• WANTED: Happy members who like to discuss audio and other topics related to our interest. Desire to learn and share knowledge of science required. There are many reviews of audio hardware and expert members to help answer your questions. Click here to have your audio equipment measured for free!

Review and Measurements of Benchmark DAC3

amirm

Founder/Admin
Staff Member
CFO (Chief Fun Officer)
Joined
Feb 13, 2016
Messages
45,830
Likes
255,915
Location
Seattle Area
This is a review and detailed measurements of Benchmark DAC3 and headphone amplifier. The unit is on kind loan from member Dallasjustice and is highly anticipated evaluation by myself and I am sure many of you. As of this writing the DAC3 retails for USD $2195 from Benchmark site. Others discount it to $2,000 plus shipping. So not cheap by budget standards. But a huge bargain compared to high-end DAC which start at $10K and go up.

For way of comparison, I am also measuring the RME ADI-2 Pro using my Audio Precision APx555 analyzer. As you see below, the two units are pretty much the same width and height.

Benchmark DAC3 and RME ADI-2 Pro Review and Measurements.jpg


Depth-wise though, the Benchmark is almost twice as deep. Still, both comfortably fit on the desk.

From user interface point of view, the DAC3 has clear LEDs which i like but it still doesn't hold a candle to RME's nice displays with full spectrum view and such as you see in the picture. This is very helpful in measurement because it is a nice confirmation of what is being sent to it.

All the settings in RME ADI-2 Pro (and by association in RME ADI-2 DAC) are "soft" in that you can change them in menus. The Benchmark has far fewer settings and what it has requires opening the case and moving jumpers. While that is a more purist implementation, clearly not as user friendly. What is NOT user friendly is the interface in RME. In both ADC+DAC combo that I am testing here and its DAC-only version, they are maddening to navigate. If you tolerate it, you have many cool options including parametric (?) equalization.

I tested the RME ADI-2 Pro using its ASIO drivers. For benchmark I used ASIO4ALL after it plugged and played in Windows as a normal sound card.

The volume control on DAC3 changes the output balanced connections (in addition to headphones I assume). On the RME by default, it only changes the headphone output. I am sure there is a way to change that in the menus but I am not brave enough to figure it out. :)

The DAC3 volume acted interesting in that small changes would make no difference and then all of a sudden the value would change. Not sure what explains this. Analog pot driving a digital attenuation? It has a max and min so it is not a rotary encoder.

Testing these two devices is quite daunting given no less than three outputs: unbalanced, balanced and headphone. For this review, based on interest of its owner, I am only focusing on balanced output performance. I aim to at least cover the headphone performance before having to return the unit in the next day or two.

EDIT: A second loaned unit from Benchmark performed even better. Please see: https://www.audiosciencereview.com/...easurements-of-benchmark-dac3.3545/post-91447

Measurements

Please note that all measurements are made with USB input of the DACs. Two posts down there is a comparison of Toslink to USB.

I always like to match levels of DACs where possible to make a fair comparison. Alas, even though Benchmark documents that the unit by default outputs 24 dBu (12 volts RMS), it puts out copious more output than that. I had to use a "pad jumper" inside the unit to dial in -10 dB of attenuation to match that of RME ADI-2 which indeed maxes out at 24 dBu. In this dashboard view of DAC3, I have one channel with default output and the other with -10 dB:

Benchmark DAC3 Max Volume Dashboard Measurement.png


As you see, the non-attenuated channel outputs nearly 20 volts RMS, corresponding to whopping +-28 volts or peak to peek! :eek: With that kind of output swing if you are bored, you can hook up the DAC3 to drive your Christmas tree lights! :D

Dallasjustice uses the much higher output to good results in room equalization. The most effective way to deal with troughs in room response is to simply bring down the levels of all the other frequencies. This gives you a flat response but now you have much less gain to drive your speakers. Having so much headroom and output capacity in the DAC3 helps mitigate this.

For the rest of these tests, I ran with the attenuation at -10 dB to match the RME as mentioned. Note that performance without this attenuation is different (sometimes better, sometimes worse). Such is life when we try to make apples vs apples comparison.

