Fiio K11 R2R DAC/HPAMP measurements

[cay-uwe]

I also made some basic measurement of the FiiO K11 R2R DAC using the SPDIF Coax input with a CD-Player attached to it that delivered several of the measuring signals. For low level signal I used dithered sine.

This is the linearity diagramm I have got at 1kHz 0, -30, -60 and -90dBFS. Lower was not included since I was using PCM 16bit / 44.1kHz signals.

Tendency is similar to the measurement at the 1st post here, except I get lower signal levels at -90dBFS.

A spectral analysis of a 1kHz sine at 0dBFS, 6dBV or 2Vrms at the RCA PRE OUT looks as follows.

Notice that I am using dBV scaling and not dBFS, meaning the output level is about 6dBV which correlates to 0dBFS. Like the measurement shown initially I also detected several harmonics above 1 kHz and several other distortion because of oversampling and undithered signals. Below 1kHz one sees some spikes that may be power supply related or intermodulation effects due to the higher level distortions over 1kHz.

I marked with orange the distortion levels that are defenitely audible, while the green line marks a typicall limit from which on for most listeners it gets inaudible.
The measured THD and THD+N are above the audible range and SNR is close to that what is considered as minimum.

To analyze influences of power supplies and other sources a measurement at -90dFS is of interest.

Again, be aware that the scaling is dBV for the level. I marked with a red point the level that should be reached when using a -90dBFS sine. That is pretty well reached, a bit lower as one can see in the linearity diagramm, but it is fine. One detects no spikes or disturbences above an below 1kHz, meaning that the power supply included works very good and causes no interferences.

Croos talk from left to right channel is actually very good as one can see from the next chart.

When applying 0dBFS or 6dBV on the left channel one measures a signal of -118dB lower at the right site. That is actually a very good value ensuring practically no crosstalk.

Last but not least the frequency response using the USB input and ASIOall drivers on my Windows 10 laptop.

The blue curve shows the frequnecy response in OS mode ( Oversampling ) which is practically very good, showing only minor drop levels at the lower and upper bandwith. The red curve shows the frequency response in NOS mode ( non oversampling ) with the typical level drop at higher frequencies.

I have not measured the headphone outputs since I would use this unit as a DAC preamplifier.

The listening experience is interesting wise good as well. Two listeners that have visited me found the sound very pleasing. One liked it more as the implemented DAC ESS Sabre in my cocktail Audio X40 music server, while the other person did not hear any meaningful difference between both...

Yes, I calibrated the listening audience best possible. There has been a difference in output level of about 0.08dB

 

[dtaylo1066]

SINAD in the 70s is "SIN-full."

Lots of noisee spikes or distortion across many frequencies/harmonics.

I tend to like well executed tube amps, so I have concluded my brain, to some degree, must like second or third harmonic distortion, which some tests by Nelson Pass have proven to be true for some listeners.

But I would not buy a DAC that offers SINAD at this level. If you want an R2R dac for whatever reason, there are far better offerings, though they will cost you more, such as Harmony, Soekris or Denafrips.

If you think you might like second order distortion, build one of these simple kits and you will find out and save yourself the cost of a DAC.

H2 Harmonic Generator

This is Nelson's DIY H2 Harmonic Generator, Version 2. This kit consists of one PCB and two selected J113 JFETs, good for stereo operation. You can read about the project here: H2 Harmonic Generator DIY H2 Harmonic Generator Version 1 DIY H2 Harmonic Generator Version 2 After reading these...

diyaudiostore.com

 

[yourmom210]

I also made some basic measurement of the FiiO K11 R2R DAC using the SPDIF Coax input with a CD-Player attached to it that delivered several of the measuring signals. For low level signal I used dithered sine.

This is the linearity diagramm I have got at 1kHz 0, -30, -60 and -90dBFS. Lower was not included since I was using PCM 16bit / 44.1kHz signals.

View attachment 504323
Tendency is similar to the measurement at the 1st post here, except I get lower signal levels at -90dBFS.

A spectral analysis of a 1kHz sine at 0dBFS, 6dBV or 2Vrms at the RCA PRE OUT looks as follows.

View attachment 504325
Notice that I am using dBV scaling and not dBFS, meaning the output level is about 6dBV which correlates to 0dBFS. Like the measurement shown initially I also detected several harmonics above 1 kHz and several other distortion because of oversampling and undithered signals. Below 1kHz one sees some spikes that may be power supply related or intermodulation effects due to the higher level distortions over 1kHz.

I marked with orange the distortion levels that are defenitely audible, while the green line marks a typicall limit from which on for most listeners it gets inaudible.
The measured THD and THD+N are above the audible range and SNR is close to that what is considered as minimum.

To analyze influences of power supplies and other sources a measurement at -90dFS is of interest.

View attachment 504327
Again, be aware that the scaling is dBV for the level. I marked with a red point the level that should be reached when using a -90dBFS sine. That is pretty well reached, a bit lower as one can see in the linearity diagramm, but it is fine. One detects no spikes or disturbences above an below 1kHz, meaning that the power supply included works very good and causes no interferences.

Croos talk from left to right channel is actually very good as one can see from the next chart.

View attachment 504328
When applying 0dBFS or 6dBV on the left channel one measures a signal of -118dB lower at the right site. That is actually a very good value ensuring practically no crosstalk.

Last but not least the frequency response using the USB input and ASIOall drivers on my Windows 10 laptop.

View attachment 504330
The blue curve shows the frequnecy response in OS mode ( Oversampling ) which is practically very good, showing only minor drop levels at the lower and upper bandwith. The red curve shows the frequency response in NOS mode ( non oversampling ) with the typical level drop at higher frequencies.

