Building a Portable Headphone Amplifier Based on the AN5272 Chip

momaka · 06-12-2012, 10:45 PM

Since I listen to music quite a lot using my headphones, I figured a project like this was in order. Up until this point, I have been using the audio amp on a pair of Saturn SP632-A speakers to drive my headphones. The sound from that amp doesn’t sound that great (lots of hiss and hum noise), but the only reason I still use it is because onboard audio simply doesn’t provide enough volume for me (and those motherboards that can often distort badly on full volume). Moreover, I often find that the bass from onboard audio to be seriously lacking. And this is the second reason why I kept using the amp on my Saturn speakers – bass and treble tone controls.

Now I’m sure I could probably modify and improve the amp in those Saturn speakers, but a while back I found a broken (literally) Panasonic TV on the road and snatched the boards from it since it was already busted open in pieces. Among other useful “goodies” on the main board was an AN5272 amplifier. Now, I really wouldn’t have thought about this project at all if it wasn’t for Th3_uN1Qu3’s thread below, which really inspired me (thanks man

)

http://www.badcaps.net/forum/showthread.php?t=16299

So here it goes, I’ll be building me a headphone amplifier soon.
Now, I’ve read up quite a bit on amplifier theory since the last time I posted on that thread above. However, I still feel somewhat inexperienced when it comes to amplifiers, so I would like to know if there are any improvements that I can make to my circuit below. It’s basically an almost identical copy to the application circuit in the AN5272 data sheet. The only changes I have made so far is lower the gain, Gv, from 30 dB (Av = 20 V/V) to ~16.58 dB (Av = 4.255 V/V) and add tone controls (those LPF and HPF boxes found on the inputs).

The circuit actually works – I already tested it on a breadboard. It sounds okay too, and a huge portion of the hiss that I mentioned here is now gone – all thanks to the much lower gain. Believe it or not, the circuit on the TV board had an even higher gain than the suggested one in the data sheet – it was set to 33.4 dB (29.4 V/V). No wonder it was picking up all sorts of random signals

.

Overall, I’m satisfied with the way it sounds, though. However, there are still a few things that bother me…

1) The bass is still not as good as that from my Saturn speakers. I have only a few explanations for this: at the time I was thoroughly testing the amp, I was using a pair of 330uF caps for the outputs. With my Sony MDR-023 headphones (measured about 28 Ohms impedance), that gives me a low cut-off frequency of about 17.2 Hz according to this calculator:
http://sim.okawa-denshi.jp/en/CRtool.php
Originally, I was planning on using the original 16V 1000uF caps that came with the TV audio circuit. However, would there be an audible difference if I used 16V 1500uF caps? With those, the -3 dB point is shifted to ~3.8 Hz (vs. 5.7 Hz for the 1000uF caps).

Also, I just noticed some of the changes that Th3_uN1Qu3 has made to his amplifier from the original LM1877 suggested circuit. One of them is the feedback caps – they have been changed from 10uF to 100uF. Does this play a role in making the amplifier attenuate the low frequencies less? The suggested feedback caps from the AN5272 data sheet (C9 and C10 on my circuit) are also 10uF. Should I bump them to 100uF too?
The suggested input coupling caps (C3 and C4 on my circuit) also seem tiny – put the 10uF caps in there maybe?

2) There’s still some very slight hissing from the amp, despite lowering the gain so much. I wonder if there is a way to reduce it even further. The power supply I was using for the testing (and plan to use for the finished amplifier) is a 19.5V 3.16A genuine HP power adapter that probably belonged to a laptop. Should I use something else?
Also, would a linear regulator help reduce the hissing? I have 24M12 and 78M15 regulators available.

Lastly, are there any improvements I can make to the Low Pass and High Pass Filter tone controls (perhaps any way to make the LPF boost the low frequencies even further)? I pretty much took the design for those from my Saturn speakers. The only thing I changed is that I didn’t put a 470 Ohm resistor between the LPF and the output like my Saturn speakers did – but I simply did that because I forgot to. With the tone controls, it seems that the output audio level drops quite a bit - any way to improve that? Or is it pretty much a given trade-off?

