I've been frustrated with many of the common DIY headphone amplifiers designs, primarily because they all require two batteries to obtain split supplies, and many of them either use op-amps in situations where they were not intended or ugly discrete messes for the requisite gain. I have also been frustrated by many of the common designs not using surface mount parts, so after some puzzling, I have started to come up with a design, and I was curious about what anyone thought.
It would use a TI TPA6120A2 class-AB headphone amplifier power op amp, with +-15v rails supplied by a TI TPS65130 positive & negative switching power supply controller (...that will be adequately filtered and running at 1.2MHz, so I see no conceivable way it would degrade the audio quality to any meaningful degree...). I have also been poking at thinking about adding some kind of DAC, particularly one that could run hosted from an Android smartphone...but I haven't come to conclusions yet on that. It will be powered by a single lithium polymer cell (~3.7 volts) and use the TPS65130 to boost it to the rails. It will then include a MCP73841 Microchip lithium polymer charger to then recharge the cell!
All pretty simple, although the switcher is pretty messy to get the split rails. Quite a handy chip though, if in a bear of a no-lead package. Everything will be surface mount, and in the end, it should be two inches square or less.
Thoughts? Anybody want one (ha!)?
It would use a TI TPA6120A2 class-AB headphone amplifier power op amp, with +-15v rails supplied by a TI TPS65130 positive & negative switching power supply controller (...that will be adequately filtered and running at 1.2MHz, so I see no conceivable way it would degrade the audio quality to any meaningful degree...). I have also been poking at thinking about adding some kind of DAC, particularly one that could run hosted from an Android smartphone...but I haven't come to conclusions yet on that. It will be powered by a single lithium polymer cell (~3.7 volts) and use the TPS65130 to boost it to the rails. It will then include a MCP73841 Microchip lithium polymer charger to then recharge the cell!
All pretty simple, although the switcher is pretty messy to get the split rails. Quite a handy chip though, if in a bear of a no-lead package. Everything will be surface mount, and in the end, it should be two inches square or less.
Thoughts? Anybody want one (ha!)?
Last edited:
You should try a search on the forum here for TPA6120. There have been a boatload of posts about the chip over the past few years and you will come up with a good week's worth of reading! 
Here is a typical thread to get you started:
http://www.diyaudio.com/forums/head...amp-there-anyone-can-give-me-some-advice.html
In there P-A mentions his QRV09 TPA6120 based headphone amp. RocketScientist/NwAvGuy, the fellow who made the O2 headphone amp, did a review of that QRV09 here:
NwAvGuy: QRV09 DIY Headphone Amp
Generally, as I recall, folks say the chip sounds fine but the one hiccup is a minimum of 10 ohms in series is needed for stability. Not that big of a deal if you are powering high impedance headphones, but the damping factor may start to suffer with really low impedance cans.
Here is a typical thread to get you started:http://www.diyaudio.com/forums/head...amp-there-anyone-can-give-me-some-advice.html
In there P-A mentions his QRV09 TPA6120 based headphone amp. RocketScientist/NwAvGuy, the fellow who made the O2 headphone amp, did a review of that QRV09 here:
NwAvGuy: QRV09 DIY Headphone Amp
Generally, as I recall, folks say the chip sounds fine but the one hiccup is a minimum of 10 ohms in series is needed for stability. Not that big of a deal if you are powering high impedance headphones, but the damping factor may start to suffer with really low impedance cans.
I just had an idea for an alternate suggestion since it sounds like you may have surface mount layout skills. Consider dropping the TPA6120 and instead build a version of opc's Wire amplifier here (all surface mount)
http://www.diyaudio.com/forums/head...igh-performance-headphone-amplifier-pcbs.html
Which uses LME49600 buffer chips (250mA each and vanishing low distortion) in a loop with a LME49990 (also has vanishingly low distortion). BUT, in your case, add that dual-rail DC-to-DC power supply you proposed. The Wire power supply is all linear on a different PCB. I've noodled around trying to power one of my Wire boards with a +/-15Vdc DC-DC dual rail converter (about $45 or so at Mouser for that wattage) and fit it all into an Altoids can. I think it would fit, from the mechanical measurements. I'm not sure that 2 square inches would be enough heat sink though for the Dpak LME49600s, depending on how hard they are driven. Would have to go through the thermal calculations.
