Hey all,
I am having a go at building a class-D amp but would like to incorporate some volume controls. I've been obsessing a bit over getting really nice logarithmic response in my volume control and so that lead me to Baxandall's active volume control circuit as described briefly here (http://www.ti.com/lit/ug/tidu034/tidu034.pdf) along with numerous other places.
My amp design is consists of a single sided supply and so I am wondering if anyone has ever adapted Baxandall's circuit for a single supply. If so, how did the schematic change for you? Any pitfalls to look out for? I have a 12V rail for pre-amp any other control. The power amp stage is 36V.
Also, I am not tied to the idea of Baxandall's solution so if anyone has any advice or any other solutions I'd gladly accept it 🙂 What would you do?
If it helps, the topology of my design is basically this:
SingleEndedInput->VolumeControl->SEtoDiffPreamp->PowerAmp->Speaker
Thanks in advance!
I am having a go at building a class-D amp but would like to incorporate some volume controls. I've been obsessing a bit over getting really nice logarithmic response in my volume control and so that lead me to Baxandall's active volume control circuit as described briefly here (http://www.ti.com/lit/ug/tidu034/tidu034.pdf) along with numerous other places.
My amp design is consists of a single sided supply and so I am wondering if anyone has ever adapted Baxandall's circuit for a single supply. If so, how did the schematic change for you? Any pitfalls to look out for? I have a 12V rail for pre-amp any other control. The power amp stage is 36V.
Also, I am not tied to the idea of Baxandall's solution so if anyone has any advice or any other solutions I'd gladly accept it 🙂 What would you do?
If it helps, the topology of my design is basically this:
SingleEndedInput->VolumeControl->SEtoDiffPreamp->PowerAmp->Speaker
Thanks in advance!
Better to use B.Putzeys' balanced volume control if you want a balanced output to feed your balanced input of the preamp.
The JLH uses a subtraction to generate the output.
Gain/Attenuation looks very good from -0dB to -20dB. That would give good channel matching as you change volume. But below -20db the attenuation skyrockets. I think that means the tolerance for channel matching goes out the window.
I think you will find that at attenuations below -30dB, that balance is way off centre. I habitually use attenuations in the range -20dB to -40dB with the typical +28dB power amplifiers that I use.
I do not recommend you try to cobble together an inverting opamp with variable gain.
Better to use a non-inverting Buffer after a conventional vol pot using an audio law (logarithmic) track. You can use the passive balanced output shown in the AN003 application note. It allows a SE source to feed into a balanced connection that is just as effective as any actively balanced connection. (it's what B.Putzeys uses in his implementation).
section 2.4 A simple alternative.
The JLH uses a subtraction to generate the output.
Gain/Attenuation looks very good from -0dB to -20dB. That would give good channel matching as you change volume. But below -20db the attenuation skyrockets. I think that means the tolerance for channel matching goes out the window.
I think you will find that at attenuations below -30dB, that balance is way off centre. I habitually use attenuations in the range -20dB to -40dB with the typical +28dB power amplifiers that I use.
I do not recommend you try to cobble together an inverting opamp with variable gain.
Better to use a non-inverting Buffer after a conventional vol pot using an audio law (logarithmic) track. You can use the passive balanced output shown in the AN003 application note. It allows a SE source to feed into a balanced connection that is just as effective as any actively balanced connection. (it's what B.Putzeys uses in his implementation).
section 2.4 A simple alternative.
Last edited:
Thanks Andrew!
These suggestions are great and definitely enlightening. I am not quite sure how to adapt them to a single sided supply however. At this point, I am creating a balanced signal by cascading two inverting opamps and the outputs from each stage end up being 'differential'. This is pretty straightforward and easy to do with a single sided supply. It should work well enough I believe. Not sure how much I can say about the performance ahead of time.
It's pretty interesting going through some of books and searching the internet...I was able to find some good single sided supply LPF, EQ, crossover, tone, etc circuits but nothing much for volume control to work off of. Perhaps it really is hard to do? Or it's just so blindingly obvious that...well I really am blinded by it. I could always just precede my inverting opamp stages with a log pot as a means of volume control...I guess I'll live without a really nice log volume response 🙂. In this case I'd probably just separate L and R controls and do the channel matching with my ear so I don't have to hope for the best with a ganged pot.
