputs located away from the high level tape monitor and effect loop inputs.
The PAS2 shorts all the inputs not used with a rotary selector switch. this was fine in 1960 with tube equipment, which runs fine powered up into a short. In 2010, with CD and MP3 players having potentially overcurrent sensitive output devices, shorting the unused inputs is not so smart. Isolating the low level (magnetic phono cartridge and dynamic mike) inputs with a grounded bulkhead from the high level inputs (FM, CD, MP3) may eliminate crosstalk if using the turntable with mag phono cartridge, and a high level input is left turned on.
commercial mixers do not have bulkheads, but have the low level dynamic mike inputs located away from the high level tape monitor and effect loop inputs.
Last edited:
Does adding muting to the unused input/s help cut down on crosstalk?
If this does help, can we use a muting resistor instead of a dead short, so easing the load on the unused source?
How big can that muting resistor be before the crosstalk approaches what it was before adding the muting.
Am I just chasing my tail?
If this does help, can we use a muting resistor instead of a dead short, so easing the load on the unused source?
How big can that muting resistor be before the crosstalk approaches what it was before adding the muting.
Am I just chasing my tail?
You should here absolutely nothing from an unused input fed with its maximum allowable signal level, even with the volume turned fully up.
Open circuit switching elements (relays, switches etc) still couple the signal via stray capacitance. Its always audible in my experience. The method I use is series switching in combination with shunt to ground for unused inputs. These all feed a virtual earth input stage so there is no "actual measurable voltage" present to couple capacitively. Use a non virtual earth input stage and you can still get traces of breakthrough... yes its many times better than conventional switching... but the virtual earth stage kills it all.
Open circuit switching elements (relays, switches etc) still couple the signal via stray capacitance. Its always audible in my experience. The method I use is series switching in combination with shunt to ground for unused inputs. These all feed a virtual earth input stage so there is no "actual measurable voltage" present to couple capacitively. Use a non virtual earth input stage and you can still get traces of breakthrough... yes its many times better than conventional switching... but the virtual earth stage kills it all.
The reasoning behind this would be avoiding potential potential problems with sources, right? A device floating somewhere could cause nasty signal spikes upon switching if the ground contact lagged behind the signal contact for some reason.
Disconnecting grounds would have the advantage of potentially breaking a bunch of ground loops though, which are quite common in practice. Maybe one could leave things connected by a 1k-10k resistor, just enough to keep sources from floating away into outer space?
Obviously presenting a high-impedance input to each ground connection seems even smarter.
Last edited:
Post #2 here,
http://www.diyaudio.com/forums/soli...fet-amplifier-designed-music.html#post1452488
You can add as many inputs as you want by duplicating the input section (one resistor and two FET's per input)
Post #201 explains the FET operation,
http://www.diyaudio.com/forums/soli...-amplifier-designed-music-11.html#post1569428
The FET's could also be replaced by CMOS switches such as 4016/4066 etc.
http://www.diyaudio.com/forums/soli...fet-amplifier-designed-music.html#post1452488
You can add as many inputs as you want by duplicating the input section (one resistor and two FET's per input)
Post #201 explains the FET operation,
http://www.diyaudio.com/forums/soli...-amplifier-designed-music-11.html#post1569428
The FET's could also be replaced by CMOS switches such as 4016/4066 etc.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.