I have not measured the headphone outputs since I would use this unit as a DAC preamplifier.

The listening experience is interesting wise good as well. Two listeners that have visited me found the sound very pleasing. One liked it more as the implemented DAC ESS Sabre in my cocktail Audio X40 music server, while the other person did not hear any meaningful difference between both...

Yes, I calibrated the listening audience best possible. There has been a difference in output level of about 0.08dB

Great work, thank you for sharing these results!

 

[cay-uwe]

SINAD in the 70s is "SIN-full."

These are statements that I find meaningless since no explanation is provided to justify it.

That is why I try to relate measurements to audible levels, like THD < 0,01% is for nearly all listeners inaudible, or SNR > 80dB is considered as very good, etc.

I measured a SNR = 84.29dB for the FiiO K11 R2R. Technically seen not the best, but still above the level that is considered to be critical.

The critical thing about SNR or SINAD is that one should not look to just the tested device, but more important to the actual setup, where one will interconnect it with preamplifiers, amplifiers, active loudspeakers, and other devices. When doing so, the other devices SNR or SINAD will lower the total SNR or SINAD of the setup.

Fortunately in this case, the FiiO K11 R2R can act as an preamplifier and in home setups one will only need an amplifier or active speakers to get a useful setup.

Let us assume one decides to use an tube amplifier, which usually have SNR in the range of 85dB, and that is already very good for that type of amplifier, then the total SNR in conjunction with the FiiO K11 R2r will be something around 81.62dB.

As one can see the resulting SNR of the setup is lower as the SNR of the individual devices and is very close to what is considered inaudible or very good.

If it is combined with modern Class-D amplifiers, having SNR beyond 100dB, then the resulting SNR will be dominated by the lowest value, in this case SNR ~ 84dB.

As more devices are interconnected, as lower the resulting SNR or SINAD will be and that is why it is important to have a look to it.

Lots of noisee spikes or distortion across many frequencies/harmonics.

That is true and I agree that it does not look very nice, also considering that the THD and THD+N are above of what is considered to be audible. In such a case I find it important to get listening impressions to check how relevant that is. As you can read, no one of the listeners complaint about distorted sound, and one even liked it better as the cocktail Audio X40 I have, which measures far better as the FiiO K11 R2R. The other listener did not even hear a difference ...

Just for the readers undestanding, I bought this unit because I got it very cheap and tought I will have a bit of fun with it

 

[dtaylo1066]

These are statements that I find meaningless since no explanation is provided to justify it.

That is why I try to relate measurements to audible levels, like THD < 0,01% is for nearly all listeners inaudible, or SNR > 80dB is considered as very good, etc.

I measured a SNR = 84.29dB for the FiiO K11 R2R. Technically seen not the best, but still above the level that is considered to be critical.

The critical thing about SNR or SINAD is that one should not look to just the tested device, but more important to the actual setup, where one will interconnect it with preamplifiers, amplifiers, active loudspeakers, and other devices. When doing so, the other devices SNR or SINAD will lower the total SNR or SINAD of the setup.

Fortunately in this case, the FiiO K11 R2R can act as an preamplifier and in home setups one will only need an amplifier or active speakers to get a useful setup.

Let us assume one decides to use an tube amplifier, which usually have SNR in the range of 85dB, and that is already very good for that type of amplifier, then the total SNR in conjunction with the FiiO K11 R2r will be something around 81.62dB.

As one can see the resulting SNR of the setup is lower as the SNR of the individual devices and is very close to what is considered inaudible or very good.

If it is combined with modern Class-D amplifiers, having SNR beyond 100dB, then the resulting SNR will be dominated by the lowest value, in this case SNR ~ 84dB.

As more devices are interconnected, as lower the resulting SNR or SINAD will be and that is why it is important to have a look to it.



That is true and I agree that it does not look very nice, also considering that the THD and THD+N are above of what is considered to be audible. In such a case I find it important to get listening impressions to check how relevant that is. As you can read, no one of the listeners complaint about distorted sound, and one even liked it better as the cocktail Audio X40 I have, which measures far better as the FiiO K11 R2R. The other listener did not even hear a difference ...

Just for the readers undestanding, I bought this unit because I got it very cheap and tought I will have a bit of fun with it

I respect your opinion and appreciate your measurements. In my audio chain it would be the lowest SNR component, as I have a modern class D Hypex amp and a DAC with 115 SINAD. Can or would I hear any difference. Perhaps not. And many claim the FiiO to be a very pleasant to listen to DAC, if not stellar, as is the belief by some of R2R DACs. Also, as I have indicated previusly on this forum, all this high SNR/SINAD winds up coming out of speakers with higher levels of distortion and frequency ups and downs than what feeds them. And then there is the room they are playing in. And the vagaries of one's hearing, certainly mine at age 69.

People should buy and listen to what they like. I would have liked the folks at FiiO to put more into the design to increase that SNR/SNAD number, as others have done that with R2R DACs.

At the price point, if one wants to see what the buzz about R2R is, this would seem a good candidate, given it has many quality features.

In addition, I was being a bit tongue in cheek with the juxtaposition of SINAD and SIN Full.

Most important is for you to enjoy your component.

 

[cay-uwe]

all this high SNR/SINAD winds up coming out of speakers with higher levels of distortion and frequency ups and downs than what feeds them.

I am in the loudspeaker development / production business and that is absolutely true. Even very good loudspeaker have distortions that are 10 - 100 times higher as the FiiO K11 R2R. Another reason why I look a bit relaxed to all that extremely good numbers one gets today even from cheap electronic devices.