[obobskivich]

Geez!  Amen to that! Much easier to copy/reply to, too! :confused_face_2:   Although, in his defense, a lot of it is Head-Fi's current forum software.

Boo Huddler!

Simple.  You need capacitors so large on the output of OTL amps for low impedance phones that they're simply impractical, not even manufactured in those sizes, even.  (All because of the RC circuit and the resulting cutoff frequency.)  Most mfrs don't even try.  I submit that's a more common reason of limiting low headphone impedance on OTL amplifiers, rather than claiming any instability in low impedance loads.  The mfrs simply don't want to be accused of having an amplifier without bass, which would happen when paired with most low impedance phones.

This may very well be true - I know some manufacturers have out and out said it's a stability concern (Schiit has, for example), but it certainly could be true of other builders (what, manufacturers lying? impossible!). But it isn't like transformers don't exist. (e.g. the WA6 and the Cayin are both suitable for lower Z cans - they both cost a lot though (last I looked the Cayin had just bumped up on $900)). I know that even with some SS amps though (as I mentioned above), the manf is saying it's a no-go, but it's usually *very* low impedance (usually <16R, which is getting into speaker territory) - there's very few full-size headphones that fit into that (I can name a total of four; only one is currently in production - none are very popular). In those cases I'm thinking it's a current demand concern (the ironic part is, the four headphones I can think of, also happen to be near the top of the list in terms of sensitivity :rolleyes - just like cheap full-size amps and 4R speakers. Regardless, a cheapo-deluxe OTL is not something I'd want to play around with.

This is just my personal opinion, but I have had a pair of K701's for many years.  I think perhaps the tube hybrid is among the best for proper feeding of such headphones - although I would limit the feedback involved, if possible.  The high current and somewhat higher than normal output impedance from a typical tube hybrid tends to give the K701's ample current while not emphasizing the harshness that can occur with SS, super-low-output impedance amps.

I will say that I never found the 701 "additionally harsh" with pure SS amps; they did nice with the CI VHP (which is TPA6120 based). But I could certainly see someone wanting to warm them a bit (EQ, amp, whatever). I know Music Hall has a hybrid based on that same chip, forget which tubes it uses - might be worth looking at in addition to the Lyr (it runs around the same price), assuming you want tubes in the chain somewhere. Would drive basically any dynamic headphone out there with ease too.

 

[J Bones]

Quote:

If you don't embed your replies in the quotes, it's much easier to read.

An example of an OTC/WOT amp would be the Woo WA6, and the transformer does put the price up (my understanding is the WA6 is just the WA3 with an OT). I'm not sure what Zout actually is on that amplifier, but with the transformer involved, it's suitable down to 16R (and Jack has confirmed this). Output impedance doesn't absolutely dictate what you can plug into though (for example many receivers and full-size power amps will have an output impedance of ~400-600R on the headphone jack, as a result of series resistors between the power amp and the jack (to protect both sides), and they will get along fine with more or less anything (but you might get some hiss (from some amps) with extremely sensitive and low impedance headphones that also offer some isolation (like IEMs or cans like the ESW9)).
Yeah, by and large trial and error is probably your best bet (especially if you subscribe to "all amps are inherently different sounding" - no harm if you do, it's just that if you do, all the charts and graphs in the world won't tell you what you want to know). Personally I found the 701 to be fairly unpicky, as long as the amplifier in question could provide enough power (so basically any dedicated headphone amplifier, stereo receiver, etc was suitable, but portable devices, PC soundcards, etc tended to be hit and miss). With something like Lyr you should not want for power, and the same would be true if you had a good stereo receiver, or even a "lesser" desktop headphone amp (like the FiiO E9); so it comes down to other considerations, like aesthetics, ergonomics, price, etc. Aside from a select few amplifiers that might be right at the lower bound of their stable range, and amplifiers that cannot provide enough power, I don't think you're going to run into any legitimate compatibility concerns though. If you're looking at true OTL I'd say ask the manufacturer before you buy, but otherwise it's pretty much free-range (I honestly can't think of a production SS amp that isn't going to handle 62R). You'll probably have a better time of finding tubes in a hybrid (like Lyr) than in an OTL for these cans though - but in that $400-$600 range there's some OTC/WOT amps (like WA6) that you could also consider. If that's a direction you want to go. There's plenty of hybrids around that price too, and plenty of pure SS amps (and some that cost less than that range and are good performers as well). Plus various full-size components. Really you have a lot of options.

