• 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.

Benchmark HPA4 based on THX AAA Amplifier

RayDunzl

Major Contributor
Central Scrutinizer
Joined
Mar 9, 2016
Messages
6,802
Likes
2,899
Location
Riverview, FL
#41
Local Experiment:

I have a DAC2, the jumpers are factory set, no attenuation. The amps have 26.4dB gain.
I set my preamp wide open (it might add a little gain itself), and attenuated the DAC2 to about a 70dB SPL in-room listening level.

The knob was about 9 o'clock (max is at 3 o'clock)
Measured about 40dB attenuation.

That should lop 7 bits off the digital signal, leaving 25 bits to work with in the DAC.

I'm pressed to see where that would create a problem.
 

Rthomas

Member
Patreon Donor
Joined
Nov 8, 2017
Messages
27
Likes
11
#42
I Pre-ordered one of these yesterday. Looking forward to trying it out as a Pre-amp replacement in my system. If it works as expected it will have saved me a bundle on the "High End" Pre I was intending to pickup.

In my view the benefit of using this device after a DAC like the DAC3 DX (which I own) is that it can address gain mismatch issues as well as digital VOL resolution related losses which can occur when not running the VOL knob at a high enough position when trying to use the DAC3 to directly drive a Non-Benchmark Amp.

In my setup while using the highest Analog Attenuator jumpers (-20db) there is still too much gain in the system (26db) to achieve a normal listening level. In my setup the volume knob of the DAC3 needs to sit at roughly 9'oclock for a normal listening level but at that position there is not enough voltage output yet to drive my amp to its full potential. Anything more than 10 o'clock on the VOL knob of the DAC3 and I start to approach Rave/Party Mode which is great if that is the mode you are trying to be in at all times.

Benchmark says the VOL knob should be at 12 o'clock or so with the -20db jumpers in place to achieve roughly 1.95v output. When I use the VOL knob at 9 o'clock the voltage is under what I need and its audibly recognizable. According to Pass Labs I should be feeding my Amp (XA60.8) at least 1.8vrms to take advantage of the Dr Jeckyl & Mr Hyde like behavior these XA Amps are capable of when they switch into A/AB Mode. The Spec sheet says only 1.1 vrms is needed for this Amp but that is based on only outputting the 60w which the Amp is advertised to supply into 8ohm. To take advantage of the additional output power (as seen in the Stereophile review) more than 1.1vrms is required.

Anyway, I'm looking forward to giving this unit a try. If its as transparent as I expect it to be, based on Benchmarks history of such, it will be a real shocker to the "High End" market. I was very impressed with the sound of the DAC3 in my system so I didn't hesitate to be an early adopter and try this HPA4 also.
If you're in the USA please send your DAC3 to Amir for measurements. A kind request on behalf of all of us outside the USA. :D
 

RayDunzl

Major Contributor
Central Scrutinizer
Joined
Mar 9, 2016
Messages
6,802
Likes
2,899
Location
Riverview, FL
#43

cjf

Active Member
Joined
Apr 20, 2016
Messages
129
Likes
21
Location
Somewhere
#44
Local Experiment:

I have a DAC2, the jumpers are factory set, no attenuation. The amps have 26.4dB gain.
I set my preamp wide open (it might add a little gain itself), and attenuated the DAC2 to about a 70dB SPL in-room listening level.

The knob was about 9 o'clock (max is at 3 o'clock)
Measured about 40dB attenuation.

That should lop 7 bits off the digital signal, leaving 25 bits to work with in the DAC.

I'm pressed to see where that would create a problem.
Now try that with the DAC Direct to your Amp using just the DAC's Digital VOL and Output Gain capabilities. In doing so, do you still enjoy all your typical listening scenerios without that external PRE in the Chain? Does it sound the same, better or worse?

The experiment above is still using your external PRE's voltage/drive and Analog VOL capabilites to push your amplifiers is it not? Just burying its VOL knob addresses any low voltage issue since that burden is no longer on the DAC.
 

RayDunzl

Major Contributor
Central Scrutinizer
Joined
Mar 9, 2016
Messages
6,802
Likes
2,899
Location
Riverview, FL
#45
(from memory)

I'm confused because (as I read your post), when you turn down the DAC (reduce its output voltage) its voltage isn't high enough to drive the amp into its Class A/B range, but you want to add the HPA4 to do what? Boost the voltage without increasing the volume? Maybe your post (or my reading of it) just isn't clear.