Overall, I’m not expecting miracles from the sound of this amp. As was pointed out to me by the Th3_uN1Qu3, the distortion of this amp is quite “lousy”. Then again, my Saturn speakers employ a KIA6283A amplifier IC, which has very similar specs to the AN5272 in terms of distortion. And TBH, I really don’t mind how the amp in those speakers sound. Of course if there is a way to improve my amp, let me know.

Sorry for all of the noob questions. I don’t really expect anyone to answer all of them unless they would like to. Most likely I will just experiment some more with the amp to see if I can answer some of those questions myself. However, if anyone doesn’t mind answering a few of them, I would be very grateful. I would actually like to finalize and build the amp in a few days if possible. I have all of the parts ready and a case as well. Won’t give too many details away, but the case has some see-through plastic, so I’m thinking of putting some LEDs in there too

.

Thanks for reading everyone.

budm · 06-13-2012, 12:35 AM

Did you check out this site:
http://sound.westhost.com/

You should have Buffer amp in front of the Bass/treble control to provide low source impedance for the tone circuit, may be have a little gain to compensate for the signal loss due to the filters.

999999999 · 06-13-2012, 01:15 PM

I don't mean to be negative about the project but to go to all this effort you really out to use a better performing IC than AN5272, and you don't need 4W for most cans including yours. Generally speaking, the higher the output power capability of an audio amp IC, the less detailed the sound will be, especially the treble.

The topic is quite broad for a single forum thread but there are great resources for building headamps at audio enthusiast site forums in addition to the westhost site budm linked, especially http://www.head-fi.org/f/6/diy-do-it...lf-discussions

So far as the output caps go, the ideal would be a split rail linear regulated power supply (with a different IC) which can do away with the need for output coupling caps but otherwise, yes you should go with something other than the laptop switching PSU. For your < 32 ohm cans you don't need much voltage nor do you need anywhere close to the 3A the laptop PSU is capable of.

A basic 12V DC, > 200mA wall wart (unregulated type) with a typical datasheet provided circuit after it with an LM7812 or LM317 regulator set to 12V should suffice. For high impedance cans you might want more voltage, it'd be a more versatile amp if you picked a 24V wall wart and regulate down to somewhere between 18V and 24V, depending on the max voltage the amp IC can tolerate.

You can get away with regulating down to the same voltage the wart is spec'd to output because a headamp is such a small load that a typical size unregulated wall wart will be floating at a higher voltage than the forward drop across the regulator IC.

momaka · 06-20-2012, 04:11 PM

Well, I did some experimentation with some of the caps and resistors, and this is what I found:

- The feedback caps (C9 and C10 on my schematic) should not be changed. I tried 100uF and 47uF, and both make the amp produce an awfully loud popping noise when it is turned on (highly annoying IMO). So I'll be leaving them alone at 10uF, as per the data sheet.
- Changing the output caps (C5 and C6) from 1000uF to 1500uF didn't seem to make much of an audible difference (at least I couldn't hear it).
- Changing the ripple filter cap (C2) from 100uF to 220uF didn't seem to do anything.
- Changing the input coupling caps (C3 and C4) from 1uF to 47uF didn't seem to make much of a difference in the audio either. (I guess you can call me tin-eared now

).
- I also played around with the gain resistors (R3 and R4). Sure the noise is much lower with the 4.7 kOhm resistors in there, but so is the gain. And with the LPF and HPF filters active, I have to turn up the volume knob almost all the way up to get to that loud-enough level that I'm normally used to. At that point, the output becomes too distorted IMO. So I decided to increase the gain back up a bit. Both 3 kOhm and 2.2 kOhm resistors give me enough gain without increasing the noise too much..
- Another thing I tried is swapping the 20 kOhm feedback resistors (R1 and R2) with 10 kOhm ones. The data sheet says not to do that, but I did it anyways. (Also swapped the gain resistors to 1 kOhm each to keep the overall gain more or less the same) Result: I *think* noise decreased a bit but I'm not 100% sure. So I'll probably end up leaving it at 20 kOhms.

