Cayin IHA6 Measurements

This unit was loaned to me to test by a supporter.
Content and analysis made possible thanks to support from https://headphones.com, and https://patreon.com/goldensound supporters.

Chit-Chat:

The IHA-6 is a discrete, fully balanced headphone amplifier with some rather interesting features. Primarily the fact that it offers a variety of output impedances to suit/tune your headphone, and a ‘high current’ mode that promises to better power current-hungry planar magnetic headphones.

Coming in at $949, it’s a relatively expensive amplifier, though much more affordable than the likes of a Ferrum OOR or Enleum AMP-23R. However, given the extremely high quality construction of the chassis, with a beautiful and robust feeling sandblasted finish, as well as what certainly appears to be more than thorough and high quality internal design and selection of components, its not hard to see why this costs more than the likes of a Singxer SA-1 or Topping A90.

(Please forgive the poor quality of the photos in this post, having some camera issues which will be sorted soon)

The internals of the IHA-6 are almost entirely discrete components, with a high quality shielded transformer, separate power supplies thereafter for the left and right channels, various safety/protection implementations, and looking on the underside of the device, a large bank of transistors which sink their heat to the chassis of the device. Quite a large number of them for a headphone amplifier!

The white substance is thermal paste to allow for better heat transfer to the chassis.

Interestingly three of the transistors seem to be different to the others, I’m not entirely sure what the differences are as the part numbers are not visible.

One unique feature of the IHA-6 is that whilst the balanced output provides a very low output impedance of 0.15Ohms, the single ended outputs provide impedances of either 10 or 120 Ohms. This will mean that with dynamic driver headphones, the frequency response will be altered, most often giving an increase in midbass frequencies (though it is dependent on the impedance curve of the headphone).

This is facilitated by the inclusion of 10 Ohm and 120 Ohm resistors in the signal path which are visible near the outputs of the amplifier.

This is something that can be subjectively quite enjoyable with many headphones. Various models such as the Sennheiser HD800 or the ZMF Verite have a reputation for sounding better on high OI sources, so it’s nice to have this option on the amp, and I wish more of them did!

Measurements:

Test Setup:

– Audio Precision APx555 B-Series Analyzer
– Amplifier was warmed up for 6 hours before measuring
– AudioQuest Mackenzie XLR interconnects
– Neurochrome Headphone Dummy Load
– Results shown are using the standard current mode unless otherwise specified (as it performed better in this mode), however full reports in various configurations are available below
– Full reports containing additional data and test configurations are attached
– Exact analyzer configurations for each measurement are detailed in the full reports
– All measurements shown below are taken with a 32 ohm load in low gain unless otherwise specified

Full Reports:

Reports available here (These contain 4V unity gain, 700mV and 50mV test results, please check you’re looking at the right ones when comparing!):

Output impedance: 0.15 Ohm across full range (Unless using the 10 Ohm or 120 Ohm outputs)

Power:

The spec says that the amplifier will do 5W @ 32 Ohm in ‘high current’ mode, and 7W @ 32 Ohm in the standard mode.
My testing showed that this was very accurate, and the amplifier begins to clip slightly above each of these figures.

When using a 300Ohm load, the amplifier delivered over 1.2W without issue, and had to be switched to high gain to get it to clip at all. With a 4V DAC and the amp in low gain, the volume could be maxed without reaching a limit in low gain.

With a 32 Ohm load, distortion begins rising at about 0.2W.

This does mean that despite what the name might imply, ‘high current’ mode actually results in less power available, not more.
So, if it’s not increasing available current, what is it doing?

I reached out to Cayin to ask, and they informed me that it was altering the bias on the transistors. So the amplifier is biased further into class A when high current is selected, however, it is not running full class A. Both modes are class AB, just with a different bias.
This can be verified by monitoring the power consumption of the device when idle vs under load.
A full class-A amplifier such as the HeadAmp GS-X MK2 pulls the same wattage from the wall regardless of whether it is sitting idle or pumping 7W into the load.

However when idle, the IHA6 pulls about 19W from the wall in std current mode, and 20W in high current mode. But ramps up to over 50W as the amplifier delivers more power to the headphones/load. So it is definitely class AB in both modes of operation.

‘High Current’ is not actually increasing the amount of current available to drive the headphones, but rather just adjusting the static/quiescent current for the transistors. So it’s mostly a slight change in sound and distortion profile, rather than making the amp able to drive different loads better.

They also confirmed that the IHA6 is NOT a zero-feedback amplifier, and is using feedback in both modes of operation, which there seemed to be some confusion about online. (Even though the IHA6 has never been marketed as zero-feedback so I’m not sure how this rumour started.)

1khz 4V input 4V Output 32 ohm (Std Current):

At high output levels like this the IHA-6 is doing a nice job, and also performing significantly better than the spec would imply. The spec claimed a THD+N of 0.02%, but we are seeing 0.0015%. Over 13x lower distortion, and the other channel is over 24x lower!

1khz 4V input 4V Output 32 ohm (High Current):

Swapping to high current we see a slight reduction in performance, and distortion profile now has equal 2nd and 3rd order harmonics, rather than being more 2nd order originally.

1khz 4V input 4V Output 300 ohm (Std Current):

With a 300 Ohm load, we see quite a bit better distortion performance, as is typical due to the fact that the amplifier is having to supply less current for the same output level. It’s effectively not needing to work as hard.

But of course, you’re unlikely to be listening at 4V output, so lets have a look at a more realistic 700mV output level.

1khz 4V input 700mV Output 32 ohm (Std Current):

Performance here is still very good, with all harmonics and PSU noise being below -110dB.

1khz 4V input 700mV Output 32 ohm (High Current):

Once we swap to high current, we actually see a slight improvement in performance now, with each harmonic being a few dB lower than in Std Current mode.
It seems that you might get better performance with easier to drive headphones in the High Current mode, but better performance with harder to drive headphones in the standard current mode.
Though try both for yourself and see which you like with your headphones.

THD+N / Frequency:

Total harmonic distortion gets higher as you move into the higher, or lower frequencies. The behaviour does change slightly depending on whether you’re in high or std current mode.

But whilst it’s not unusual to see a rise into the higher frequencies (typically providing a subjectively sweeter or warmer sound, Rebel Amp being a good example), I’ve not seen many amps where it also rises into the lower frequencies. This could potentially explain the slight softening effect this amp seemed to have on macrodynamics and percussion in my subjective experience, but you’ll have to let your ears decide.

Interestingly in Std Current mode one channel performs slightly differently to the other. But in high current mode both are the same.

IMD (SMPTE) vs Output Level:

Crosstalk:

Crosstalk is quite high, and remained at this level even with lower signal levels.

1khz Square Wave:

Multitone:

One channel has higher distortion than the other in std current mode. But the multitone overall mirrors the THD+N vs freq behaviour we saw earlier, as expected.

Overall, from an objective standpoint the IHA-6 seems to be fine. Not necessarily pushing the boundaries of performance, as other discrete amplifiers such as the more expensive Ferrum OOR, or less expensive Singxer SA-1 outperform it in almost all areas, but also potentially offering a sound signature that will be preferable to many due to the rising distortion in high and low frequencies giving a slightly laid back sound, whilst leaving midrange performance to shine through.

Though the added features such as the high output impedance offerings, and ability to slightly tune the sound using the high/std current option, as well as the fantastic build quality and aesthetics, will be big attractions for many.