The second part of your post addressed losing resolution due to digital attenuation (I think).

As for losing bits, my experiment was to attenuate the DAC, maybe 20dB more that you do in the digital attenuation, since I've not set attenuation with the jumpers, to see how it sounds.

Seemed ok to me, nothing obviously wrong.

I'll repeat with direct XLR to the amps... I suspect I'll have a little less attenuation at the DAC, since the gain of the pre will be gone, and have 26 or maybe 27 of its internal 32 bits still active.
 

RayDunzl

Major Contributor
Central Scrutinizer
Joined
Mar 9, 2016
Messages
6,802
Likes
2,899
Location
Riverview, FL
#46
Made the change, preamp out, DAC2 using XLR to amplifiers, volume knob is at 10 o'clock, and nothing seems diluted to me except volume.

The DAC2 volume knob has marks corresponding to a clock face like the DAC3.

In-room is 76 dB Lzeq. A moderate afternoon level.

upload_2018-4-27_18-14-4.png


The difference in output of the DAC between nine and ten o'clock is about 10dB.

The preamp spec shows 12dB gain, so, that's about right (just under 9 o'clock with pre wide open, 10 o'clock without, for similar in-room levels).

I added the preamp for flexibility - various sources, two destinations, experiments, but I didn't get it to improve on the DAC direct.

Your mileage may vary, I get to be wrong, I don't have the Golden Ears, but I just don't see where you'll gain (bad pun) much over the DAC3 direct to Amp setup for the cost.

If you want one, go for it.

That's all...

---

Belated Opinion:

Assuming you're buying the HPA4 for use as a preamp and not for its headphone jack, I'd probably spend the $3000 on an AHB2 instead.
 
Last edited:

cjf

Active Member
Joined
Apr 20, 2016
Messages
129
Likes
21
Location
Somewhere
#47
What I would like to know but couldn't squeeze an answer from Benchmark when asked is, what kind of voltage do these DAC's output, RMS, when the VOL knob is sitting at 9 o'clock or so while also using the -20db jumpers in a DAC direct to amp scenario. Knowing that would tell the true tale of whether or not the DAC alone should be able to drive a given amp into all loads/outputs its capable of.

I agree that in this particular case using these DAC's (model 2 & 3 w/32bit layout) Digital Bit chopping shouldn't be an issue but the downside of not having some kind of digital readout of what the actual attenuation level is at a given position leaves one guessing if they have crossed the threshold or not of allowable digital vol reduction to need to worry about Bit chopping. An example would be, if I see I'm only sitting at -45dbfs its probably safe to assume no Bits are being chopped..yet. But, at that level, is the DAC supplying the minimum voltage required to drive the amp appropriately.

I'm kind of assuming here but I would like to think a proper external PreAmp is always supplying at least 2v output via XLR to feed an amp no matter what its VOL position is. Then as its turned up more and more, that number increase even further up to whatever (4-36v). So it seems at least. Low volume levels with my external PRE seems to sound much more powerful then it does when going DAC direct which, in my case at least, forces me to operate much lower on the VOL scale. Given the fact that the Benchmark (or most any DAC without proper vol control) requires one to be at almost 3/4 the way up to provide just under 2v output my assumption is that below those levels the voltage output is way down there on the scale still.
 

RayDunzl

Major Contributor
Central Scrutinizer
Joined
Mar 9, 2016
Messages
6,802
Likes
2,899
Location
Riverview, FL
#48
I'm kind of assuming here but I would like to think a proper external PreAmp is always supplying at least 2v output via XLR to feed an amp no matter what its VOL position is.
Totally wrong.

Volume control limits the voltage sent to the amplifier.

Same signal displayed two ways.

Top: Attenuation in Volts (scale 0 to 1 on the left - multiply by 2 for maximum of 2V)
Bottom: attenuation in dB

upload_2018-4-27_20-19-34.png


One bit of digital signal is "lopped off" with each -6dB step. Maximum voltage is reduced by 1/2 for each -6db step. Note the first step, goes from 1 to 0.5, second step to 0.25

At the far right side there are still 25 of 32 bits of digital resolution.