Quote:

Originally Posted by 999999999

I don't mean to be negative about the project but to go to all this effort you really out to use a better performing IC than AN5272

Yeah I know, but I already had these parts laying around.

Since I had nothing better to do on Sunday, I decided to build the headphone amp. Pictures attached. There are still a few things that need to be done to it, like add a ON/OFF switch, re-wire the DC power jack (I only put this one temporarily), put the whole thing in a case (I have it already), and make a knob for that volume pot.

Note that I didn’t add the bass and treble tone controls. Decided I’ll play a bit more with them on the breadboard before implementing them. So far, the headphone amp sounds okay. Yes, I know I won’t win any Hi-Fi contests with it, but it more or less does what I need it to: amplify sound from my computer. Like I mentioned, most computer audio cards just don’t have enough power to drive my headphones loud enough for me. If I need less distorted but more quiet sound, I can just plug my headphones back to my computer again (or actually, now that I had that thought, I can wire a switch on the headphone amp that just bypasses it).

At least the parts are of high quality

. The IC itself is made by Panasonic, the caps are Panasonic, and even the DC jack came from a Panasonic Toughbook

. The heat sink is copper and also of high quality (any guesses where it might be from?

).

Most parts were all reclaims too (take that, Energy Star!). RoHS-compliant as well! (ugh

. I only used Lead-free solder since someone gave me a really big spool of it long time ago – hence the cloudy joints in case you were wondering why. It didn’t help that I only have a 35W unregulated Radio Shack iron.)

As far as what I used in the final design:
C9 and C10 --> 10uF 50V
C5 and C6 --> 1000uF 16V
C2 --> 100uF 25V
C3 and C4 --> 10uF 50V
R3 and R4 --> 3.3 KOhm SMD resistors
R1 and R2 --> 2 pairs of 10 KOhm SMD resistors in series (for a total of 20 KOhm)
VR1 --> 5 KOhm potentiometer since that’s all I had (works fine)

Quote:

Originally Posted by budm

Did you check out this site:
http://sound.westhost.com/

Thanks for the input. Yeah, I found that site a few months ago. Very good stuff on there.
At one point, I was actually ready to scrap this project and try to build a class-A headphone amp similar to the one shown there, but I decided against that since I didn’t feel like I have enough knowledge yet. Perhaps the next one.
...
But first, finish this one. Case coming soon.

momaka · 07-15-2012, 01:39 AM

Seems like there isn't a whole lot of interest in this thread, but here's the update regardless:

I’ve been using that headphone amp with my computer for the last few weeks. Surprisingly, it sounds a lot better than I though it would.

At first, I had it hooked just straight to my computer without the analog tone controls. But after listening to music for a few hours, I found the bass and treble to be a bit on the lacking side at low volume levels so I had to keep the volume high (which gets a bit annoying after a few ours of music). So then I hooked the tone controls (which are still on my breadboard, btw

). Since then, I haven’t really looked back to not using them. I can now turn the volume low and there is still a good amount of bass, even with my “cans”. (They actually sound a lot better than what most people would think. I have several pairs of other similarly sized headphones, and these other pairs do indeed all sound like cans. Still way way, waayy much better than ear buds, though).

Anyways, I think setting the gain to ~6 V/V (~19.56 dB with volume stage) (via R3 and R4 = 3.3 KOhm) was a good move rather than having it at ~4.26 V/V (~16.58 dB with volume stage) the way I had it in the hand-drawn schematic with the original 4.7 kOhm resistors. Noise is actually much quieter than from my Saturn speakers. I can actually hear my computer’s fans run again – and this is with the volume knob turned about halfway. As far as sound level goes – it’s almost equally loud as that from my Saturn speakers when the volume knob is turned about halfway. All of this without changing any of the internal volume settings in Windows. So all in all the volume is the same, but the noise is much less (due to lower gain). Also, that copper heat sink really is doing its job well – it gets warm after a few hours of use, but no less than the IC itself. Funny, cause I actually (re)used thermal compound from whatever was leftover on a Pentium 4 CPU

. The copper heat sink is also a re-used item – it came from a dead Sony PS3 GPU (now you don’t have to guess anymore

).

