Schiit Mjolnir headphone amplifier - Page 106

Quote:

Originally Posted by Solude 

No because the noise is doubled too ;)  That said the output level of balanced input is the same since the Bryston drops the negative leg.  Ie 3 pins enter, two get grounded, 1 continues through the amp.  I'd have to look again but its also possible Bryston sums the input, same difference.

No, no, no!

The Bryston's balanced input is differential:  i.e. it "reads" or "sees" both the non-inverting and inverting input and converts it into a SE signal.

This is a Common Mode Noise rejecting input.

Quote:

Originally Posted by SoupRKnowva 

 

it could be that its a differential input stage, so its outputting a single ended signal regardless to go to the voltage gain stage?

Right, the volume control then goes to two output stages: a non-inverting and an inverting output stage, hence the BHA-1 can drive either SE or balanced phones.

Obviously the Balanced output has +6dB more output voltage.

Summed, dropped, call it what you want, the signal whether single ended or balanced becomes single ended.  Akin to saying diner is fresh even though its been frozen and reheated :p

And no we aren't talking about the Mjolnir which is the only actual balanced front to back amp available that isn't a bridged amp.  Unless I'm forgetting one?

Anyway enough derailing.  Schiit delivered an awesome focused product that no one is going to touch at, near or up to 3x its price.  The new GS-X might, but its intro price is $2500 so um ya.  Only thing I wish it had was a pre-out defeat switch because power amps with remotes?  Rare ;)

Quote:

Originally Posted by SoupRKnowva 

 

Not quite, most dacs ive looked at that have extensive specs listed, show the balanced output having at least a couple db higher SNR. Like my D2 for instance, 123 vs 126 db single ended vs balanced outputs.

In principle, the signal is doubled while the noise, being random, doesn't exactly add together. At best, this will give you a 6 dB better S/N ratio. Of course, when we're talking about multiple devices, this is only true if the noise is input noise on the receiving device (if the other device is noisy, then the differential signal will be delivering a 6 dB louder copy of an already-noisy signal and you won't gain anything). This isn't usually why people use differential equipment.

One reason that differential is popular is that, on any truly balanced connection (whether the equipment itself is differential inside or not), noise that is common to both sides of the signal is cancelled out. This means, with interconnects, that any hum picked up by the balanced cable, being picked up equally by both wires, cancels out and disappears. (This is only important if your wire is picking up hum.)

The other big reason is that, with two equal amplifiers, we expect the distortion in each half to be equal. Therefore, since they are connected to the load out of phase, the DISTORTION in the amplifiers should also cancel out - giving you a much lower distortion overall. This may or may not work out in practice but, with a good design, a differential amplifier will have much lower distortion...

In the case of a DAC, you usually get a few dB better S/N using the differential output... because, internally, what you usually have is two separate DAC chips decoding each half of the signal separately. Since both are delivering the same desired audio output, but the noise and distortion tend to cancel, you should get at least slightly better S/N and distortion figures.

All in all, though, there are so many other design considerations that make more of a difference that it isn't at all wise to assume that a particular differential amplifier will outperform a different unbalanced one. Many differential designs are otherwise flawed, and many that actually work well otherwise fail to realize any of the POSSIBLE benefits of being balanced. Likewise, some amps perform very poorly in unbalanced mode, but really do improve a lot when used balanced. You rally have to assess each product on its own merits.

Quote:

Originally Posted by Solude 

Summed, dropped, call it what you want, the signal whether single ended or balanced becomes single ended.  Akin to saying diner is fresh even though its been frozen and reheated :p

 

And no we aren't talking about the Mjolnir which is the only actual balanced front to back amp available that isn't a bridged amp.  Unless I'm forgetting one?

 

Anyway enough derailing.  Schiit delivered an awesome focused product that no one is going to touch at, near or up to 3x its price.  The new GS-X might, but its intro price is $2500 so um ya.  Only thing I wish it had was a pre-out defeat switch because power amps with remotes?  Rare ;)

The Balanced converted to SE converted back to balanced argument is a red herring, a straw man, etc.

So what if it is converted to SE?

I would argue that comparing the Mojo topology to the BHA-1 topology is irrelevant.

Just get the one you prefer the sound of. 

Quote:

Originally Posted by Chris J 

 

The Balanced converted to SE converted back to balanced argument is a red herring, a straw man, etc.

So what if it is converted to SE?

 

I would argue that comparing the Mojo topology to the BHA-1 topology is irrelevant.