 
Thanks for the help man, I really got a good feeling about the Lyr reading some of the reviews on here about it and specifically it's performance with the Q701, I just couldn't figure out why it might be a good match. I'm actually going to take a look at some of the other OTC/WOT amps like the Woo to see how much they might all differ but I think I've a good select few amps to truly trust and delve into in order to find the one that'll best suite me.
 

Originally Posted by tomb /img/forum/go_quote.gif
 
All I cited as an example was a voltage threshold (and how you could easily dismiss portables and solid-state devices with high-impedance headphones).  It does not follow at all that a portable amp would be able to meet the power requirements.  Voltage is only one variable in the power equation.

 
Like I said above this is what tripped me up. I wouldn't consider Q701's or anything around their resistance to be high-impedance, but many have told me a portable amp will do no justice for the cans and that a desktop amp is necessary (I was always assuming they were making this claim off the fact that the Q701's were high-impedance in their eyes and this required more voltage to operate the way they were intended). But really, I think testing out a few amps will be my only option to know what really works best.
 

 
Simple.  You need capacitors so large on the output of OTL amps for low impedance phones that they're simply impractical, not even manufactured in those sizes, even.  (All because of the RC circuit and the resulting cutoff frequency.)  Most mfrs don't even try.  I submit that's a more common reason of limiting low headphone impedance on OTL amplifiers, rather than claiming any instability in low impedance loads.  The mfrs simply don't want to be accused of having an amplifier without bass, which would happen when paired with most low impedance phones.

 
I can't figure out whether the Lyr has blocking capacitors or how big they are, either.  Just looking at it, it appears they have a solid-state output buffer whose offset is zeroed out with an opamp.  If that's the case, then they may not have any blocking capacitors at all.
 
They seem like good guys, though, why not just send them an e-mail or PM and ask them directly?

 
I think an email is in order 

 
 

Low output impedance can be achieved a couple of ways - no output capacitors and/or lots of feedback.  As for the transformer(s) (it has two in the pics), it appears to me that one is the PT (power transformer), while the other is a transformer for the heaters.  Many times, those functions are combined into one transformer.  However, to keep things simple and cheap, it might've been easier to use two transformers.  This is because the high-voltage transformation needed for the tubes are in contrast to the typically low-voltage and high-current heaters in the tubes.  It appears the Lyr is designed around the 6DJ8/6922 family of tubes.  Those tubes might have 200V on the plates, while the heaters need only 6.3V but at 300ma each.  IOW, it's two different optimizations in the way the transformers are wound and it may have worked out cheaper to have two separate ones, despite the increased cost in iron.

This is just my personal opinion, but I have had a pair of K701's for many years.  I think perhaps the tube hybrid is among the best for proper feeding of such headphones - although I would limit the feedback involved, if possible.  The high current and somewhat higher than normal output impedance from a typical tube hybrid tends to give the K701's ample current while not emphasizing the harshness that can occur with SS, super-low-output impedance amps. 

 
I guess I could have just taken a look at the pictures myself lol. But yeah I do see that...doesn't the Lyr though already have a pretty low output impedance?

---

 

[tomb]

Quote:

Originally Posted by J Bones /img/forum/go_quote.gif
 
I guess I could have just taken a look at the pictures myself lol. But yeah I do see that...doesn't the Lyr though already have a pretty low output impedance?

Yes, but it also says Zero Feedback - that's good!  As I posted earlier, it appears that there's an opamp that's used as a "servo" to zero out the offset at the outputs.  That means you can design the amp circuit without using blocking capacitors on the output.  The result is a much lower output impedance and a sound that's not as "filtered" as happens with capacitors on the output.  The relay-delay is still necessary, however, because the opamp does not operate to zero the offset in zero time - it takes a bit to settle out before damaging voltage at the output disappers.
 