So, everything is almost done now - even the case for the amp. Normally I’m modest, but IMHO I think I did a really good job this time - the case looks quite cool. I figured, if it can’t sound (as) good, at least it will sound so in style . Pictures are coming very soon, I promise. I just need to finish the knob for the volume control – yes, it’s all handmade. I actually had to make a custom tool to make it

. It was worth it, though. Made a few other knobs as well. Just can’t decide which one to use.

All that’s left to do now is integrate the bass and treble controls inside the case (there’s still some space on the PCB). But I want to play around with them a bit more. I actually have figured out how they work now (or at least I think I do, based on my experiments so far). Will add that in my next post in case anyone is building an amp and wants to implement these in their design. In any case, I'll probably post pictures before I integrate the tone controls.

momaka · 07-28-2012, 11:23 PM

Time for pictures! (finally)

Lots of blue colors in this one

http://www.badcaps.net/forum/attachm...1&d=1343539304

What is that blue object? Let's see it from a little closer...
http://www.badcaps.net/forum/attachm...1&d=1343539304

There we go, turned on the lights...
http://www.badcaps.net/forum/attachm...1&d=1343539304
http://www.badcaps.net/forum/attachm...1&d=1343539304

Turned off:
http://www.badcaps.net/forum/attachm...1&d=1343539304
http://www.badcaps.net/forum/attachm...1&d=1343539304

The back:
http://www.badcaps.net/forum/attachm...1&d=1343539304

Just a note, the color of the LED in the amp is a lot more cyan-ish than it appears. I guess my camera couldn't pick up all of the colors. It is one of those opaque blue LEDs.
I'm thinking about using a different color LED, but haven't got to trying it yet. I have several I want to try actually. Red and orange may look fine, but I probably won't use them since it gives me a feeling as if something inside my amp was on fire.

More pictures coming in a moment...

momaka · 07-28-2012, 11:48 PM

The top side:
http://www.badcaps.net/forum/attachm...1&d=1343540760

And here's the case all empty:
http://www.badcaps.net/forum/attachm...1&d=1343540760

As you can see, the construction is quite simple. There actually aren't any screws or anything like that to hold the white piece to the clear piece. It is simply held by some small friction contacts in the clear piece. I don't think I can make it more simple than that. As for the case, this is what it's from:
http://www.badcaps.net/forum/attachm...1&d=1343540760

When I saw it several years ago, I knew I could use it for something.
Here's a few more shots, this time with a different volume knob:
http://www.badcaps.net/forum/attachm...1&d=1343540760
http://www.badcaps.net/forum/attachm...1&d=1343540760
http://www.badcaps.net/forum/attachm...1&d=1343540760

This was the first knob I made for it. I ended up making more, though, because I just thought it was too big. I don't know, what do you think? I can easily swap them since they are held by a screw to the pot shaft. The small one was supposed to have an acrylic piece like the medium one, but I didn't feel like cutting one for it.

Speaking of which, this is the tool I had to make to cut those round acrylic and wood wheels:
http://www.badcaps.net/forum/attachm...1&d=1343540760

Yes, it looks ghetto but it works fine. I got the idea from way back in engineering class in high school. Back then, my friend and I made a custom RC car as one of the projects for that class. It looked cool, so we decided to make a custom turntable for it too. Decided we were going to use acrylic for the rotating base of the turntable.
We had a professional configurable wheel cutter for our drill press, but it was too small for our purpose. It could only do about 6 inch diameter and we needed a 10" diameter circle. So we made our own and used the cutting blade from the professional one. It worked but it looked scary - having a rotating object that swings about 1/2 a foot each way at several hundred RPM on the drill press isn't something that looks exactly safe.

This one is much smaller and it's well balance, so it's quite safe. My corded drill had no problems with it whatsoever. I would NOT recommend a battery drill, though - especially if it's low RPM.
--------------

So any thoughts/suggestions about the amp?