The top graph is representative of the maximum voltage sent to the amplifier as you attenuate the signal (turn the volume control counterclockwise).

The bottom graph is representative of the level change of what you hear.

The voltage decrease may look dramatic, but, that's how it works. Using 2v above, the decrease would be from 2v to 1/64 volt: 2, 1, 0.5, 0.25, 0.125, 0.0625, 0.03125, 0.015625 volts max as you turn down the volume in 6dB steps.

For equivalent loudness from the speakers, the DAC or the Preamp will send the same voltage level to the amplifier.

Drops of 10dB are generally considered to be sensed by the listener as "half as loud", or gains of 10dB "twice as loud".

3dB is noticeable, but not a major difference, like, "turn it up/down a little bit" . That doubles/halves power.
 
Last edited:

RayDunzl

Major Contributor
Central Scrutinizer
Joined
Mar 9, 2016
Messages
6,802
Likes
2,899
Location
Riverview, FL
#49
In your case, the Amplifier applies 26dB gain, so, those volts (from low to high) going into the amplifier, become

0.311
0.623
1.247
2.49
4.98
9.97
19.95
39.90 volts

coming out of the amplifier.

And, if the speakers are 8 ohm, the power output is

0.012
0.048
0.194
0.775
3.100
12.42
49.70
199.0 Watts

Double the watts for 4 ohm speakers.

---

Note also that your comfortable listening level as you described above is likely to be in the general area of single-digit watts from the amplifier.

60 Watts (your target for the Jekyll/Hyde A->A/B amplifier transition) is likely to be quite loud.
 
Last edited:

cjf

Active Member
Joined
Apr 20, 2016
Messages
129
Likes
21
Location
Somewhere
#50
Thanks for taking the time to lay out these details Ray.

I'll reply after I've had a chance to digest it a bit.
 

RayDunzl

Major Contributor
Central Scrutinizer
Joined
Mar 9, 2016
Messages
6,802
Likes
2,899
Location
Riverview, FL
#51
One last note...

Volume changes don't alter the shape or content of the signal (within reasonable limits), it just changes the "size".

Here's a zoom (in time) of part of the original, and the same signal attenuated by 42dB, zoomed in by time and amplitude. Same shape, different excursion (see scale on left). The third image shows the attenuated signal without the amplitude visual zoom.

upload_2018-4-27_22-24-50.png
 
Last edited:
Joined
Apr 28, 2018
Messages
70
Likes
110
Location
Columbia, SC
#52
Great discussion here. I ordered an HPA4, mainly because I have so many headphones and wanted to see if I can hear a difference.

I see the specs for the HPA2 circuit in my DAC3 DX are:
  • THD < -108 dB (0.0004%) into 300 Ohms
  • THD < -100 dB (0.0010%) into 30 Ohms
  • - 3 dB Bandwidth exceeds 0.1 Hz to 500 kHz
  • Output Impedance, near 0 Ohms
  • Maximum Output Power, 1.25 Watts into 30 Ohms
  • Maximum Output Voltage, 9.2 Vrms into 300 Ohms
Versus the new HPA4:
  • THD < -125 dB (0.00006%), with a statement that:
    • The HPA4 maintains less than 0.00006% THD, even while delivering a full 6 W output into a 16-Ohm load.
    • And from an application note: THD measures better than the -125 dB (0.00006%) limits of our test equipment.
  • SNR > 135 dB, A-weighted
  • SNR > 131 dB, unweighted, 20-20 kHz
  • Frequency Response - 0.003 dB at 10 Hz, -0.001 dB at 20 kHz
  • - 3 dB Bandwidth exceeds 0.1 Hz to 500 kHz
  • Output Impedance, near 0 Ohms
  • Output Noise < 2.45 uV at Unity Gain, 20-20 kHz
  • Maximum Output Power, 6 Watts into 16 Ohms
  • Maximum Output Voltage, 11.5 Vrms into 300 Ohms
  • Crosstalk < -133 dB @ 1 kHz, -115 dB @ 10 kHz (XLR4)
Given the THD improvement, it would seem that I am gaining about 25dB or so of headroom for my Focal Utopias and LCDi4's. I am also gaining a slightly higher output voltage and more output power. Damping factor should remain about the same.

