All Discrete Class-A Headphone Amp

mnturner

Member
2010-05-07 2:47 am
Hey guys,
just got a pair of AKG K701's which sound great but are a little lacking in absolute volume when driven by my sources (mostly PC soundcard), so I'm going to build an amp for them. Going for extremely low distortion and decided to go single ended class-A because I could. Won't be portable due to power consumption, but if it was made push-pull and the quiescent current was reduced a bit it might be practical.
Replace output transistors with whatever you like, I'm just using these because I have them. Paralleling them probably isn't necessary, I'm just doing it to ensure against and beta droop induced distortion (however unlikely that is at these low levels of current).
Also use whatever CCS for the output stage you like, I'll either go discrete ring type or more likely LM317 as CCS.
This will be powered by bipolar LM317 regulated 12v rails.
Spice gives promisingly low distortion figures.
Any thoughts on this design? Based around a D. Self type front end.
 

Attachments

  • Class-A headphone amp.jpg
    Class-A headphone amp.jpg
    171.6 KB · Views: 3,628

mnturner

Member
2010-05-07 2:47 am
Hey thanks for the reply,
Where do you think a bootstrap is necessary? On the input?
The series miller comp caps are twin pole compensation; it's what I'm using right now on my class B 130 Watt power amp and it seems to work perfectly.
When I make the PCB I'll include provision for the second output device, if sharing is an issue I'll either increase the emittor resistors or remove the second device as at these power levels its obviously not strictly necessary unless I decide to significantly increase the quiescent current.
 
Hi mnturner
A bootstrap is typically applied to the output as in 5.22 (b) of the following diagram which is what I pictured when I viewed your schematic. Its purpose is to traditionally increase AC swing, however as you are only driving headphones increased output swing may not be necessary, A bootstrap might alleviate the need for the second output device and the problems that go along with current sharing
using a second device.

I would concentrate rather researching possible benefits of properties of a bootstrap lowering distortion, Cheers / Chris

Distortion In Power Amplifiers

"When a VAS-buffer is added, the drop in distortion is dramatic, as it is for the beta-enhancement method. The gain increase is ultimately limited by Early effect in the cascode and current-source transistors, and more seriously by the loading effect of the next stage, but it is of the order of 10 times and gives a useful effect. Fig 12 plots the distortion of a model amplifier with 100 Ohm input pair degeneration resistors, showing the extra distortion from a simple VAS. However, the beta-enhanced version has the THD submerged in the noise floor for most of the audio band, being well below 0.001%. I think this justifies my contention that input-stage and VAS distortions need not be problems; we have all but eliminated Distortions 1 and 2 from the list of eight."
 
I built a low power amp almost identical to this, but with a conventional darlington output. I went with a lower feedback network impedance and a few other values differed slightly, but the performance was first rate and distortion numbers generally under .001% across the audio band. The key to achieving those numbers in the real physical design has a lot to do with how/where you ground various parts of the circuit. My transistor compliment was MPSA56 and MPSA05 for the small stuff, and output darlingtons consisting of 2N2218+TIP32A and 2N4037+TIP31A. It ran on +/- 18VDC. (I'd dump the 3055s in a heartbeat!)
 

mnturner

Member
2010-05-07 2:47 am
Sounds like complementary feedback pair or Sizklai pairs (or however you spell that guy's name). That's my preferred output stage topology, I've used it in a couple of class B power amps.
Did your headphone amp use complementary devices with a bias setting?
I thought about that but in this design I've just used the single ended output because it avoids having to use a Vbias transistor etc. to simplify PCB layout at the expense of efficiency.
I have a feeling using only one polarity of output device with a constant current load to set the quiescent current might possible be more linear too, but I'm not sure about that and spice sims haven't shown much difference either way.
 

mnturner

Member
2010-05-07 2:47 am
I got impatient and couldn't be bothered making a PCB for the amp straight away, so in the meantime I knocked this together on perfboard. Regulated +/- 13v supply with ~50VA +/-12Vac trafo scavenged from and old integrated amp, 2200uF per rail before regulators and 470uF low ESR caps after regulators and WIMA polycarb input caps. The gain of about 4 which seems to be about right for my sources and these headphones.
I reduced the quiescent current to 125mA because I've use a pretty small heatsink (which is still only barely adequate), but even at this level the amp can still put about 480mW into my phones (more than enough to blow them up).
And all made from stuff out of the junk box
Anyway, it sounds much better than trying to run the phones straight from source and lets me get the phones up to ear-shattering levels if I'm not careful with level controls :p
 

Attachments

  • head_amp_opamp.jpg
    head_amp_opamp.jpg
    104.8 KB · Views: 2,703
Mnturner, we seem to have tread the same ground. Ignoring the errors in the schematic, I've built this same circuit for headphone use, but with a power MOSFET device and better opamp. I've posted it here previously to mostly laughs and derision. Some interesting things happen in that circuit, especially if you use a MOSFET. The output of the opamp operates at the gate turn-on voltage of the MOSFET, so it doesn't pass through zero until the level gets up higher, so for small signals there's no crossover in the opamp. I used another MOSFET with fixed gate voltage for the current sink, rather than an IC, but either works. A large heat sink is essential because there's no temperature compensation and below a certain current MOSFETS are positive just like a BJT (depending on type). I ran a hefty enough bias that I could also drive speakers at moderate levels. I'm not a big believer in "amplifier sound" but I think this design sounded incredibly good compared to anything else I've ever plugged phones into.
 

mnturner

Member
2010-05-07 2:47 am
I actually did build it in non-inverting configuration (for no particular reason), there is no pole in the feedback loop; gain compensation comes from the cap from opamp output to -ve input (without which it oscillates). There's a little ringing on square wave which will be minimized i think by adding a small lead cap across the feedback resistor.
Yeah in this build with bipolar outputs the opamp's outputs sit at about 1v dc; the output transistor's Vbe + the voltage drop across the 1000 ohm base resistors, so the opamp is in class A up to that point.
I like the principle of this design, it would beat the he'll out of those single MOSFET buffer amps people seem crazy about at the moment. Needs no output cap and the output stage (which is already quite linear) is enclosed within the opamp's feedback loop, so essentially you end up with only the distortion of the opamp.
I have some irf switching fets, I might try them out at some point.
 
I built mine with a 5534, but today I'd try a National LME opamp. The fets I used were IRF530 because those were what I had a few of. This was at least 15 years ago, so there must be better parts. I'll look for the schematic this weekend and post it if I can find it. Maybe my blameless design too. It would be interesting if somebody wanted to simulate it, since I know what it actually does in real life if properly constructed.
 

mnturner

Member
2010-05-07 2:47 am
Hi Patrick,
I'll Probably try this circuit out with MOSFETs, either some irf something's or exicon audio MOSFETs I have lying around. the DAO follower you linked has no voltage gain, which I need as none of my sources drive the phones as loud as I'd like. This necessitates some voltage gain stage of some sort, and if an opamp is used, as in this case, I don't see any point in not including it in the feedback loop which reduces the significant THD of a single FET follower and reduces output impedance to very low levels.
I'm thinking about building a tube based version at some stage which might use a FET follower similar to the one you linked for the output stage though.
Thanks for the reply