Op-amp to headphones: Forget Matching?

[DigiTan]

I'm working on this box to combine the audio from my computer, desk television, and cordless phone into one stereo headset. I have single-rail op-amps like the TL084 and LM386 available. And I want to have one of those op-amps drive the headphones directly if I can get the impedance and distortion right. Otherwise, I'll put something else there. I have a few questions:

1. What's the typical power output from a headphone driver? I can only guess that these are 16Ω to 32Ω headphones. I heard they usually draw around 40mW.
2. Is a dual-rail system really worth the trouble?
3. Does the output impedance of the headphone driver really need to be matched?

The reason for that last question is I read some conflicting articles over whether max power transfer was really necessary with modern op-amps.

Finally, one basic question... In the classic inverting configuration (**broken link removed**), the output resistance is the same as the feedback loop's resistance right?

 

[audioguru]

The minimum load resistance for most opamps is 2k ohms. A few opamps can drive 600 ohms.

Headphones are made in impedances from 8 ohms per ear to 600 ohms. The 8 ohms and 16 and 32 ohms ones need a little power amplifier to drive them. The 600 ohms ones need plenty of voltage to drive them.

The output impedance of new amplifiers never matches speakers and headphones. Old vacuum tube amplifiers matched the impedance of the output of the amplifier to the speaker.
The negative feedback of an amplifier or opamp reduces the few ohms for an amplifier or 50 ohms for an opamp open loop (no feedback) output impedance to 0.04 ohms for an amplifier or 0.00025 ohms for an opamp at low frequencies. Then the speaker or headphones has its resonances damped very well.

The LM386 is a little power amplifier, not an opamp. It is designed to drive a speaker (or headphones) with impedances down to 8 ohms. It has built-in negative feedback and its inputs work at 0V. It is not the best. It is noisy (hisss) and has a little amount of distortion.
Look in Google for The Headbanger headphones amplifier.

The site devoted to headphones amplifiers is Headwize:
**broken link removed**
They have one headphones amplifier that uses an OPA2134 dual opamp that has distortion of only 0.00008%. Then they added parts at its output to make it operate in class-A "for reduced distortion".

 

[dknguyen]

They have one headphones amplifier that uses an OPA2134 dual opamp that has distortion of only 0.00008%. Then they added parts at its output to make it operate in class-A "for reduced distortion".

My common sense is tingling. I detect a hint of sarcasm.

 

[audioguru]

My common sense is tingling. I detect a hint of sarcasm.

"Audiophiles" are crazy about reducing distortion much lower that can be heard and lower than can be measured with very sensitive equipment.

They think that very expensive wires make the sound better.
They think that wierd pointed feet on speakers make them sound better.
Some even think that distorted old vacuum tubes sound good.

The OPA2134 dual opamp has distortion of 0.00008%. Most people will not be able to hear distortion at 0.1%. I think I can hear 0.2%. Some people can hear 0.5% distortion.
Audiophiles are crazy to try to reduce 0.00008% distortion.

 

[picbits]

Probably the same Audiophiles who pay $1000 for a meter of cable ......

 

[dknguyen]

Oops. I forgot to add the graphic. Here it is!

 

[Speakerguy]

It should also be mentioned that none of these people biasing them into class A have any measurement apparatus to see if it in fact improves the performance. It's not like they all own Audio Precision's.

Me, I wouldn't mess with what TI/Burr Brown does.

 

[Speakerguy]

Also, just buy a TPA6120 chip and forget about it. That's the best that's out there currently IMO. I would go dual supply just for pop reduction.

 

[audioguru]

The distortion war is on!
National Semi's LM4562 dual opamp can drive 600 ohms with only 0.00003% distortion.

 

[Speakerguy]

Bob Pease likes the LM4562 IIRC.

The TPA6120 will do like two watts into 32ohms on +/-15V supplies. Waaay more power than anyone needs.

 

[audioguru]

I think 2W on headphones will blow your head off.

 

[walshlg]

ROTFLOL this thread is hilarious!

 

[DigiTan]

How about this single-rail/dual-rail business? All of my op-amps here are single-rail devices? Does dual-rail offer enough advantages to justify another purchase, or should I give my current system the go-ahead? I have a dual-rail inverter available, but I was hoping to save it for later.

 

[audioguru]

Any opamp will work from a single supply or a dual polarity supply if it is biased correctly. Audio power amps like an LM386 use a single supply.

A single supply causes many opamps to make a very loud POP when the power is turned on, when the power is turned off, when the input is connected and when the headphones are connected. Opamps with dual polatity supplies do not POP.

 

[DigiTan]

What causes the singles to pop while the duals do not? Can offset biasing or some other adjustment hope to fix it? I mean, the pop isn't critical, but I'd like to avoid it.

 

[Speakerguy]

On turn-on, the DC blocking caps have to charge. While that is happening, the voltage at the output (biased to 1/2 VCC with no signal) shows up on the output and exponentially decays. Sounds like a 'pop' and you'll get it in any single-supply application unless you deal with it in some fashion. A relay can be used to disconnect the output jack from the ckt until the caps are charged so you avoid the pop.

 

[bananasiong]

**broken link removed** has got it solved.

 

[audioguru]

An amplifier with a single supply has an input capacitor , output capacitor and maybe a bias voltage filter capacitor. The output makes a sudden POP sound when they charge or discharge.

With a dual polarity supply, the capacitors are not required.

 

[DigiTan]

I'll look into the Maxim chips. Will probably have to request demos though, due to budget problems.

Would it be possible to cancel that by putting a small capacitor parallel with the feedback resistance? Since the gain usually has that -Zf/Zin look to it, it seems like the cap would give the amp 0 gain at start-up. Could this possibly fix it? Or could a filter at the output stop it?

 

[audioguru]

A capacitor across the feedback resistor cuts high frequencies.
POPs are a DC problem not a high frequency problem.

You cannot filter out a POP. it is not a single frequency. It is a sudden DC level shift with a lot of harmonics.