I am hoping I hear a decent improvement given the amount of money I just spent on my pre-order, but maybe not as the HPA2 was a very good headphone amp to start with. I do remember going from my old Emotiva XPR amps to the AHB2's was about a 25dB gain in THD, and I definitely heard that difference in my theater (night and day to me, but my speakers are very sensitive).
 

RayDunzl

Major Contributor
Central Scrutinizer
Joined
Mar 9, 2016
Messages
6,802
Likes
2,899
Location
Riverview, FL
#53
Last edited:

RayDunzl

Major Contributor
Central Scrutinizer
Joined
Mar 9, 2016
Messages
6,802
Likes
2,899
Location
Riverview, FL
#54
Note also that your comfortable listening level as you described above is likely to be in the general area of single-digit watts from the amplifier.
Hooked up the 'scope to the speaker leads. I see 3 volts peak at this moderate listening level.

upload_2018-4-28_14-18-24.png


3Vpk = 2.12Vrms

That's around 1.12 watts with my 4 ohm speakers, 2.25 watts peak

Take away 26.4 dB gain for the amp, that's about 0.14Vpk for the output of the DAC going into the amp.
 
Joined
Apr 28, 2018
Messages
70
Likes
110
Location
Columbia, SC
#55
No doubt that if I hear a difference it will be from lower distortion. The damping factor can't be that much better than the HPA2 (which was already very well designed), and I agree I don't need the extra power (although your chart assumes an ideal amp and constant driver impedance as you likely already know).

I think what we need to see are THD measurements of both amps at a low output like 3 mW (not at full power). The AHB2 was different than most amps in that it had ultra low distortion even below 1W, so hopefully this headphone amp behaves the same.

I believe having sensitive headphones will help to highlight the lower distortion of the HPA4, as this was my experience when I changed from Emotiva's XPR amps to the AHB2's. I also experimented with full active tri-amping using both amps, and the THX AAA design in the AHB2 can't be beat and showed a night and day difference attached to sensitive Klipsch Palladium speakers, especially when I tri-amped them too.

Wrt to price, you can without a doubt hear a difference when using headphones of this caliber (unfortunately). I am a patient buyer so I was able to find each on sale new for ~40% off. The price is worth it to me knowing that I can use them and sell them for a only $100-200 loss, but headphone prices need to come back down. No way would I ever pay retail price and lose over a thousand later when I sold them.

 
Last edited:

RayDunzl

Major Contributor
Central Scrutinizer
Joined
Mar 9, 2016
Messages
6,802
Likes
2,899
Location
Riverview, FL
#56
I think what we need to see are THD measurements of both amps at a low output like 3 mW (not at full power).
I live in the real world, too, and try to keep things rational, at least in my own mind.

I'm not so much a headphone listener, don't need to be.

and showed a night and day difference
upload_2018-4-28_14-51-0.png
 
Joined
Apr 28, 2018
Messages
70
Likes
110
Location
Columbia, SC
#57
So, lets look at the AHB2 versus some other amps for 1kHz and see what distortion looks like at very low levels.

Doing some rough estimates based on this white paper: https://benchmarkmedia.com/blogs/application_notes/interpreting-thd-measurements-think-db-not-percent...

For simplicity, just pick the lowest measured THD under 1W from the Stereophile measurements below:
.0005% THD => -106dB (Benchmark AHB2, $3k)

.006% THD => -84dB (Mark Levinson No 536, $15k each)

.02% THD => -74dB (Emotiva XPA Gen 3, $1k)

Now, assume we want a playback level of 80dB, which means we should plan for 100dB with peak.
100dB + each of the calculated values above yields:
-6dB Benchmark (won't hear any distortion)
+16dB Mark Levinson
+26dB Emotiva

This kind of illustrates what I heard when I swapped my amps out. I think of it like this, your ear can hear different frequencies and different levels for those different frequencies at the same time. So if you have a background singer and a main piano track, you are hearing multiple frequencies at once (and some are at low levels) and so these ultra low wattages become very important for background detail. Another place this really came into play was hearing the echo and decay of the room in the recording, artists breathing, etc... These affects were all masked before with my previous amps (although they were decent amps for the price).

It is true that this is all environment dependent. If a room's noise floor is 40dB with nothing turned on it will be harder to hear this difference. I have my theater very well treated and use horn speakers too which focus to the main listening position. I can definitely hear someone beside me talk when I play something at normal level, so to me the distortion difference I saw between amps must be audible.

