• WANTED: Happy members who like to discuss audio and other topics related to our interest. Desire to learn and share knowledge of science required as is 20 years of participation in forums (not all true). Come here to have fun, be ready to be teased and not take online life too seriously. We now measure and review equipment for free! Click here for details.

Review and Measurements of HifiBerry DAC+ Pro XLR

amirm

Founder/Admin
Staff Member
CFO (Chief Fun Officer)
Joined
Feb 13, 2016
Messages
20,594
Likes
25,379
Location
Seattle Area
#1
This is a review and detailed measurements of the HifiBerry DAC+ Pro XLR. It was kindly purchased by a member and drop shipped to me. It costs USD $59.90 which is quite cheap for balanced output. I think the cheapest stand-alone DAC with XLR output that I have tested was US $150 (Massdrop Grace SDAC Balanced). Even including the cost of the Raspberry Pi, you are less than $100 here and you get streaming functionality to boot.

Here is what it looks like:

HIFIBERRY DAC+ PRO XLR Audio review.jpg

The giant XLR connectors dwarf the rest of the board. They are sturdy though and had no trouble supporting the weight of my XLR cables although the rest of the unit as you can imagine, is quite unstable. You would need to weigh it down somehow to keep it from following the XLR cable to the ground or wherever it goes. :)

The Pro designation in the name means that it has dual clock oscillators instead of using the poor quality clock that the Pi generates (they are the two rectangular parts near center right). XLR means exactly that. Instead of the RCA jacks in regular HifiBerry DAC+ Pro which I just reviewed, you get balanced XLR outputs. This is nice because it sharply reduces the possibility of nasty ground loops which are hard to deal with it when they bite you.

The heart of the device is the TI PCM5242 DAC chip with balanced outputs. Its rated distortion and noise is 94 dB (SINAD). Let's see if the actual implementation gets there.

DAC Audio Measurements
Here is our dashboard view:
HIFIBERRY DAC+ PRO XLR Audio Measurements.png


Output voltage of 4.2 volts is slightly higher than nominal value of 4 volts we like to see which is fine. Alas, total SINAD (signal over noise and distortion) falls short of TI spec of 94 dB. But maybe TI means distortion alone in which case, we are there with third harmonic peeking its head to that level in the FFT graph above.

The measured SINAD places the XLR version just behind the RCA version:

Best Audio Dacs Measured and Reviewed 2019.png


That is not a significant factor though. Compared to Grace SDAC, we have a shortfall of 5 to 6 dB.

Jitter performance is good and similar to RCA version:
HIFIBERRY DAC+ PRO XLR Jitter Audio Measurements.png


Multitone performance is again similar but perhaps a hair worse:
HIFIBERRY DAC+ PRO XLR Multitone Audio Measurements.png


Filter response is identical:
HIFIBERRY DAC+ PRO XLR White Noise Filter Response Audio Measurements.png


Conclusions
The HifiBerry DAC+ Pro XLR has performance that is essentially identical to RCA version of it. You pay $20 more which to me is a reasonable insurance to pay against ground loops. Sadly the inclusion of XLR doesn't bring better fidelity.

Overall, based on low cost of the solution I am going to recommend the HififBerry DAC+ Pro XLR even though I wished it had better measured performance.

------------
As always, questions, comments, recommendations, etc. are welcome.

Have to do a bunch of system upgrades and that is costing money. So appreciate any donations toward that using:

https://www.audiosciencereview.com/forum/index.php?threads/how-to-support-audio-science-review.8150/
 

msmucr

New Member
Joined
Aug 16, 2019
Messages
4
Likes
14
Location
Prague, Czech Republic
#2
That lower stopband attenuation initially looked to me like overload at internal digital path. Quite a few DAC chips has suffers from that, when it's being fed by high level white noise. Sometimes just small attenuation of test signal by 2-3 dB will significantly improve attenuation of its reconstruction filter. Similarly such lack of internal headroom is indicated by clipped overshoots of reconstructed square wave at DAC output, especially with minimum phase filters.

But in this case, the measured response apparently corresponds to TI chip datasheet figures.
http://www.ti.com/lit/ds/symlink/pcm5121.pdf (p. 9) So DAC performs exactly as initialized.
However it also supports longer FIR filter with higher latency and steepness, rated to 100dB of SB attenuation. Chip at this HAT is in software mode and initialized as via I2C bus from RPi by respective Linux kernel module.
Just quickly looking to few relevant pieces of code..
https://github.com/raspberrypi/linu...486cc335224882/sound/soc/codecs/pcm512x.h#L62 (register is there)
https://github.com/raspberrypi/linux/blob/rpi-4.19.y/sound/soc/bcm/hifiberry_dacplus.c#L145 (we can add modification to init there)

So for anyone bit geeky and interested, it could be quite easy to mod by adding of line to setup respective register values and recompile the module.

I had RCA version of the HAT, but gave it as a gift to my ex GF as part of the "Spotify Box", so can't really try to do that ATM. Getting that box back for playing would definitely involve some unrelated talking ;)

Anyway, thanks Amir for the measurements.