Just get the one you prefer the sound of. 

From what Chris J says, the BHA-1 has a real balanced input. The main benefit of a balanced input is that it has common mode rejection. This means that any signal equally present on both input lines is rejected (the music is carried as a *difference* between the two inputs). The main benefit of a balanced input is that any hum picked up by the interconnects (which is picked up equally and in phase by both signal leads) is cancelled and so disappears. This is strictly a matter of the *input circuitry* being balanced.

The main benefit to differential amplifier sections is that, since you are using two of the same amplifier out of phase, the distortion and nonlinearities of both cancel out, presumably giving you lower overall distortion and possibly lower noise. This is quite separate from the benefits of a balanced connection.

Maintaining a fully balanced fully differential signal path all the way through isn't always a plus, however. True, you get a fully symmetrical signal path, which should lead to lower distortion. You may, however, actually end up with more noise. Let's assume you DO have hum on those input lines. If the balanced input is converted to single ended, the hum is cancelled out at that stage. If, instead, the device is "fully differential", then each half of the amplifier will amplify the hum that was present on its half of the input. If everything works perfectly, the hum will indeed be cancelled at the output. But, if there are flaws elsewhere, or you connect the output to a downstream component that doesn't have a real balanced input, then the hum will suddenly reappear at that point (since it never had a chance to really be cancelled). In that case, a device that had a balanced input, but converted the signal to unbalanced at the input, will actually deliver *less* noise.

(You have the hum flowing through both halves of the amplifier, with the assumption that it will be cancelled at the output.... when maybe it would have been better to eliminate it at the input stage.)

kLevkoff

The real point of a balanced input is to have common mode rejection. This means that noise common to both inputs is rejected, which is why balanced connections are immune to hum picked up in the interconnects.

Quote:

Originally Posted by kLevkoff 

 

Warning: Spoiler! (Click to show)

 

 

From what Chris J says, the BHA-1 has a real balanced input. The main benefit of a balanced input is that it has common mode rejection. This means that any signal equally present on both input lines is rejected (the music is carried as a *difference* between the two inputs). The main benefit of a balanced input is that any hum picked up by the interconnects (which is picked up equally and in phase by both signal leads) is cancelled and so disappears. This is strictly a matter of the *input circuitry* being balanced.

 

The main benefit to differential amplifier sections is that, since you are using two of the same amplifier out of phase, the distortion and nonlinearities of both cancel out, presumably giving you lower overall distortion and possibly lower noise. This is quite separate from the benefits of a balanced connection.

 

Maintaining a fully balanced fully differential signal path all the way through isn't always a plus, however. True, you get a fully symmetrical signal path, which should lead to lower distortion. You may, however, actually end up with more noise. Let's assume you DO have hum on those input lines. If the balanced input is converted to single ended, the hum is cancelled out at that stage. If, instead, the device is "fully differential", then each half of the amplifier will amplify the hum that was present on its half of the input. If everything works perfectly, the hum will indeed be cancelled at the output. But, if there are flaws elsewhere, or you connect the output to a downstream component that doesn't have a real balanced input, then the hum will suddenly reappear at that point (since it never had a chance to really be cancelled). In that case, a device that had a balanced input, but converted the signal to unbalanced at the input, will actually deliver *less* noise.

(You have the hum flowing through both halves of the amplifier, with the assumption that it will be cancelled at the output.... when maybe it would have been better to eliminate it at the input stage.)

 

kLevkoff

 

 

 

 

The real point of a balanced input is to have common mode rejection. This means that noise common to both inputs is rejected, which is why balanced connections are immune to hum picked up in the interconnects.

 

 

It simply boils down to implementation.

Beyond obvious (and legitimate) electrical characteristics like capacitance, resistance, and inductance, most of the various myths and claims about cables are just that, myths.

I just received my HE-500's and they came with some nasty clear, braided, teflon-insulated "magic wire" (I see they sell for about $150 here and about). They're stiff, and, if you tap on the wire, you can actually hear it in the earpiece ("mechanical microphonics" I suppose). I'm sure it's positively wondrous, but when I move I can hear it clattering about in my ear. Such silliness...... I wouldn't pay you $5 for them (although they are pretty and silver, and my cat seems to be entranced by them). I would have cheerfully paid $50 more for some nice flexible rubber-coated plain old copper wire instead - and I don't even care how many nines it is or how many inches long the crystals are.