Let's see how the two do in jitter and noise department:
Benchmark DAC3 Jitter and Noise Measurements.png


Starting with RME ADI-2 Pro, we see the same problem with random low frequency noise around our main tone at 12 kHz, resulting in that "skirt" around it. RME has since upgraded the clock both the Pro and DAC version so that problem is no longer there. I confirmed that in the DAC review but not in the Pro as I don't have that version as you see in that review:

index.php



So that part is a "non-problem." The RME has slightly higher noise floor still. But the DAC3 has those two odd peaks. They are not symmetrical in level so likely not jitter. They are proportional to the position of the volume knob so the problem is prior to that stage. This is unexpected in this class product and from a company like Benchmark.

The news becomes quite positive for Benchmark DAC3 as we proceed with other tests such as intermodulation (SMPTE) relative to level:
Benchmark DAC3 IMD distortion Measurements.png


The DAC3 has a clear advantage of about 6 dB lower distortion+noise. It starts to clip sooner than RME but still at vanishingly small level of -110 dB.

THD+N measurements relative to frequency show similar advantage for DAC3:
Benchmark DAC3 THD+N distortion Measurements.png


Looking at the spectrum of that at 1 kHz, we get this:
Benchmark DAC3 RME ADI-2 Pro 1 kHz noise and distortion Measurements.png


Applying perceptual modeling to this is hard visually. They both have the same second harmonic distortion but from there on, one gets ahead and then the other. Over time I hope to create a mathematical model for this to generate better figure than THD+N which summs the power of all of those.

Slightly higher noise level is apparent here as with our jitter and noise analysis.

Switching to our favorite test, Linearity, we get this:
Benchmark DAC3 Linearity Measurements.png


The DAC3 produces output essentially the same as the analyzer itself. I have put in the "19 bit" marker for historical reason (0.1 dB of variation) but really, this is as good as we can measure. The RME ADI-2 Pro lags a bit here but we know that the DAC version is superior, producing similar output to DAC3. Again from my previous review:

index.php


As a side-note, see how the measurements of the RME ADI-2 is very similar using my old analyzer (AP 2522) versus new (APx555). On APx555 I have increased the resolution some so that we can see finer variations but if I use the coarser ones used in the 2522 the results match very closely. So some progress in finalizing my APx555 settings to make them comparable to the older analyzer both as a sanity check and ease of comparison. Similar progress was made in IMD measurements I post above.

I think I will stop here as to not overwhelm you with too much data. As mentioned, there will be more to come.

For now, if you have any questions and concerns please raise them quickly as I will be returning DAC3 soon.

EDIT: headphone power and output impedance measures in this post and the one after it: https://www.audiosciencereview.com/...easurements-of-benchmark-dac3.3545/post-86011

Conclusions
The Benchmark DAC3 as expected is a state-of-the-art digital to analog converter. Other than one set of noise spikes in jitter test, the rest of the measurements show exceptional performance. No glaring faults are seen at all. Its higher output level can be useful in room EQ applications to boot.

So of course the DAC3 goes on my recommended list.

Considering the price though, the RME ADI-2 DAC retails at less than half the DAC3 price. While the DAC3 probably beats it a bit in objective measurements, I can't really justify the additional cost of it over that unit. So if money is no object and you have no use for the additional features of ADI-2 DAC, by all means get the DAC3. Otherwise, the RME ADI-2 DAC remains my pick for "expensive" DACs.

-------------

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

If you like this review, please consider donating funds for these types of hardware purchases using Patreon (https://www.patreon.com/audiosciencereview), or upgrading your membership here though Paypal (https://audiosciencereview.com/foru...eview-and-measurements.2164/page-3#post-59054).
 
Last edited:
Here is the unbalanced tests from Benchmark DAC3. It is not the full suite due to shortness of time but should give you a flavor of what is going on.

First, dashboard:
Benchmark DAC3 Unblanced RCA Dashboard Measurement.png


As with balanced the DAC3 has healthy level of drive even out of its RCA plugs at over 3.1 volts. So it should have no difficulty driving any unbalanced inputs on your pre-amp or amplifier.

Next, let's look at Jitter and Noise:
Benchmark DAC3 Jitter and Noise Measurement.png


Here I am comparing the two outputs: balanced vs unbalanced. Same distortion spikes exist over USB. Noise floor is nicely lowered together with the lower output so dynamic range more or less remains the same. I wonder if they attenuate and convert the balanced output to unbalanced and hence the reduction in noise level.

THD+N likewise looks like this:

Benchmark DAC3 THD+N Measurement.png


Here, we do see slightly degraded performance due to higher noise levels. So maybe the dynamic range is a bit lower. :)

Now here is where it gets interesting: linearity:

Benchmark DAC3 vs Topping D50 Linearity.png

Unlike balanced output, the unbalanced response is considerably worse with one channel worse than the other.

