The Lavry DA11: For your ears only

[Edwood]

Great to see an update to the DA10.

But why still no proper volume knob? Even with a rotary encoder, it would still be much better than a toggle switch.

That toggle switch stands in the way of the DA11 being a perfect standalone product.

-Ed

[USAudio]

A fellow Washingtonian welcome to you Dan and thanks for creating a product specifically for the needs of headphone users.

Can you speak at a high level about the design of the headphone amplifier, how well it will do with high impedance headphones like the HD650, etc.

Thanks!

[jude]

Lavry DA11 just arrived. Here are some photos (click on them for larger versions):

[s1rrah]

Originally Posted by jude

Lavry DA11 just arrived. Here are some photos (click on them for larger versions):

Very pretty machine Mr. Lavry has built ...

The "image" feature is very intriguing to me ...

If I weren't mad in love with my Havana DAC, I'd certainly be ordering ...

Be sure to give us some sonic impressions later!

Best.
.joel

[USAudio]

Originally Posted by jude

Lavry DA11 just arrived. Here are some photos (click on them for larger versions)

Jude, what are your initial impressions of the quality of manufacture of the unit? Solidity?

Any listening impressions you can share when you've had a chance to listen to the unit for awhile are also appreciated.

Also, it looks like there are no feet under that chassis???

[Dan Lavry]

Originally Posted by Edwood

Great to see an update to the DA10.

But why still no proper volume knob? Even with a rotary encoder, it would still be much better than a toggle switch.

That toggle switch stands in the way of the DA11 being a perfect standalone product.

-Ed

Here is why:

The matching between channels is far better (!) then a dual analog pot,
And it holds its accuracy over the whole range. Say you want 60dB range. That means 1000 to 1 resistance range. Now take 2 pots and mount them on a single shaft (dual pot). The mechanical tolerance plus the resistor matching needs to be at least 10000 to 1 to get 10% matching over the range... The results are way off the mark. I could add fuel on the fire by mentioning that the pot resistance is semi logarithmic curve...

Now, with my scheme, you get .1dB absolute accuracy over the whole range, and the matching between channels is even better then that.

Also, unlike analog potentiometers, when you set the gain with a digital display, you get reliable and predictable level (such as 56 mean 24dBu, 55 means 23dBu...). When you move an analog pot, you lose the calibration. Some people want such accuracy (mastering engineers, hearing tests labs and others)

Digital attenuation also provides accuracy but digital attenuation causes loss of bits (around 1 bit loss for each 6dB attenuation). When attenuating by analog means, you do not truncate digital bits.

Also, the digital controlled analog attenuation does not have the wear and tear of a mechanical pot:
With analog pot, the signal itself passes through the pot contact (friction between a conductor and the resistive material).
With a switch, the contact resistance can deteriorate to 100 Ohms (don't worry it will not happen), and everything still operates like new. The switch does not pass analog, it is just an "up one" or "down one" command.

Rotary encoder is also a friction based device, and while not as bad as a pot, it does pass the signal and it does wear out over time. Doing a rotary for 56 positions at 1dB steps may not be practical. Some analog attenuators use 2 rotaries, such as for 1dB step and the other for 10dB steps. Say your setting is at 20dB and you want to change it to 19dB. You first need to attenuate to 10, then you need to do 9 steps to 19dB... This is not user friendly way... The switch is better - note that the switch controlled volume speeds up when you hold the switch up or down.

The digitally controlled analog attenuation is much costlier then a dual analog pot. I chose that method for the above reasons.

I will be out for a few days, going to give a lecture at AES (Audio Engineering Society) in Mexico City.


Regards
Dan Lavry

[thomaspf]

Hi Dan,

I agree with everything on the digitally controlled analog volume control except the rejection of a volume knob. My family still struggles with the toggle on my DA10 in the living room. I guess the remote on the DA11 is one answer but I have to say I personally also like the tactile experience of a volume knob.

I don't pretend to be a DAC designer but a simple design alternative here could be something similar to the Volumite controller from Twisted Pear Audio. Any pot combined with a small micro controller. If you do not like the friction based pot there are friction-less optical rotary encoders that you could interface to the micro controller and make it do any resolution you like in firmware. Fast turns might give you larger jumps slow turns go step by step. Just a thought ...

