Monitoring Controller "Class A" - Help us & Join us to design it !

[moamps]

If the user adjusts the volume of output 2, that value is remembered, so if the user switches to output 1 and then back to 2, the latest value of output 2 is used.

This will not work correctly, IMO.

And, how do you switch the inputs?

 

[Monte McGuire]

I think it will work correctly. You just sum all of the volume control outputs and either use volume commands or mutes sent to the volume control chips to implement 'switching'. There won't be so many channels as to cause problems from the fact that all of the chips are essentially active at the same time. Plus, if there's some silly reason why one would need to mix sources together, this would be easy to do by altering the software, but impossible if the sources were actually switched.

Further, it's usually a good idea to do some sort of a volume ramp up or down before doing a mute or a switch anyway, just to avoid abrupt changes or clicks.  I think the chips have a gain setting as well as a mute, and while it's good to ramp down to a low gain to mute, the mute itself should also probably be asserted, just to avoid crosstalk issues. IIRC, most of these chips have separate volume and mute features, so this should be simple enough.

 

[moamps]

I think it will work correctly. You just sum all of the volume control outputs and either use volume commands or mutes sent to the volume control chips to implement 'switching'. There won't be so many channels as to cause problems from the fact that all of the chips are essentially active at the same time. Plus, if there's some silly reason why one would need to mix sources together, this would be easy to do by altering the software, but impossible if the sources were actually switched.

Further, it's usually a good idea to do some sort of a volume ramp up or down before doing a mute or a switch anyway, just to avoid abrupt changes or clicks.  I think the chips have a gain setting as well as a mute, and while it's good to ramp down to a low gain to mute, the mute itself should also probably be asserted, just to avoid crosstalk issues. IIRC, most of these chips have separate volume and mute features, so this should be simple enough.

I have no problem with that. It's over-designed and IMO has practical purpose only if multi channel audio is monitored.

I have problem with this:

If the user adjusts the volume of output 2, that value is remembered, so if the user switches to output 1 and then back to 2, the latest value of output 2 is used.

Imagine situation where you are listening a quiet  signal with main and nearfield speakers, so the volume of both PGAs is for example at -6dB. Then you are returning to the main speakers and next ten minutes you are working only on them on some loud signal where your  PGA is at -18dB. Then you switch to your nearfield speakers to check the sound there, and they (and you) go overloaded because the second PGA is left on -6dB.

To conclude, the listening volume should be IMO  propagated thru all PGAs. Eventually there can be a small fixed offset if someone like to equalize SPL at listening position with monitor controller instead on output amplifiers.

 

[drask]

We choose to use relay to select input and output.

 

[Andy Peters]

I think it will work correctly. You just sum all of the volume control outputs and either use volume commands or mutes sent to the volume control chips to implement 'switching'. There won't be so many channels as to cause problems from the fact that all of the chips are essentially active at the same time. Plus, if there's some silly reason why one would need to mix sources together, this would be easy to do by altering the software, but impossible if the sources were actually switched.

Further, it's usually a good idea to do some sort of a volume ramp up or down before doing a mute or a switch anyway, just to avoid abrupt changes or clicks.  I think the chips have a gain setting as well as a mute, and while it's good to ramp down to a low gain to mute, the mute itself should also probably be asserted, just to avoid crosstalk issues. IIRC, most of these chips have separate volume and mute features, so this should be simple enough.

I have no problem with that. It's over-designed and IMO has practical purpose only if multi channel audio is monitored.

Well, it's not monitoring multi-channel audio. It is designed to have four sets of stereo inputs and several outputs. Each output has a mux for input select. Each output has a volume control, and the volume control is also the mute. What's "over-designed" about it?

I have problem with this:

If the user adjusts the volume of output 2, that value is remembered, so if the user switches to output 1 and then back to 2, the latest value of output 2 is used.

Imagine situation where you are listening a quiet  signal with main and nearfield speakers, so the volume of both PGAs is for example at -6dB. Then you are returning to the main speakers and next ten minutes you are working only on them on some loud signal where your  PGA is at -18dB. Then you switch to your nearfield speakers to check the sound there, and they (and you) go overloaded because the second PGA is left on -6dB.

Well, honestly, that's a usage issue.

To conclude, the listening volume should be IMO  propagated thru all PGAs. Eventually there can be a small fixed offset if someone like to equalize SPL at listening position with monitor controller instead on output amplifiers.

That's a small matter of firmware.  There is nothing preventing the gain setting being propagated to all of the outputs.