As a way of reference, I put the Topping D50 through the same test and as before, it has textbook output to -120 dB.

Since these are new Linearity tests on my new Audio Precision APx555, I thought I do a sanity check and look at the waveforms at -120 dB with the analyzer filtering in place:

Benchmark DAC3 vs Topping D50 -120 dB sine wave measurement.png


While you can't see the linearity error itself, you can easily see the varying levels in the output of Benchmark DAC3 on the right. Clearly one or the other channel output is in error and that is what the Linearity measurement showed. The Topping on the other hand, most of the time produces identical output (once in a while one channel would be a bit lower).

This is more remarkable than it seems since the output of the Topping D50 is lower (1.9 volts RMS vs 3.1 on DAC3). All else being equal, it is harder to get a clean lower voltage output than higher (with respect to noise).

This is a very surprising result. I will leave the door slightly ajar on instrumentation issues due to newness of the linearity test on the APx555 which I have developed. But the arrows certainly point to less than optimal implementation of unbalanced output from DAC3.

Summary
While the benchmark DAC3 generally measures well on its unbalanced output, reproduction of low level signals is not very accurate. There certainly is no excuse for a $250 DAC to beat it in such tests as linearity.

So sadly, I cannot recommend the Benchmark DAC3 for unbalanced output.

As always, comments, correction and criticism are welcome.
 

Attachments

  • Benchmark DAC3 vs Topping D50 Linearity.png
    Benchmark DAC3 vs Topping D50 Linearity.png
    49 KB · Views: 3,215
Last edited:
Was the jitter test done with spdif? I have a feeling those spikes on the dac3 are from the src4392.
 
I am a bit cofused. Is this a comparison of the Benchmark and ADI or two ADI models? Perhaps it is not finished and the answer will be revealed.
 
Was the jitter test done with spdif? I have a feeling those spikes on the dac3 are from the src4392.
Sorry. Forgot to mention that all measurements were with USB input.
 
I am a bit cofused. Is this a comparison of the Benchmark and ADI or two ADI models? Perhaps it is not finished and the answer will be revealed.
The measurements conducted were side by side with adi2 pro. I no longer have the dac only version of RME so I referenced my older review for that. Sorry for the confusion.
 
Can you do a THD measurement for the headphone out with a dummy and real headphone load, as you already did for a few other devices?
 
TOSLINK...

This is same procedure from me: Will you show any Toslink comparisons, especially jitter related?
 
Alas, even though Benchmark documents that the unit by default outputs 24 dBu (12 volts RMS), it puts out copious more output than that.

Maybe not exactly...

Press the DIM/MUTE button for 3 seconds, It will go into "bypass the volume knob" mode for calibrated output.

(On my DAC2, the volume knob will turn itself to the calibrated output level, at about 8/10 dots on the knob and lights a yellow LED. DAC3 might operate differently, it may not move the knob)

Otherwise, as I read it, if you turn the knob full on you can get 27.5dBu.

1529878830216.png


Page 17 of the manual:

1529878936475.png
 
Last edited:
My understanding ASIO4ALL is more or less a hack. Has it been established to be a reliable ASIO emulation interface for this application? I'm pretty sure native high-quality ASIO driver must be available from Benchmark, why not just use it to avoid any possibility ASIO4ALL somehow affects the results?
 
My understanding ASIO4ALL is more or less a hack. Has it been established to be a reliable ASIO emulation interface for this application? I'm pretty sure native high-quality ASIO driver must be available from Benchmark, why not just use it to avoid any possibility ASIO4ALL somehow affects the results?
Generally asio4all is hit or miss in my experience.
 
My understanding ASIO4ALL is more or less a hack. Has it been established to be a reliable ASIO emulation interface for this application? I'm pretty sure native high-quality ASIO driver must be available from Benchmark, why not just use it to avoid any possibility ASIO4ALL somehow affects the results?
Indeed, they have an ASIO driver at benchmarkmedia.com/drivers
 
Nice review @amirm! I hope you post some pictures of the DAC3 insides too. We enjoyed our visit in your kingdom. You and Lynn made my family feel at home. Here is a shot of the boys and I hiking in the snow near Paradise, Mt. Rainier.
85A3E0F0-150B-43BF-8FF5-DE84A693C910.jpeg
 
Back
Top Bottom