Cayin RU6 R2R Dongle Review (and Measurements: shit happens)

I don't think it sounded better at lower volumes using the analog attenuation because this is how I used it with various headphones. I set the volume around 35 for Grados. I didn't use digital volume control at all.

WIth respect to measurements, here is the -60db attenuation using the analog volume control (instead of digital) compared to using -60dbFS digital attenuation with the volume fully cranked up:

Cayin RU6
-60db using analog volume control
64k FFT, 90kHz bandwidth
SE output


Cayin RU6
-60dbFS with volume maxxed out to 100
64k FFT, 90kHz bandwidth
SE output


This result is better, but limited by noise. Note that to achieve -60db, I needed to adjust the volume to 14/100, which is about 20 steps lower than what I would listen to most headphones with.

Notes:

• SE output
• 90kHz bandwidth
• FFT size is 64kHz (read here on FFT gain and how it plays tricks).

 

Here are some linearity measurements. This seems consistent with the -70db, -80db, -90db FFTs earlier. Basically as the signal gets lower, the RU6 outputs a signal lower than it's supposed to. I do wonder if the AP is having trouble locking on to the signal from -60dbFS and up. The 1kHz signal from the RU6 isn't steady and jumps around (massive jitter). It's also possible that the resistor ladder is inaccurate. This would kind of a like "jitter" in the volume domain instead of time domain.

Cayin RU6
Linearity (Relative Output Measured vs Intended Output)
x3 measurements superimposed
Bandpass Level Sweep (1/24 octave selectivity bandpass)
Vol set to 100


Cayin RU6
Linearity (Relative Output Measured vs Intended Output)
x3 measurements superimposed
Bandpass Level Sweep (1/24 octave selectivity bandpass)
Vol set to 40 (simulating real world use)


For comparison, this what a good linearity measurement should look like.

 

This breed is well known and one that I've run into before many years ago when discrete resistor DACs were in their infancy: discrete ladders with cheap resistors. There is a reason why Soekris goes with better than 0.05% resistors in their entry level going up to 0.02% and 0.01%.

As I've said. The sound is unique and different from every other dongle out there. This in itself will attract a lot of people to it, like a circus sideshow.

Some possibilities:

1. This unit is defective or the wrong parts (even more horrible tolerance resistors got fed into the SMT PCB assembly machine). It happens.
2. I've heard lot of stuff before including discrete ladder DACs with low tolerance resistors which I never liked. I recognize this sound.
3. The sound is unique and different from every other dongle out there. This in itself will attract a lot of people to it, like a circus sideshow.
   
4. One man's shit is another man's treasure. (Hey, I like Grados).
5. I have a larger burden of responsibility to warning people away from the next FOTM.

 

This wasn't the loaner unit. It was more or less brand new unit purchased by a member, used for no more than a day, and is now sitting on my desk. Now we do have the loaner, which I am sure will make it to me eventually.

I'll take as an AIO measurements at 0, -10, and -20dBU at various loads.

 

The jumpiness could also be the L and R signals being multiplexed into a single R2R ladder. This explains the lag between the L and R signals and jitter.



I can't get a stable frequency reading. LOL, this is visible jitter - gross errors.

 

Use as a AIO driving headphones/IEMS.

Cayin RU6
0dbu into 300-ohms
Adjust via volume
DAC output set to -1dbFS (benefit of the doubt to avoid anomalous behavior at -0dFS)


Cayin RU6
-10dbu into 300-ohms
Adjust via volume
DAC output set to -1dbFS (benefit of the doubt to avoid anomalous behavior at -0dFS)


Cayin RU6
-10dbu into 32-ohms
Adjust via volume
DAC output set to -1dbFS (benefit of the doubt to avoid anomalous behavior at -0dFS)


Cayin RU6
-20dbu into 32-ohms
Adjust via volume
DAC output set to -1dbFS (benefit of the doubt to avoid anomalous behavior at -0dFS)


Finally, I noted the channel disparity increases at lower impedance loads.

 

Here is OS mode. I literally have no clue what the oversampling rate is supposed to be. 10x?

Cayin RU6
1kHz -100db
44.1kHz sampling rate
90kHz bandwidth linear scale


500kHz bandwidth linear scale


1MHz bandwidth

 

There could be something wrong with this. One channel seems to struggle under load. The more difficult to load, the worse it gets.

Cayin RU7
OS 44.1kHz
37Hz into 300-ohms, -10dBU output
-1dbFS, output set by volume



Cayin RU7
OS 44.1kHz
37Hz into 32-ohms, -10dBU output
-1dbFS, output set by volume

 

Here is the lag in NOS
1kHz square wave

 

Cayin RU6
Frequency Response (NOS)
-1dbFS, volume set to 68
300-ohm load