I am designing a DAC using CP2114 + PCM1794A and the audio outputs will be balanced. I am now struggling with how to implement the volume control. I will mostly use the digital volume control on my Mac but I still want to have a control over it with a normal pot.
There are nice PGAs but the problem is they are all digitally controlled and I don't want to put an MCU there. A suitable voltage controlled gain amplifier would be the ideal solution I think.
The obvious solution is to use a potentiometer to attenuate the voltage after the DAC I/V stages and then buffer it and feed to output connectors. However, I have now four voltage lines to be controlled because the signalling is balanced. Options for this:
So what, how and why?
There are nice PGAs but the problem is they are all digitally controlled and I don't want to put an MCU there. A suitable voltage controlled gain amplifier would be the ideal solution I think.
The obvious solution is to use a potentiometer to attenuate the voltage after the DAC I/V stages and then buffer it and feed to output connectors. However, I have now four voltage lines to be controlled because the signalling is balanced. Options for this:
- Use a four gang potentiometer to attenuate each signal and feed this buffered to the output connector.
- Use a two gang potentiometer to attenuate single-ended signals (after conversion), convert back to differential and feed this buffered to the output connector.
- Use a two gang potentiometer to attenuate single-ended signals after I/V stage (use only Iout+ and connect Iout- to GND), convert these to differential and feed to the output connector.
So what, how and why?
Last edited:
The good news is that balanced signal interfaces are not the same thing as differential signal processing. You can have all the common-mode noise rejection performance of a balanced signal interface while utilizing single-ended signal processing within the DAC analog stages.
I suggest obtaining a single-ended signal from each channel, whether by post I/V combining of phases via differential amplifier, or simply via the method listed in your option #3 above. However you obtain a single-ended signal it is not necessary to convert the signal to differential afterward.
Attenuate the two channels using a two gang volume pot. The output of the pots. will need to be buffered to implement a balanced interface using only a two gang pot. The balanced signal interface is then formed by feeding each single-ended buffer output to one phase of an XLR connector contact via an isolation resistor (typically, of 10 to 100 ohms). Connect the opposite phase contact of the XLR connector to signal ground through an equal value isolation resistor. The signal interface output is now balanced, even though the signal processing is being conducted single-ended.
I suggest obtaining a single-ended signal from each channel, whether by post I/V combining of phases via differential amplifier, or simply via the method listed in your option #3 above. However you obtain a single-ended signal it is not necessary to convert the signal to differential afterward.
Attenuate the two channels using a two gang volume pot. The output of the pots. will need to be buffered to implement a balanced interface using only a two gang pot. The balanced signal interface is then formed by feeding each single-ended buffer output to one phase of an XLR connector contact via an isolation resistor (typically, of 10 to 100 ohms). Connect the opposite phase contact of the XLR connector to signal ground through an equal value isolation resistor. The signal interface output is now balanced, even though the signal processing is being conducted single-ended.
Last edited:
Well I have stuck to this differential signalling but maybe without a good reason. Yes I know balanced isn't necessarily differential but is the balanced I/O in audio usually differential? PA perhaps yes? Any real benefits other that good immunity in noisy environments?
Why should I or should not use the both polarities of the current outputs of the PCM1794A? I mean is it so that I can really connect the other to ground? I cannot think a case why I shouldn't unless I want the differential output and then again I am not sure if there would be any real benefits in differential over single-ended. Effectively they are ground-connected anyway in the datasheet's examples.
Why should I or should not use the both polarities of the current outputs of the PCM1794A? I mean is it so that I can really connect the other to ground? I cannot think a case why I shouldn't unless I want the differential output and then again I am not sure if there would be any real benefits in differential over single-ended. Effectively they are ground-connected anyway in the datasheet's examples.
An externally hosted image should be here but it was not working when we last tested it.
Ok I can reply to my own question: it is because those current outputs have offset. The diff-to-single-ended stage removes this offset.
An externally hosted image should be here but it was not working when we last tested it.
The benefit of differential signal processing really has little to do with noise immunity, at least, until the differential phases are subtracted from each other, usually producing a single-ended result. Depending on how the differential outputs are created, the primary benefit with regards to differential DAC outputs is an +3dB improvement in SNR. There might also be some reduction in even-order distortion products. Common-mode noise rejection, however, is more related to having a balanced impedance signal interface.
Again, the main benefit of taking both output phases is a 3dB improvement in SNR. I have long listened to an experimental DAC of my own design which utilizes the PCM1794A in a single-ended mode (one phase output, while the opposite phase is simply grounded), so, I'm rather familiar with the excellent subjective sound quality of which it is capable. I block the D.C. offset resulting from the PCM1794A's output bias current with an coupling capacitor, post I/V conversion of course.
Last edited:
Tortuga Balanced LDR Volume Attenuator
I use Tortuga LDR volume attenuators after my DSP biamp crossover with 4 channels soon to be 6 channels. I added LDR output buffers of my own design, however, I understand that Tortuga Audio is developing a buffer option. The sq of buffered LDRs is the best I have ever heard from a volume control device. The Tortuga LDR can be implemented as fully assembled units or as DIY boards and components.
I use Tortuga LDR volume attenuators after my DSP biamp crossover with 4 channels soon to be 6 channels. I added LDR output buffers of my own design, however, I understand that Tortuga Audio is developing a buffer option. The sq of buffered LDRs is the best I have ever heard from a volume control device. The Tortuga LDR can be implemented as fully assembled units or as DIY boards and components.
In case anyone is interested, here is my schematic that isn't ready but there you can see the volume control scheme I am now planning to use. So single-ended signalling with 2-gang pot. E.g. pages four and five show it for the right channel (page five is common for both channels). Please do not comment on missing power supply connections etc...it isn't ready 
Just to show the volume control solution.
Just to show the volume control solution.- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.