I have no idea of how the HPA4 will behave versus HPA2, but given that swapping all my amps out with one of the same topology, I am really hoping for some audible distortion improvement.

image: https://www.stereophile.com/images/1115BAHB2fig04.jpg




Fig.4 Benchmark AHB2, sum of distortion harmonics (dB ref. fundamental) vs 1kHz continuous output power into 8 ohms.


image: https://www.stereophile.com/images/1115BAHB2fig05.jpg




Fig.5 Benchmark AHB2, sum of distortion harmonics (dB ref. fundamental) vs 1kHz continuous output power into 4 ohms.


image: https://www.stereophile.com/images/1115BAHB2fig06.jpg




Fig.6 Benchmark AHB2, bridged-mono mode, sum of distortion harmonics (dB ref. fundamental) vs 1kHz continuous output power into 8 ohms.

Read more at https://www.stereophile.com/content...er-amplifier-measurements#wA2SWl8iAhXKwekc.99

image: https://www.stereophile.com/images/617ML536fig04.jpg




Fig.4 Mark Levinson No.536, distortion (%) vs 1kHz continuous output power into 8 ohms.


image: https://www.stereophile.com/images/617ML536fig05.jpg




Fig.5 Mark Levinson No.536, distortion (%) vs 1kHz continuous output power into 4 ohms.

image: https://www.stereophile.com/images/817Emotivafig04.jpg




Fig.4 Emotiva XPA Gen3, distortion (%) vs 1kHz continuous output power into 8 ohms.


image: https://www.stereophile.com/images/817Emotivafig05.jpg




Fig.5 Emotiva XPA Gen3, distortion (%) vs 1kHz continuous output power into 4 ohms.

Read more at https://www.stereophile.com/content...er-amplifier-measurements#C4U8dJhrEy1EWAex.99

Read more at https://www.stereophile.com/content...er-amplifier-measurements#HA5O0PiPD8e4tGTt.99
 

RayDunzl

Major Contributor
Central Scrutinizer
Joined
Mar 9, 2016
Messages
6,802
Likes
2,899
Location
Riverview, FL
#58
I don't know!

Interesting experiments for you to perform.

In June, or whenever you get the unit.
 

cjf

Active Member
Joined
Apr 20, 2016
Messages
129
Likes
21
Location
Somewhere
#60
In your case, the Amplifier applies 26dB gain, so, those volts (from low to high) going into the amplifier, become

0.311
0.623
1.247
2.49
4.98
9.97
19.95
39.90 volts

coming out of the amplifier.

And, if the speakers are 8 ohm, the power output is

0.012
0.048
0.194
0.775
3.100
12.42
49.70
199.0 Watts

Double the watts for 4 ohm speakers.

---

Note also that your comfortable listening level as you described above is likely to be in the general area of single-digit watts from the amplifier.

60 Watts (your target for the Jekyll/Hyde A->A/B amplifier transition) is likely to be quite loud.

OK so I'm not sure what the reason is as to why with the external PreAmp in the chain compared to DAC direct there is a quite noticeable difference, for the better in terms of SQ and perceived output power, if the source of these differences are not due to a lack of voltage being supplied to the Amps while using the DAC directly. Could it simply be due to the added Salt/Pepper the external PreAmp is adding to the signal?

The sound levels with the DAC direct config seem more like an On/Off switch after a certain point compared to the external Pre with its Analog VOL control. The external Pre's various VOL positions are much more linear and smooth. I think its this less then linear VOL control with the DAC directly that results in one minute it being too low and the next minute, with just an increase of one additional detent now being too loud, by a lot.

I think in my case the purchase of the HPA4 is still well justified as I also have more then one source and I expect that the HPA4 will be more transparent than my existing PreAmp.

The DAC Direct config sounds great when its at its too loud for normal listening levels position, very clear and dynamic sounding. Its certainly better then with my external Pre in the chain at those times but its during the normal listening level time period where I feel the external Pre pulls way ahead in terms of SQ compared. My hope is that with the HPA4 in the chain I will be able to find that proper listening level while also maintaining the better clarity that the DAC Direct configuration offers. Ideally without needing to go into full Rave mode to get there.
 

Similar threads

Top Bottom