So I'm thinking of designing a headphone amp with two independent outputs. Essentially I would like to be able to connect two different headphones to the same amp with independent volume control. I havent decided on any specific design or topology but I was wondering what would be some of the problems I could have.
I know volume control would be digital (something like the DS1802 or PGA2310) since it would be remote controlled. It also means I would need a microcontroller and a separate voltage regulator for it.
So any opinions, warnings, suggestions or links to similar projects would be greatly appreciated. Note that this setup is for TV/Movie use so hi-fi is not needed.
I know volume control would be digital (something like the DS1802 or PGA2310) since it would be remote controlled. It also means I would need a microcontroller and a separate voltage regulator for it.
So any opinions, warnings, suggestions or links to similar projects would be greatly appreciated. Note that this setup is for TV/Movie use so hi-fi is not needed.
For the MCU You need a separate regulator but if You use a separate transformer too then you can implement stand-by mode and switch the power of the amps with relays. You also can control a relay for delay on the outputs too. This way You also have a mute function for each headphone.
There are plenty of functions you can do with the MCU. Auto turn on (on audio signal from the TV), adjusting the volume on different conditions (eg. one channel on tv is louder than the other, enviromental noise)... etc. And You already mentioned the remote control... which can be IR or RF.
You didn't mention if already decided on the MCU. I work with atmel mcu-s, In this case You can do all with an Atmega48 which is about $3.
There are plenty of functions you can do with the MCU. Auto turn on (on audio signal from the TV), adjusting the volume on different conditions (eg. one channel on tv is louder than the other, enviromental noise)... etc. And You already mentioned the remote control... which can be IR or RF.
You didn't mention if already decided on the MCU. I work with atmel mcu-s, In this case You can do all with an Atmega48 which is about $3.
Last edited:
I'm not really concerned with the micro aspect at the moment. Just wanted to know of any problems that could occur with having two output channels from a single input.
Especially power supply wise, which is why I mentioned the micro since there would be an asymmetrical load on a split supply.
Thanks for some micro ideas. Mute is handled by the volume control chip, both chips I mentioned have separate mute functions. I'll prolly use an AVR, I originally wanted to use an arduino so I have some code written for one. Now that I'm starting from scratch and I plan on using a PCB, I can use a standalone chip.
Especially power supply wise, which is why I mentioned the micro since there would be an asymmetrical load on a split supply.
Thanks for some micro ideas. Mute is handled by the volume control chip, both chips I mentioned have separate mute functions. I'll prolly use an AVR, I originally wanted to use an arduino so I have some code written for one. Now that I'm starting from scratch and I plan on using a PCB, I can use a standalone chip.
There shouldn't be any noteworthy problems with having two independent amps in one box. You'll want to use a buffer at the input, but that's about it. Otherwise it's just the usual star grounding / star power, yadda yadda.
When using a PGA, your distribution of levels will be different from a configuration using a volume pot. In a headphone amp you will not want to use any more than unity gain following it, and even then it probably will not be completely hiss-free on very sensitive IEMs.You can try a little resistive attenuator switched in with a Reed relay or somesuch, between PGA and output buffer.
When using a PGA, your distribution of levels will be different from a configuration using a volume pot. In a headphone amp you will not want to use any more than unity gain following it, and even then it probably will not be completely hiss-free on very sensitive IEMs.You can try a little resistive attenuator switched in with a Reed relay or somesuch, between PGA and output buffer.
^I do appreciate it though.
Anyway, why do guys think about power supply design. I'm thinking a simple voltage doubler with regulators on each rail powered from a wall transformer (like the O2). The power supply will power the micro with its positive rail so the load wont be symmetrical. Plus the amp load will vary whether 1 or 2 of the outputs will be used. Right now I'm thinking of a simple 'Cmoy' style configuration will suffice since I dont need exceptional hi-fi.
As far as your concerns with the PGA, I dont quite understand what you mean by having a unity gain after it. I havent used the PGA before or have looked at its datasheet even but I need more than unity gain after it since this is an amplfier. I havent decided on using the PGA either since I already have DS1802 digipot chips on hand and code written for them. The problem I had with them is their single supply and having to bias the signal between that. The PGA seems to be more usable in this scenario (though costly especially since I need two)
One other note is that I'll also have a simple mux to use as a source selector. My original design had one output of the mux going to the amp and other just going straight through (output to the surround sound). This seemed to work fine.
On one last note. What's the best way to simulate a power supply load? Just a resistor in parallel (in what range) or what. If it makes any difference, I use multisim since thats what I use for school.
Anyway, why do guys think about power supply design. I'm thinking a simple voltage doubler with regulators on each rail powered from a wall transformer (like the O2). The power supply will power the micro with its positive rail so the load wont be symmetrical. Plus the amp load will vary whether 1 or 2 of the outputs will be used. Right now I'm thinking of a simple 'Cmoy' style configuration will suffice since I dont need exceptional hi-fi.
As far as your concerns with the PGA, I dont quite understand what you mean by having a unity gain after it. I havent used the PGA before or have looked at its datasheet even but I need more than unity gain after it since this is an amplfier. I havent decided on using the PGA either since I already have DS1802 digipot chips on hand and code written for them. The problem I had with them is their single supply and having to bias the signal between that. The PGA seems to be more usable in this scenario (though costly especially since I need two)
One other note is that I'll also have a simple mux to use as a source selector. My original design had one output of the mux going to the amp and other just going straight through (output to the surround sound). This seemed to work fine.
On one last note. What's the best way to simulate a power supply load? Just a resistor in parallel (in what range) or what. If it makes any difference, I use multisim since thats what I use for school.
If the O2 got by with a supply like that, chances are your amp also will. Unsymmetrical loading from micro + PGA digital supply may be a concern but I have no experience with how critical that might be here - we're not exactly talking huge currents either. A resistor may do in balancing currents.
The PGA23xx series already provides a maximum gain of +31.5 dB, which is more than ample by headphone standards.
The output buffer as used in the O2 should do a fine job here.
The DS1802 seems to be a "digital pot" only. Not bad, but yes, signal biasing (and prior attenuation, ca. 3..4 dB) would be a concern.
A 4066 or somesuch? People considered them good enough for hi-fi 30 years ago at least, so they can't be that bad.
A resistor in the range of (expected supply voltage) / (expected supply current) will frequently do just fine, depends on how accurate you want to get though.
So far this is my biggest concern since this is where my last design failed, I got it working with bandaid fixes but I wasnt satisfied.
I didnt even consider the gain from the PGA. I really need to sit down and read it's datasheet before I even consider anything. I have little time to do any of this so I'm still in the theoretical stages of this build.
As a curiosity, what if this wasn't a headphone amp and I needed more gain that the PGA. How do big power amplifiers handle volume control digitally?
That's good to hear. It was in my last design and it worked just as advertised and I had no audible noise.
Alright thanks, I wanted to simulate the loading effects of the power supply.
Those will obviously have some (fixed) extra voltage gain following the PGA, though definitely less than you normally see in power amps. Some 15..16 dB would normally do, while still giving low noise levels... 60 µVrms is just fine for speaker amps.
After all, their loads are in the 80..100 dB SPL / 2.83 V / 1 m range (71 .. 91 dB SPL / 1 V), while most headphones range from about 88 to 115 dB SPL / 1 V, with very sensitive IEMs at 130+ dB SPL / 1 V and very insensitive models at 74 (K1000) or 77 dB SPL / 1 V (HE-6) being the extremes.
I'd very much hope so. It's just a bunch of FET switches, nothing much to generate noise other than added series resistance (drain-source r) there.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.