I tested 8 different dual op-amps that I had in stock for what might are arguably the most important parameters in a battery powered headphone amp: Maximum undistorted output voltage into a realistic load; maximum current drain at that load; quiescent current drain.

The tests were performed using a 2kHz sine wave from my function generator. The load for the amplifier was 100 ohms for each channel. The circuit incorporates 10 ohm series resistors to prevent instability from driving overly capacitive cables. The power was supplied from a three-terminal regulated supply (a 7808 I had laying around with 2 diodes in series with its ground pin to bring the output voltage up to 9.18V). The current measured was the total drain of the circuit, including 0.6mA/ch. of Class-A biasing and 0.8mA for the power-on LED. Distortion was "measured" by looking for flat-topping of the output with only one cycle at maximum vertical resolution possible showing on my trusty THS-720A 100 MHz digital storage oscilloscope. The output level of the function generator was adjusted until severe flat-topping occured, then backed off until it went away. Results, then, are most useful for the relative comparison of each op-amp, not as absolute results, per se. The op-amps are listed in decreasing order of peak-to-peak output voltage.

LMH6672M: Iq = 30.1mA; Imax = 58.1mA; Vpp = 6.08V

LMH6655M: Iq = 21.5mA; Imax = 50.8mA; Vpp = 5.88V

AD8620AR: Iq = 13.5mA; Imax = 44.6mA; Vpp = 5.84V

NE5532AP: Iq = 16.3mA; Imax = 42.5mA; Vpp = 4.84V

LM6172AC: Iq = 12.1mA; Imax = 37.2mA; Vpp = 4.48V

TSH22IN: Iq = 12.8mA; Imax = 37.1mA; Vpp = 4.32V

LF412ACN: Iq = 11.4mA; Imax = 33.6mA; Vpp = 3.80V

MC1458CP: Iq = 8.3mA; Imax = 27.4mA; Vpp = 3.12V

Of course, the last two aren't indicated for the application, especially not the last one?!?!, but I included them to show what performance a generic FET-input amp, the LF412, and a generic bipolar-input amp, the infamous 1458, would give.

The real shocker for many will probably be the NE5532's results. Lots of people turn their noses up at this op-amp because it's been around for so long and used in so many devices (hell, it's used for every gain stage in my M-Audio Audiophile sound card), but it delivers high drive capability without the high price.

Eventually I'll get around to running RMAA 5.1 on all of them, as well as posting some of the notes I took while A/B testing them (as well as A/B testing the Class-A bias circuit which seemed to obliterate the sonic differences between the op-amps, as long as the output level was such that all of them could drive the load without problem!)

looks pretty nice jeff! Be nice if you could spare some moments to test some op-amps used a lot here: (i know you don't need this, but people can add their own)

OPA2134/2
AD861/20
OPA637
OPA627
AD823

g

AD843 please

Quote:



A common definition for clipping in an amplifier is when distortion reaches 1%. This assumes that in normal operation your circuit generates less than 1% distortion, of course. Enter RMAA.

Quote:



Believe me, I saw this coming a mile away!

Time really isn't the factor in testing them; takes all of about 15 seconds to test one (a minute if I have to solder an soic package to a Brown Dog). The real problem is that while I have many op-amps laying around, few of them happen to be the types so loved here. I tend to add a couple of new types everytime I have to order parts anyway, so more will be tested... eventually

Quote:



lol, well you said it!