Michal
 

Joachim Herbert

Member
Patreon Donor
Joined
Jan 20, 2019
Messages
47
Likes
63
#3
Now this might be a challenge for @IVX. A high performance shield for the raspberry pi. While there are offerings they seem to be based on lowend chips. Just imagine a competently designed board based on one of the newer ess chips plus custom housing. Plus spdif out, just in case. Streamer of my dreams.
 

somebodyelse

Active Member
Joined
Dec 5, 2018
Messages
226
Likes
159
#4
Now this might be a challenge for @IVX. A high performance shield for the raspberry pi. While there are offerings they seem to be based on lowend chips. Just imagine a competently designed board based on one of the newer ess chips plus custom housing. Plus spdif out, just in case. Streamer of my dreams.
The Katana and ApplePi are still in the top quarter of all DACs tested. I don't think they have the SPDIF output though.
 

Joachim Herbert

Member
Patreon Donor
Joined
Jan 20, 2019
Messages
47
Likes
63
#5
@somebodyelse Both are unbalanced. Do not like the optimized for sound approach the katana take, plus I'd rather have integrated circuits instead of discrete op amps. I obviously am no audiophile.
 
Joined
Aug 20, 2019
Messages
4
Likes
5
#6
@amirm How did you power the DAC during the test? Through the PI or directly to the DAC?
I have experienced differences in the sound depending on where I connect, and with what power supply (Cheap noisy 5V power vs a "more expensive" IFI iPower supply) .
 
Joined
Mar 18, 2019
Messages
39
Likes
39
#8
I can understand you @Joachim Herbert . I also like ICs better. There are really good measuring devices on ASR, that just use one or two Operational Amp IC per output.

And another thing, what I don't like from their website Allos website:
Katana gives you the choice between 2 stages:
"Pure THD+N":
At -112.9 Thd+n 1Khz A weighted, this stage will satisfy those that want no distortion in music. Board sounds very good.
"Sound quality" stage: 100.5 Thd+n a weighted 1Khz, this board has been optimized for sound quality. While sacrificing some thd+n numbers we found the subjective experience to be, in a word, closer to analog sound.
"Sound quality stage" smells like...

But they show the THD+N and with Pure stage it is a really good DAC for a RPi, as measurements on ASR show.
 
Last edited:

amirm

Founder/Admin
Staff Member
CFO (Chief Fun Officer)
Joined
Feb 13, 2016
Messages
20,594
Likes
25,379
Location
Seattle Area
#9
@amirm How did you power the DAC during the test? Through the PI or directly to the DAC?
I have experienced differences in the sound depending on where I connect, and with what power supply (Cheap noisy 5V power vs a "more expensive" IFI iPower supply) .
Through Pi. Since measurements match chip company data with lab supplies it should not make a difference.
 

somebodyelse

Active Member
Joined
Dec 5, 2018
Messages
226
Likes
159
#10
But in this case, the measured response apparently corresponds to TI chip datasheet figures.
http://www.ti.com/lit/ds/symlink/pcm5121.pdf (p. 9) So DAC performs exactly as initialized.
However it also supports longer FIR filter with higher latency and steepness, rated to 100dB of SB attenuation. Chip at this HAT is in software mode and initialized as via I2C bus from RPi by respective Linux kernel module.
Just quickly looking to few relevant pieces of code..
https://github.com/raspberrypi/linux/blob/7ee8d72eeb56b85aaad25f368b48https://github.com/raspberrypi/linux/blob/7ee8d72eeb56b85aaad25f368b486cc335224882/sound/soc/codecs/pcm512x.c6cc335224882/sound/soc/codecs/pcm512x.h#L62 (register is there)
https://github.com/raspberrypi/linux/blob/rpi-4.19.y/sound/soc/bcm/hifiberry_dacplus.c#L145 (we can add modification to init there)
It may already be exposed - it looks like there are controls defined for most of the adjustable things in the base pcm512x driver, so unless it's overridden elsewhere (mixer quirk for hifiberry?) you can probably already change it at runtime.
https://github.com/raspberrypi/linu...86cc335224882/sound/soc/codecs/pcm512x.c#L463
 

Jimster480

Major Contributor
Patreon Donor
Joined
Jan 26, 2018
Messages
1,083
Likes
577
Location
Miami
#11
I don't see the purpose in this product considering the poor performance... I guess if you are just driving speakers with it for outside this can work fine? Otherwise Its not competitive in any manner when you can spend a bit more for so much better performance.
 

mk1classic

New Member
Joined
Aug 20, 2019
Messages
4
Likes
5
#12
I don't see the purpose in this product considering the poor performance... I guess if you are just driving speakers with it for outside this can work fine? Otherwise Its not competitive in any manner when you can spend a bit more for so much better performance.
It's one of very few Raspberry Pi HAT DACs with XLR output, so it's kind of unique.
 

Similar threads

Top Bottom