Great to have you post here

Cheers

Thomas

[Edwood]

Originally Posted by Dan Lavry

Here is why:

The matching between channels is far better (!) then a dual analog pot,
And it holds its accuracy over the whole range. Say you want 60dB range. That means 1000 to 1 resistance range. Now take 2 pots and mount them on a single shaft (dual pot). The mechanical tolerance plus the resistor matching needs to be at least 10000 to 1 to get 10% matching over the range... The results are way off the mark. I could add fuel on the fire by mentioning that the pot resistance is semi logarithmic curve...

Now, with my scheme, you get .1dB absolute accuracy over the whole range, and the matching between channels is even better then that.

Also, unlike analog potentiometers, when you set the gain with a digital display, you get reliable and predictable level (such as 56 mean 24dBu, 55 means 23dBu...). When you move an analog pot, you lose the calibration. Some people want such accuracy (mastering engineers, hearing tests labs and others)

Digital attenuation also provides accuracy but digital attenuation causes loss of bits (around 1 bit loss for each 6dB attenuation). When attenuating by analog means, you do not truncate digital bits.

Rotary encoder is also a friction based device, and while not as bad as a pot, it does pass the signal and it does wear out over time. Doing a rotary for 56 positions at 1dB steps may not be practical. Some analog attenuators use 2 rotaries, such as for 1dB step and the other for 10dB steps. Say your setting is at 20dB and you want to change it to 19dB. You first need to attenuate to 10, then you need to do 9 steps to 19dB... This is not user friendly way... The switch is better - note that the switch controlled volume speeds up when you hold the switch up or down.

The digitally controlled analog attenuation is much costlier then a dual analog pot. I chose that method for the above reasons.

I will be out for a few days, going to give a lecture at AES (Audio Engineering Society) in Mexico City.


Regards
Dan Lavry

A toggle switch is still a mechanical component that is subject to wear and tear. I have a rotary encoder in my Grace m902 that I use all the time, for over 3 years and it's still going strong. The only problem I have with rotary encoders is it can sometimes be misinterpreted by the digital control if you spin it too quickly, and can jump up in volume unexpectedly. Either way, a rotational control is very familiar and more ergonomic for better control.

I'm not saying a rotary encoder is perfect, but a toggle switch is even worse, IMO. It would be even more reliable to have no physical switches at all, but how practical would that be? The clicky clackity of a toggle switch really detracts from the headphone listening experience.

But I can see that the use of a remote controller would alleviate this issue.

Anyways, we can agree to disagree.

The visual display of "crossfeed" is very interesting. But coupled with variable control of the "crossfeed" is a first for an all in one unit like the DA11.

Looking forward to checking it out at CanJam.

-Ed

[jude]

So far, I'm having fun with the DA11. It's a nice-sounding unit, no doubt, but I won't say much more than that this early on.

At first, the controls/adjustments seem a bit daunting, but once you figure out (just some reading of the manual), they're second nature. The toggle volume control is (as others have pointed out) not my favorite feature of it. I'm curious to know how well the channel-matching is with it at the lowest volume levels, which I'll test with my various IEMs later.

The adjustable crossfeed (PIC) is very cool. In the past, my favorite crossfeed overall has been HeadRoom's, as it had the best soundstage coherency and shape of the ones I'd used previously, the downside to HeadRoom's being a bit of bass boost with the crossfeed on. For a long time--when one of the older-generation HeadRoom Maxes was my main amp--I rarely listened without crossfeed on.

Meier's crossfeed (at least as implemented in some of the older-generation Meier amps I have) is nice, but had the opposite effect--when on, it can give a sense of a touch of bass reduction--certainly, its effect in this regard was less than HeadRoom's in the opposite direction. It didn't soundstage as three-dimensionally as HeadRoom's, but the greater tonal neutrality was certainly a big plus.

Early on I'd say that the PIC function would sound more like Meier's than HeadRoom's in terms of tonality, but, as with different crossfeed types, it has its own signature. I don't feel comfortable quite yet characterizing its effect on soundstage, as I simply haven't run as much material through it as I'd like to come to any conclusion about that. When listening to recordings (especially from the early stereo era, where over-use of hard-left/hard-right stereo was more common), crossfeed can be indispensible, and PIC is no exception. My early impression, as far PIC goes as a crossfeed type, is that I like it a lot (again, as far as crossfeed goes, which I've not been using much of at all over the last several years).

One cool "newer" song that helps one appreciate PIC (and crossfeed in general) is Blind Melon's "Change" (especially the first 40 seconds or so of it), where things start out totally left and totally right. The left-right channel bias is so bad in that song that, for the first minute of it (through headphones), the maximum crossfeed setting of PIC is actually very helpful in making an actual image out of the beginning of the song (as opposed to a separate show for each ear).