Or something like this (http://datasheets.maximintegrated.com/en/ds/MAX5440.pdf) will likely be fine too...it only requires a single sided supply.
These suggestions are great and definitely enlightening. I am not quite sure how to adapt them to a single sided supply however. At this point, I am creating a balanced signal by cascading two inverting opamps and the outputs from each stage end up being 'differential'. This is pretty straightforward and easy to do with a single sided supply. It should work well enough I believe. Not sure how much I can say about the performance ahead of time.
It's pretty interesting going through some of books and searching the internet...I was able to find some good single sided supply LPF, EQ, crossover, tone, etc circuits but nothing much for volume control to work off of. Perhaps it really is hard to do? Or it's just so blindingly obvious that...well I really am blinded by it. I could always just precede my inverting opamp stages with a log pot as a means of volume control...I guess I'll live without a really nice log volume response 🙂. In this case I'd probably just separate L and R controls and do the channel matching with my ear so I don't have to hope for the best with a ganged pot.
Or something like this (http://datasheets.maximintegrated.com/en/ds/MAX5440.pdf) will likely be fine too...it only requires a single sided supply.
It seems that you are confusing "differential" with balanced impedance connection.
These are very different requirements. If you want balanced because some of your equipment is balanced impedance, then it's the IMPEDANCE that needs to be balanced. This has NOTHING to do with voltage levels.
there is almost no signal processing for audio that works with balanced signals. Look at B.Putzeys' volume control. You will se he has converted the balanced impedance input to unbalanced. He then controls the volume of the unbalanced signal and then converts the unbalanced signal back to balanced impedance. He used unbalanced just to change volume level. What is available on the market to do more complex processing of balanced impedance signals?
read a lot more about balanced.
Start with Rane, Jensen, D.Self and ESP.
These suggestions were great.
Although there was good stuff in all of the sources you pointed me to, this article...Balanced Interfaces was really helpful! Still alot more to learn but I think I am beginning to understand how to get this done.
Thanks again Andrew,
-AA
Although there was good stuff in all of the sources you pointed me to, this article...Balanced Interfaces was really helpful! Still alot more to learn but I think I am beginning to understand how to get this done.
Thanks again Andrew,
-AA
Thanks for that link to the Whitlock/Jensen article in the ESP site.
It seems to be a compendium of Jensen papers/articles and I have not read this version.
For anyone interested, I should also mention that a source of my confusion was building a class-D amp(well the amp is basically already made for you in the TP32xx chips) around the TI TPA32xx chips where in the AN's they show how to convert the input from single ended to differential to accommadate whats needed for the chip input. I was wrongly assuming differential and balanced were the same thing...although superficially they are similar. Check this AN out to see what I mean... http://www.ti.com/lit/an/slaa719/slaa719.pdf
Converting to different may be good enough for my purposes and so I am not sure if I am going to try and convert from single ended to a true balanced input yet. I am not sure if there is any point as I suppose being balanced depends whats going on inside the TPA32xx chip as well. Achieving just less than 0.1% THD is probably as much as I need anyway and this is supposed to be a learning experience for myself.
Still not sure the best way to achieve volume control in this context but it would act on the SE input...before it gets to the differential or balanced stage. I like Putzeys example suggested above but I dont think it really fits here and might be superfluous. And if it does, I am not sure how to adapt it and make it useful for this project.
I apologize if this is too basic for anyone and I'm not making sense. I am a newbie lol.
-AA
Converting to different may be good enough for my purposes and so I am not sure if I am going to try and convert from single ended to a true balanced input yet. I am not sure if there is any point as I suppose being balanced depends whats going on inside the TPA32xx chip as well. Achieving just less than 0.1% THD is probably as much as I need anyway and this is supposed to be a learning experience for myself.
Still not sure the best way to achieve volume control in this context but it would act on the SE input...before it gets to the differential or balanced stage. I like Putzeys example suggested above but I dont think it really fits here and might be superfluous. And if it does, I am not sure how to adapt it and make it useful for this project.
I apologize if this is too basic for anyone and I'm not making sense. I am a newbie lol.
-AA
- Status
- Not open for further replies.