 

[Andy Peters]

I think it will work correctly. You just sum all of the volume control outputs and either use volume commands or mutes sent to the volume control chips to implement 'switching'. There won't be so many channels as to cause problems from the fact that all of the chips are essentially active at the same time. Plus, if there's some silly reason why one would need to mix sources together, this would be easy to do by altering the software, but impossible if the sources were actually switched.

No, that's not what's going on here. There is no channel summing, and the volume control chips do not implement the switching. That's too complicated.

The volume controls are per output, not per input. There is one 4:1 mux per output, which selects which input will drive the output. Here is a block diagram:

Further, it's usually a good idea to do some sort of a volume ramp up or down before doing a mute or a switch anyway, just to avoid abrupt changes or clicks.  I think the chips have a gain setting as well as a mute, and while it's good to ramp down to a low gain to mute, the mute itself should also probably be asserted, just to avoid crosstalk issues. IIRC, most of these chips have separate volume and mute features, so this should be simple enough.

Before an input select is changed, I (well, the firmware will) mute the associated output and then unmute it after the change.

 

[Andy Peters]

We choose to use relay to select input and output.

So you must have a bank of relays to do this selection?

 

[moamps]

Well, it's not monitoring multi-channel audio. It is designed to have four sets of stereo inputs and several outputs. Each output has a mux for input select. Each output has a volume control, and the volume control is also the mute. What's "over-designed" about it?

You have  three  PGAs and two MUXes redundant for the same functionality.  See the attached picture.

Well, honestly, that's a usage issue.

Yeah, you will see what your customers will comment.

 

[Andy Peters]

To conclude, the listening volume should be IMO  propagated thru all PGAs. Eventually there can be a small fixed offset if someone like to equalize SPL at listening position with monitor controller instead on output amplifiers.

That's a small matter of firmware.  There is nothing preventing the gain setting being propagated to all of the outputs.

Since the remote control part of the design is still in process, I wrote a little program for the Mac which controls the hardware. It was easy enough to add a feature to enable level tracking, so when one output level is adjusted, all outputs which are set to track are adjusted as well.

Aside: it is interesting when hardware design conforms to a software-design paradigm called "MVC," for "Model-View-Controller."  The "Model," in this case, is the mux/gain-control hardware. The "view" is either the Mac-based control program or the upcoming hardware control. And the "controller" is the (software) logic which relates the two.

 

[Andy Peters]

Well, it's not monitoring multi-channel audio. It is designed to have four sets of stereo inputs and several outputs. Each output has a mux for input select. Each output has a volume control, and the volume control is also the mute. What's "over-designed" about it?

You have  three  PGAs and two MUXes redundant for the same functionality.  See the attached picture.

Well, honestly, that's a usage issue.

Yeah, you will see what your customers will comment.

The picture you posted shows a configuration which makes it impossible for more than one output to be active at a time. Which is perfectly fine, if that meets your requirements!

I was discussing this with a friend (a mastering engineer) and he asked, "Does it have a dedicated subwoofer output?" and I said no. And then after a moment, I said, "Well, there's no reason why you can't connect the sub to one of the outputs and the mains to another, and have both outputs on at the same time." (An external crossover is required for this.)

Then it's again a simple matter for the controller software/firmware to have one control which affects the level of both. And even better is the balance between the mains and the sub can be set and then the level control maintains that balance as you change the volume.

Hey, this design was basically scratching an itch. You are free to design your controller in any way you want!

 

[moamps]

The picture you posted shows a configuration which makes it impossible for more than one output to be active at a time.

You have already said that the controller isn't intended for the multi- channel audio. Maybe I'm wrong but sub out is part of multi-channel  audio.

Well, it's not monitoring multi-channel audio. ....

Then it's again a simple matter for the controller software/firmware to have one control which affects the level of both. And even better is the balance between the mains and the sub can be set and then the level control maintains that balance as you change the volume.

Now you are talking. 

Hey, this design was basically scratching an itch. You are free to design your controller in any way you want!

Of course. Regards.

 

[drask]

Any advice where to start.
Especially about input impedance circuit (op-amp) and Balanced to single-ended signal?

Thanks !

 

[moamps]

Some basic info about controller you can find here:

http://groupdiy.com/index.php?topic=26638.0

and here for input output amplifiers:

www.rane.com

for example:

http://www.rane.com/pdf/old/ava22sch95.pdf page 2

 

[drask]

Some basic info about controller you can find here:

http://groupdiy.com/index.php?topic=26638.0

and here for input output amplifiers:

www.rane.com

for example:

http://www.rane.com/pdf/old/ava22sch95.pdf page 2

Thanks