I'll play more with it, and definitely comment more about the DA11 as a DAC, especially after some new balanced cables arrived (I don't have any good balanced cables here at the moment). Of course, I'll also say more about the DA11 as a DAC/headphone amp combo later, too. I just want more time with it.

Early on, I have to say I'm having a lot of fun with it, it sounds excellent, and so have actually been carrying it back and forth between work and home since it arrived.

[fkclo]

It would be interesting to see how the PIC of Lavry DA11 compares to that of the Phonitor, which to-date is the best I have ever heard. Unfortunately, I have no plans to get a DA11 this soon so those who have the opportunity to listen to both please share your views.

F. Lo

[m0ofassa]

That has killer smooth looks jude. Congrats

[USAudio]

Originally Posted by jude

... I'll play more with it, and definitely comment more about the DA11 as a DAC, especially after some new balanced cables arrived (I don't have any good balanced cables here at the moment). Of course, I'll also say more about the DA11 as a DAC/headphone amp combo later, too. I just want more time with it ...

[Dan Lavry]

Originally Posted by Edwood

A toggle switch is still a mechanical component that is subject to wear and tear. t CanJam.

-Ed

Here is the point:

Yes, a toggle switch, or a rotary, or a relay are still a mechanical switch.
But I am NOT passing the analog signal trough the mechanical contact.

I am using the switch to generate a control signal for setting the volume level. But the analog signal (music) does not pass through a mechanical switch contact; it passes trough solid and permanent soldered contacts.

Say you have a switch connected between 0V and 5V. You also have a detection circuit that tells you if the signal is over or under 2.5V. The output of such a circuit is the same if you feed it 5V or 3V (both are over 2.5V). You get the same result if you feed the circuit 0V or 2V (both cases are under 2.5V). So you can "live with" 2V error in the signal. That is OK, because your switched based digital control circuit is only being used to send a message to the analog circuit (made with soldered connections), and the digital message is to increase volume or to decrease volume.

The switch will have to deteriorate very badly to yield 2V error. I am exaggerating the error to make a point, to show that even 2V will not do you any harm.

Now try and put some .002V AC signal on mechanical contacts that carry the analog signal (music), and you have a badly broken unit.

So yes, a switch can wear out, and so can a rotary. But that is not the point. When you pass a signal trough a switch, a slight amount of wear can impact the result a lot. Such is the case with much of the gear out there, certainly with the rotary based resistive attenuators. My method lets me bypass the wear and tear issue completely. My switch can heave ridiculously high contact resistance, even 100 Ohms (that will not happen) without impacting the sound at all.

Regards
Dan Lavry

[revenge]

Dear Mr Lavry
First of all let me tell you it's a great honour to address you. I've been using your DA-10 for ages, even after buying the CD-X2 from Naim. I have never considered selling the "black box" and have always been happy with it. It's always been around, small, modest looking, but ready to deliver a big sound in no matter what setup. In fact DA-10 is one of the best investments I have ever made. That's why I have never felt any desire to upgrade it, until now. And the reason is not the USB out, the crossfeed or the infrared remote sensor that all come with the new DA-11 (and which are of no importance to me as I am not using the computer as a source or the headphones too much) but the SQ itself. So the big question for me is: is the DA-11 better sounding than the original version? Is it an improved version or just the same DAC in shiny, new clothes?

[revenge]

Also, from a consumer point of view (as I understand DA-11 is consumer orientated), I have to confess I am a bit dissapointed by your decision to keep the XLR as the only line-out option. The Neutrik adaptors (which I already have) are a nice addition to the standard package but they also add a certain length on the rear of the unit (the Neutrik adapter plus the phono plug plus the cable end) not to mention that, no matter how good, they still degrade the sound.
And as much as I like the capability to switch off the line-out circuit when you listen through headphones, I would have liked a similar option to switch off the headphone out when not in use.
While the switch is very accurate, as you've thoroughly explained above, it is by no means easy to use. On one hand it is quite stiff and it jumps too easily from one level to another. Quite often I have to try several times until I reach the desired level. So as good as it might be in a studio orientated setup, in daily, consumer use is far from convenient. My fingers hurt after just a few attempts to set it. Luckily, most of the time I don't have to alter the volume.
While I don't like the Benchmark DAC-1 sound at all, the new Pre version is years ahead of the spartan Lavry interface in terms of convenience and, to be honest, that's what I would have expected from a consumer-orientated version of DA-10 as well.
Again, the sound of DA-10 is well worth the money and makes up for any minor niggles encountered in daily use.