I still haven't put the tone controls, but I will be doing that next. This means there are going to be more switches and knobs added to the amp - actually only 1 more knob and 2 switches.

The reason for having only 1 knob is that I'm thinking about setting the bass boost to max and thus eliminating the knob for it. The treble I will leave, though. Also, the 2 switches I'll have will do the following: one will select between direct input (i.e. no tone controls) and filtered output (tone controls enabled), while the other switch will select between bigger bass/treble boost but quieter sound output and louder sound output but with slightly less bass/treble (but again, I'll explain how those filters work in another post here another time).

Th3_uN1Qu3 · 07-29-2012, 06:26 AM

EDIT: I have no idea how, but i only read the first post.

I guess most of the design decisions have been taken by now, congrats on a job well done!

Quote:

Originally Posted by momaka

Originally, I was planning on using the original 16V 1000uF caps that came with the TV audio circuit. However, would there be an audible difference if I used 16V 1500uF caps? With those, the -3 dB point is shifted to ~3.8 Hz (vs. 5.7 Hz for the 1000uF caps).

Also, I just noticed some of the changes that Th3_uN1Qu3 has made to his amplifier from the original LM1877 suggested circuit. One of them is the feedback caps – they have been changed from 10uF to 100uF. Does this play a role in making the amplifier attenuate the low frequencies less? The suggested feedback caps from the AN5272 data sheet (C9 and C10 on my circuit) are also 10uF. Should I bump them to 100uF too?
The suggested input coupling caps (C3 and C4 on my circuit) also seem tiny – put the 10uF caps in there maybe?

I'd use the 1500uF caps for the output. In an amplifier with capacitor coupled output you have three highpass filters: Input capacitor with input impedance, feedback capacitor with feedback resistor to GND, output capacitor with load impedance. If all three solve to the same frequency you're going to get a 18dB/octave cut from there, if they're on different frequencies you're gonna get different slopes. It's a good idea to move all of them as low as practical, unless you're using the amplifier to listen to vinyl records, and need to eliminate turntable rumble.

10uF/1k that you have on the feedback is a -3dB point of 15Hz. Using a larger cap will help. As for the input caps, i don't see any mention of input impedance in the AN5272 datasheet. With the amplifier off and unplugged, measure resistance from pin 2 to ground and pin 5 to ground. Report your findings.

Quote:

Originally Posted by momaka

2) There’s still some very slight hissing from the amp, despite lowering the gain so much. I wonder if there is a way to reduce it even further. The power supply I was using for the testing (and plan to use for the finished amplifier) is a 19.5V 3.16A genuine HP power adapter that probably belonged to a laptop. Should I use something else?
Also, would a linear regulator help reduce the hissing? I have 24M12 and 78M15 regulators available.

Some amps don't like SMPS supplies - try a battery first. If it still hisses on the battery there's nothing you can do with it, leave it alone. If it does not hiss on the battery, try a pi filter on the power input.

Quote:

Originally Posted by momaka

Lastly, are there any improvements I can make to the Low Pass and High Pass Filter tone controls (perhaps any way to make the LPF boost the low frequencies even further)? I pretty much took the design for those from my Saturn speakers. The only thing I changed is that I didn’t put a 470 Ohm resistor between the LPF and the output like my Saturn speakers did – but I simply did that because I forgot to. With the tone controls, it seems that the output audio level drops quite a bit - any way to improve that? Or is it pretty much a given trade-off?.

I gotta go right now, so i'll only answer the last question - given that the tone controls are passive and pretty high impedance as far as i see, yes it is normal that the output level drops - because the input level drops. You would need more gain to make that up. If you make the filters active, then you get rid of that level loss, but you complicate the circuit because you'll need a couple opamps to do the job.

Depending on what the source can drive, lowering the input impedance of the filters will also lower the loss, you may find that an acceptable compromise since pretty much everything these days can drive headphones.

momaka · 07-29-2012, 11:28 PM

Quote: