Linearity is one of the hallmarks of a well-designed amplifier. A great deal of effort goes into keeping distortion as low as possible in the recording chain. Why should your headphone amplifier undo all this work by adding its own distortion and possible coloration?
The UHA-6 and 6S series amplifiers are designed to provide a low-distortion and low-noise output into all types of headphone loads. The graphs below demonstrate this by showing harmonic distortion across the audio frequency band of 20 Hz to 20 kHz. In order to see how the UHA-6S stacks up against the competition, I also took some comparison data on a popular DAC/Amp from a competitor. The competitor shall remain nameless, but the model I’m using for this comparison has a similar feature set, size and price point as the UHA-6S. The UHA-6S outperforms the competitor amp in all cases I tested.
My Test Setup
My audio analyzer is a Prism Sound dScope III. The DAC/Amps under test were driven by the dScope digital optical output, using a 48-kHz sample rate and word length of 24 bits. I used a single-tone stepped sweep at a level of -1 dB full-scale. The amplifier output was connected via a length of three-conductor headphone cable to a dummy load consisting of 1/4-watt resistors (same load on each channel). The analog input on the analyzer was connected across the dummy load on one channel. (I’m assuming the left and right channels have nearly identical performance.) The amps were set to low-gain mode except in the cases where I needed the extra gain to get the appropriate output level. The gain knobs were dialed to give the desired power across the load.
I decided to test at two different output powers: 1 mW and 10 mW. That may not sound like much, but even 1 mW is enough to produce uncomfortably high sound pressure levels on many headphones and earphones. (My Sennheiser HD 600, for example, has a sensitivity of 97 dBSPL/mW, which means 1 mW of applied power produces 97 dB SPL – about as loud as a “typical” car stereo at max volume. Some earphones have sensitivities greater than 115 dBSPL/mW – at 1 mW, that’s approaching the threshold of pain!) An increase from 1 mW to 10 mW is roughly equivalent to a doubling of subjective loudness. In any case, 1 mW of output power should not be beyond the expected capability of any headphone amplifier.
The graphs show distortion as a percentage of the desired signal, and the curves are in groupings of three:
THD+N (Total Harmonic Distortion Plus Noise) – this is a measure of everything added by the amplifier, including noise.
2HD (Second Harmonic Distortion) – this is the level of the second harmonic of the signal (the first overtone).
3HD (Third Harmonic Distortion) – this is the level of the third harmonic of the signal (the second overtone).
The measurement bandwidth used was DC to 80 kHz, and no weighting filters were applied.
The Results
Let’s start with a “typical” load impedance of 32 ohms, at 1 mW of power dissipation (click the thumbnail to enlarge):
Notice how the curves tend to rise at high frequency. This is typical for most types of amplifiers designs, including high-end power amplifiers for driving speaker loads. The THD+N curve is naturally higher than the 2HD and 3HD curves because it includes all harmonics and the noise floor.
Now let’s take a look at how the competitor amp does with the same load conditions:
The scale has changed, in order to try to fit the competitor’s THD+N and 3HD curves, which have a significant rise above 1 kHz. Comparing THD+N at 1 kHz, the competitor is about five times higher than the UHA-6S.
To understand what’s going on, it sometimes helps to look at the actual spectrum of the signal. On the left is the UHA-6S output spectrum for a 1 kHz signal at 1 mW into 32 ohms, and on the right is the competitor:
These are graphs of actual signal voltage level (expressed in dBV) versus frequency. They are 128k-point FFTs with 4 averages. In both graphs you can see the 1 kHz fundamental and its harmonics. Notice how the harmonics from the competitor amp are higher in level and extend all the way to 20 kHz. High-order harmonics such as this are usually considered to be an especially unpleasant form of distortion. A few other characteristics also come to light: the UHA-6S has a lower noise floor, and the competitor amp has some noticeable low-level spreading of the 1 kHz fundamental, possibly due to clock jitter on the DAC.
Now let’s step up the power to 10 mW, still into a 32 ohm load:
Note that the THD+N curve from the UHA-6S is actually lower when compared with the 1 mW curve. This is because the 10 mW signal is higher above the noise floor than the 1 mW signal (and remember that THD+N includes a measure of the total noise).
The competitor amp doesn’t fare so well:
Notice the scale on the distortion axis. It’s difficult to even show these two sets of curves on the same graph. The competitor amp reaches a THD+N of 0.22% at low frequency. The distortion harmonics of a signal at 20 Hz are only 53 dB below the fundamental signal. Surprisingly, the manufacturer claims this amp can deliver up to 100 mW into 32 ohms. I was only able to get it to 50 mW before severe clipping set in.
All other things being equal, a lower impedance presents a heavier load to an amplifier than a higher impedance. The heavier load requires more current, and not all amplifiers handle this well. Here we have the curves for 1 mW into a 16 ohm load:
The UHA-6S THD+N at 1 kHz is 0.006%, compared to 0.005% for the same power into a 32 ohm load. At 1 mW into 16 ohms, the competitor shows a THD+N at 1 kHz of 0.048%, eight times higher than the UHA-6S. The low-frequency bumps in the competitor curves indicate multiple distortion mechanisms are at work.
At 10 mW into 16 ohms, the UHA-6S is still performing quite admirably:
As with the 32 ohm load, notice that THD+N is lower compared with the 1 mW case, due to the signal level being higher above the noise floor.
At 1 mW into 100 ohms, the UHA-6S second and third harmonic distortion is down in the noise floor of the measurement instrument. Compare this to the competitor:
At 10 mW into 100 ohms, the competitor distortion is off the scale (0.7% THD+N at 1 kHz), but the UHA-6S is still looking very good:
At a power of 1 mW into a load of 300 ohms, the UHA-6S distortion is exceedingly low, but the competitor amp has all but given up:
Finally, at 10 mW into 300 ohms, the UHA-6S performance shines. The competitor amp went into clipping at 6 mW and hence is not shown:
When it comes to low-distortion performance, not all headphone amplifiers are created equal. Choosing a low-distortion amplifier means you are less likely to experience unwanted and unpleasant coloration of the sound. The Leckerton Audio UHA-6S and UHA-6 amps deliver a highly linear output into all types of headphone loads.
UHA-6S Performance Measurements: Harmonic Distortion
Linearity is one of the hallmarks of a well-designed amplifier. A great deal of effort goes into keeping distortion as low as possible in the recording chain. Why should your headphone amplifier undo all this work by adding its own distortion and possible coloration?
The UHA-6 and 6S series amplifiers are designed to provide a low-distortion and low-noise output into all types of headphone loads. The graphs below demonstrate this by showing harmonic distortion across the audio frequency band of 20 Hz to 20 kHz. In order to see how the UHA-6S stacks up against the competition, I also took some comparison data on a popular DAC/Amp from a competitor. The competitor shall remain nameless, but the model I’m using for this comparison has a similar feature set, size and price point as the UHA-6S. The UHA-6S outperforms the competitor amp in all cases I tested.
My Test Setup
My audio analyzer is a Prism Sound dScope III. The DAC/Amps under test were driven by the dScope digital optical output, using a 48-kHz sample rate and word length of 24 bits. I used a single-tone stepped sweep at a level of -1 dB full-scale. The amplifier output was connected via a length of three-conductor headphone cable to a dummy load consisting of 1/4-watt resistors (same load on each channel). The analog input on the analyzer was connected across the dummy load on one channel. (I’m assuming the left and right channels have nearly identical performance.) The amps were set to low-gain mode except in the cases where I needed the extra gain to get the appropriate output level. The gain knobs were dialed to give the desired power across the load.
I decided to test at two different output powers: 1 mW and 10 mW. That may not sound like much, but even 1 mW is enough to produce uncomfortably high sound pressure levels on many headphones and earphones. (My Sennheiser HD 600, for example, has a sensitivity of 97 dBSPL/mW, which means 1 mW of applied power produces 97 dB SPL – about as loud as a “typical” car stereo at max volume. Some earphones have sensitivities greater than 115 dBSPL/mW – at 1 mW, that’s approaching the threshold of pain!) An increase from 1 mW to 10 mW is roughly equivalent to a doubling of subjective loudness. In any case, 1 mW of output power should not be beyond the expected capability of any headphone amplifier.
The graphs show distortion as a percentage of the desired signal, and the curves are in groupings of three:
The measurement bandwidth used was DC to 80 kHz, and no weighting filters were applied.
The Results
Let’s start with a “typical” load impedance of 32 ohms, at 1 mW of power dissipation (click the thumbnail to enlarge):
Notice how the curves tend to rise at high frequency. This is typical for most types of amplifiers designs, including high-end power amplifiers for driving speaker loads. The THD+N curve is naturally higher than the 2HD and 3HD curves because it includes all harmonics and the noise floor.
Now let’s take a look at how the competitor amp does with the same load conditions:
The scale has changed, in order to try to fit the competitor’s THD+N and 3HD curves, which have a significant rise above 1 kHz. Comparing THD+N at 1 kHz, the competitor is about five times higher than the UHA-6S.
To understand what’s going on, it sometimes helps to look at the actual spectrum of the signal. On the left is the UHA-6S output spectrum for a 1 kHz signal at 1 mW into 32 ohms, and on the right is the competitor:
These are graphs of actual signal voltage level (expressed in dBV) versus frequency. They are 128k-point FFTs with 4 averages. In both graphs you can see the 1 kHz fundamental and its harmonics. Notice how the harmonics from the competitor amp are higher in level and extend all the way to 20 kHz. High-order harmonics such as this are usually considered to be an especially unpleasant form of distortion. A few other characteristics also come to light: the UHA-6S has a lower noise floor, and the competitor amp has some noticeable low-level spreading of the 1 kHz fundamental, possibly due to clock jitter on the DAC.
Now let’s step up the power to 10 mW, still into a 32 ohm load:
Note that the THD+N curve from the UHA-6S is actually lower when compared with the 1 mW curve. This is because the 10 mW signal is higher above the noise floor than the 1 mW signal (and remember that THD+N includes a measure of the total noise).
The competitor amp doesn’t fare so well:
Notice the scale on the distortion axis. It’s difficult to even show these two sets of curves on the same graph. The competitor amp reaches a THD+N of 0.22% at low frequency. The distortion harmonics of a signal at 20 Hz are only 53 dB below the fundamental signal. Surprisingly, the manufacturer claims this amp can deliver up to 100 mW into 32 ohms. I was only able to get it to 50 mW before severe clipping set in.
All other things being equal, a lower impedance presents a heavier load to an amplifier than a higher impedance. The heavier load requires more current, and not all amplifiers handle this well. Here we have the curves for 1 mW into a 16 ohm load:
The UHA-6S THD+N at 1 kHz is 0.006%, compared to 0.005% for the same power into a 32 ohm load. At 1 mW into 16 ohms, the competitor shows a THD+N at 1 kHz of 0.048%, eight times higher than the UHA-6S. The low-frequency bumps in the competitor curves indicate multiple distortion mechanisms are at work.
At 10 mW into 16 ohms, the UHA-6S is still performing quite admirably:
As with the 32 ohm load, notice that THD+N is lower compared with the 1 mW case, due to the signal level being higher above the noise floor.
At 1 mW into 100 ohms, the UHA-6S second and third harmonic distortion is down in the noise floor of the measurement instrument. Compare this to the competitor:
At 10 mW into 100 ohms, the competitor distortion is off the scale (0.7% THD+N at 1 kHz), but the UHA-6S is still looking very good:
At a power of 1 mW into a load of 300 ohms, the UHA-6S distortion is exceedingly low, but the competitor amp has all but given up:
Finally, at 10 mW into 300 ohms, the UHA-6S performance shines. The competitor amp went into clipping at 6 mW and hence is not shown:
When it comes to low-distortion performance, not all headphone amplifiers are created equal. Choosing a low-distortion amplifier means you are less likely to experience unwanted and unpleasant coloration of the sound. The Leckerton Audio UHA-6S and UHA-6 amps deliver a highly linear output into all types of headphone loads.