I've also pondered ways to add a lithium cell controlled by a lithium charge controller chip to the O2 amplifier, so that addition you are proposing to your design would certainly be interesting too. MUST use a charge controller chip with lithium, as you are proposing, too dangerous otherwise for trickle charge, etc.
The LME49600s pull a fair amount of quiescent current and the DC-DC converter will probably loose 15% or so. But still, with a lithium 3.7V cell, that should be a few hours at least on the battery.
At the very least there are probably folks out there who didn't get in on the Wire PCB group buy last year who may be interested in a group buy. I would probably have interest in a PCB group buy if that dual-rail DC-DC supply and Lithium charge controller chip were included in the design. But opc's design results were measured on a dScope (Audio Precision equivalent), which was part of the draw, in knowing for sure what the audio measurement results were...
Good luck!
Consider dropping the TPA6120 and instead build a version of opc's Wire amplifier here (all surface mount)http://www.diyaudio.com/forums/head...igh-performance-headphone-amplifier-pcbs.html
Which uses LME49600 buffer chips (250mA each and vanishing low distortion) in a loop with a LME49990 (also has vanishingly low distortion). BUT, in your case, add that dual-rail DC-to-DC power supply you proposed. The Wire power supply is all linear on a different PCB. I've noodled around trying to power one of my Wire boards with a +/-15Vdc DC-DC dual rail converter (about $45 or so at Mouser for that wattage) and fit it all into an Altoids can. I think it would fit, from the mechanical measurements.
I'm not sure that 2 square inches would be enough heat sink though for the Dpak LME49600s, depending on how hard they are driven. Would have to go through the thermal calculations.I've also pondered ways to add a lithium cell controlled by a lithium charge controller chip to the O2 amplifier, so that addition you are proposing to your design would certainly be interesting too. MUST use a charge controller chip with lithium, as you are proposing, too dangerous otherwise for trickle charge, etc.
The LME49600s pull a fair amount of quiescent current and the DC-DC converter will probably loose 15% or so. But still, with a lithium 3.7V cell, that should be a few hours at least on the battery.
At the very least there are probably folks out there who didn't get in on the Wire PCB group buy last year who may be interested in a group buy. I would probably have interest in a PCB group buy if that dual-rail DC-DC supply and Lithium charge controller chip were included in the design.
But opc's design results were measured on a dScope (Audio Precision equivalent), which was part of the draw, in knowing for sure what the audio measurement results were...Good luck!
Last edited:
Yeah, I'm aware of the ten ohms -- considering that it will be able to drive up to 15 volts, it will still easily be able to drive 750mA into ten ohm headphones -- something I'm not really intending to do -- (or own!)
Interesting idea on "The Wire", as it is basically using high performance op amps, like the rest of these. The trick is whether or not you're getting any improvements, and those have fairly high quiescent. I have an Audio Precision, so I was planning on characterizing this when I build it in a few weeks (finishing up the schematics and board now)...
However, if there's interest for a single-supply switcher variant of "The Wire", I'd be happy to lay it out and measure it, if people want one. The switcher that I'm using is about $4 in parts, a pretty far cry from the modules.
Interesting idea on "The Wire", as it is basically using high performance op amps, like the rest of these. The trick is whether or not you're getting any improvements, and those have fairly high quiescent. I have an Audio Precision, so I was planning on characterizing this when I build it in a few weeks (finishing up the schematics and board now)...
However, if there's interest for a single-supply switcher variant of "The Wire", I'd be happy to lay it out and measure it, if people want one. The switcher that I'm using is about $4 in parts, a pretty far cry from the modules.
Great news that you have an AP! I mixed that up in my post. NwAvGuy/RocketScientist, who designed the O2 amp and ODAC DAC is the one with the dScope. opc has an AP. In both cases it was great to see the actual measured results with crosstalk, noise floor, intermodulation, frequency & phase response, etc.