Because the HE-500's are orthos, and so draw significant power, I opted to make cables out of the thicker Canare Star Quad 20 gauge x 4 version (the L4E6S is 24 gauge, and so quite a bit higher resistance - the orthos actually draw some current, so the resistance would lower the damping factor). It was a whole $0.31 a foot (the shipping cost more than the wire), and costs about the same from any reputable parts house...... and those little gold SMC connectors (which certain people sell "for HifiMan cables" for $12 each) - they go for $1.50 apiece on eBay.

Almost all commercial copper wire is "six-nines" these days, including the $5 rolls of speaker wire they sell in the five-and-dime store, and the crystal structure of the copper really doesn't make it sound (or measure) any differently; the only thing it improves is certain folks' profit margins. Perhaps he's not "stating the casting process name of his wire" because he knows it doesn't matter :)

kLevkoff

 

Quote:

Originally Posted by Girls Generation 

"high condutive pure copper low capacitance stranded quad wire conductors"

 

I don't know why he's just not stating the casting process name of his wire. OCC/OFC/etc. :)

 

Further, I'm sure many here use aftermarket cables, but refrain from advertising for many different reasons. 

Quote:

Originally Posted by kLevkoff 

I opted to make cables out of the thicker Canare Star Quad 20 gauge x 4 version (the L4E6S is 24 gauge, and so quite a bit higher resistance - the orthos actually draw some current, so the resistance would lower the damping factor). It was a whole $0.31 a foot (the shipping cost more than the wire)

 

Which is what Audeze cans used to come with as well.  Not terribly flexible though so I opt for Q.  No magic, just good materials and excellent workmanship. 

FWIW, and I believe I mentioned this in a past post. The differences between SE and balanced inputs to the BHA1 were extremely minor if any (could be placebo on my part). The differences between SE and balanced inputs to the Mjolnir were minor but definitely evident. This is using the SE and balanced outputs from the PWD2. Topology in this case does matter, maybe, perhaps, depending upon your source, and only if you believe that balanced differential outputs from a DAC results in sonic benefits other than just CMRR, and that is a big if. YMMV.

Quote:

Originally Posted by purrin 

The differences between SE and balanced inputs to the Mjolnir were minor but definitely evident. This is using the SE and balanced outputs from the PWD2.

I wonder how much of this is the PWD2 single ended output.  The analog board shots aren't great but it looks like the single ended outputs are the balanced signal summed.  Which is a whole other discussion about whether using a balanced anything in single ended mode is gimp.

Quote:

Originally Posted by purrin 

FWIW, and I believe I mentioned this in a past post. The differences between SE and balanced inputs to the BHA1 were extremely minor if any (could be placebo on my part). The differences between SE and balanced inputs to the Mjolnir were minor but definitely evident. This is using the SE and balanced outputs from the PWD2. Topology in this case does matter, maybe, perhaps, depending upon your source, and only if you believe that balanced differential outputs from a DAC results in sonic benefits other than just CMRR, and that is a big if. YMMV.

So would you say that perceived difference between se and xlr is due to the PWD's output? Or is the Mjolnir's balanced input inherently "better" than rca?

I believe the PWD cheats with its SE output. It simply takes the hot signal from the balanced output for SE. So in essence, the balanced output of the PWD is technically superior to the SE output.

That being said, my impression that the Mjolnir seemed to take advantage of the balanced output more than the other amp may possibly have something to do with the fact that the Mjolnir doesn't cheat and convert the signal into SE (for the purposes of using a two-gang volume pot) and back to balanced again to feed the bridged amps. As I've noted in a previous post, the Mjolnir actually has a four-gang pot. That usually means something. Let's say I've seen the innards of the amps (BTW, the design of the Mjolnir is really something. Definitely impressive given the $750 price, and definitely not something you can buy off the shelf parts from Mouser to make yourself. There's some secret sauce going on in there and I'm not going to say.)

Finally, I should mention that my reality and direct experience with these pieces of equipment may significantly differ from others'.

getting ready to 'order up' some, well, Schiit in both the Gungnir and Mjolnir. I may place each item on a separate shelf but I may also choose to stack them.

Which unit, if any, outputs more heat than the other? DAC v.s. amp as far as what should be placed on top of the stack, vs what should be on the bottom assuming one does put out more heat than the other. I would guess the amp puts out more heat and thus should be on the top of the stack. maybe this just doesn't matter at all

thoughts? thxs

http://schiit.com/cart/images/gungnir_01_08.jpg