In your project's case measurements will be especially good to verify no switcher artifacts making it into the audio via conduction or radiated EMI. I fully expect that with a high frequency switcher chip like that it should be possible to get any converter EMI filtered out. Yeah doing the DC to DC converter yourself with the parts would really cut the cost down!
If you should make a Wire variant it may be interesting to have a goal of "portability". opc's goal with the Wire was top audio results using an external stepped attenuator and RCA connectors for lower crosstalk. But instead a RK097 series minipot and 1/8 inch PCB mount in/out audio jacks, like on the O2 headphone amp, could be used and maybe the whole thing would fit in an Altoids can or the mini^3 Hammond case. Even the battery might fit in if a cell phone flat pack batt could be used. lol - a "CMOY on steriods".
I suspect a lot of commercial portable headphone amps are using a single 3.7V lithium + dual rail DC-DC converter, since I see a lithium battery on some of the specs. But I'm not aware of many DIY designs with that. The need for all surface mount is probably what has damped some DIY enthusiasm.
In your project's case measurements will be especially good to verify no switcher artifacts making it into the audio via conduction or radiated EMI. I fully expect that with a high frequency switcher chip like that it should be possible to get any converter EMI filtered out. Yeah doing the DC to DC converter yourself with the parts would really cut the cost down!
If you should make a Wire variant it may be interesting to have a goal of "portability". opc's goal with the Wire was top audio results using an external stepped attenuator and RCA connectors for lower crosstalk. But instead a RK097 series minipot and 1/8 inch PCB mount in/out audio jacks, like on the O2 headphone amp, could be used and maybe the whole thing would fit in an Altoids can or the mini^3 Hammond case. Even the battery might fit in if a cell phone flat pack batt could be used. lol - a "CMOY on steriods".
I suspect a lot of commercial portable headphone amps are using a single 3.7V lithium + dual rail DC-DC converter, since I see a lithium battery on some of the specs. But I'm not aware of many DIY designs with that. The need for all surface mount is probably what has damped some DIY enthusiasm.
Last edited:
I forgot about this thread from last year where RocketScientist and jcx discuss a bunch of stuff about the TPA6120. They try shunting the output 10R with an inductor to lower the output impedance. Posts 73 - 120:
http://www.diyaudio.com/forums/head...tional-lme49600-reference-design-project.html
Sounds like the inductor worked somewhat to lower the output Z but caused some ringing with square wave 10kHz drive.
Also has occurred to me that just the power supply part of your design alone would be useful for other amps to add a lithium battery. The TPA6120 could simply be left off a PCB and the +/-15Vdc cabled over to power a Wire or O2 headamp. Both wouldn't need more thant 200mA max per channel, and usually a lot less for most headphones.
Happy holidays!
http://www.diyaudio.com/forums/head...tional-lme49600-reference-design-project.html
Sounds like the inductor worked somewhat to lower the output Z but caused some ringing with square wave 10kHz drive.
Also has occurred to me that just the power supply part of your design alone would be useful for other amps to add a lithium battery. The TPA6120 could simply be left off a PCB and the +/-15Vdc cabled over to power a Wire or O2 headamp. Both wouldn't need more thant 200mA max per channel, and usually a lot less for most headphones.
Happy holidays!
I like this idea and would be interested in some boards, especially if they can be offered with the really difficult part(s) pre-soldered.
Perhaps the power supply and amp sections could be arranged so it could either be used all together or separated and maybe stacked to make it fit a certain case size, or if people want to just use one part or the other.
I would also suggest that the board be a little narrower than the width of a standard iPhone/iPod so it could be built into a case that stacks nicely with the player.
Perhaps the power supply and amp sections could be arranged so it could either be used all together or separated and maybe stacked to make it fit a certain case size, or if people want to just use one part or the other.
I would also suggest that the board be a little narrower than the width of a standard iPhone/iPod so it could be built into a case that stacks nicely with the player.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.