All that said, their tongue-in-cheek description of the "dynamically adaptive" output circuitry - an "ancient Teutonic magic trick" may be more truthful than they're stating out-right.  I could very well be wrong, but it may be their way of saying that the Class A bias is set at a certain current (perhaps lower than some DIY amps, for instance) and that once that current is exceeded, the output devices will simply "switch on" to supply more current if necessary.  In terms of an amplifier circuit and the devices, this is almost like saying that a dropped ball will fall down.  I could be wrong, of course, not having seen a schematic.

 
Still, my personal opinion - not knowing anymore specific details at this point - is that the Lyr may be a very good choice to drive K/Q-701's.  Zero feedback, tube signal stage, and lots of current capability.  I would still recommend a listen or read the Lyr threads extensively.  The proximity of those transformers may cause some hum in very efficient headphones.  That probably wouldn't affect the K/Q-701 that much, but very few of us only plug one type of headphone into an amp.  Just a guess, though.  

 

[TMRaven]

Higher impedance = higher head-fi self-esteem.

 

[J Bones]

Quote:

Yes, but it also says Zero Feedback - that's good!  As I posted earlier, it appears that there's an opamp that's used as a "servo" to zero out the offset at the outputs.  That means you can design the amp circuit without using blocking capacitors on the output.  The result is a much lower output impedance and a sound that's not as "filtered" as happens with capacitors on the output.  The relay-delay is still necessary, however, because the opamp does not operate to zero the offset in zero time - it takes a bit to settle out before damaging voltage at the output disappers.
 

 
 
I understand to an extent the purpose of an opamp, but I don't see what two outputs it's offsetting and why we would want to zero these out? It makes sense though that using an opamp instead of blocking capacitors would reduce output impedance, I just don't see how the opamp performs essentially the same job.
 
 

All that said, their tongue-in-cheek description of the "dynamically adaptive" output circuitry - an "ancient Teutonic magic trick" may be more truthful than they're stating out-right.  I could very well be wrong, but it may be their way of saying that the Class A bias is set at a certain current (perhaps lower than some DIY amps, for instance) and that once that current is exceeded, the output devices will simply "switch on" to supply more current if necessary.  In terms of an amplifier circuit and the devices, this is almost like saying that a dropped ball will fall down.  I could be wrong, of course, not having seen a schematic. 

 
Now trying to wrap this around my mind, I cannot get the straight answer why this output constantly is on due to class A bias? Upon reading the following I thought I had it figured out but it still remains cloudy...how do the opamp inputs demand anything at all according to the bolded part?
 

When you put a current source on the op-amp's output inside the feedback loop, it causes the op-amp to continuously "fight" against the current source. The chip must put out at least as much current as the current source demands in order to force its output to the voltage the op-amp inputs demand. If it did not, the current source would pull the output of the op-amp to V-. This continual fight against the current source keeps the op-amp's output stage turned on all the time, with a constant current. Voilá, class A bias.

 

[J Bones]

Quote:

Yes, but it also says Zero Feedback - that's good!  As I posted earlier, it appears that there's an opamp that's used as a "servo" to zero out the offset at the outputs.  That means you can design the amp circuit without using blocking capacitors on the output...

I'm assuming by this you meant zero global feedback correct? I would think there has to be local feedback used at certain stages throughout the amplifier right?

 

[Kodhifi]

Quote:

That graph doesn't show what you think it does - that's a square wave response measurement (not impulse response), which lets us view (acoustic) phase information about the driver (because nothing is DC-to-light, you won't ever have perfect phase; square wave lets us look at the relationship between phase and frequency - a perfect square-wave would mean DC-to-light response, it could still ring though (which is what damping is meaning to correct)). Tyll has an article on InnerFidelity that goes into this in more depth - that plot can be generated from an FR plot, or vice versa (which was actually demonstrated a while ago in the Sound Science section with IIR filters too).

 
I've always interpreted those graphs to show how a driver actually responds with a given input signal. IE ideally square wave goes in, square wave comes out. But electrical signals move at the speed of light and practically no mass (that would be a whole other discussion about special relativity so for our argument it has no mass) where as a driver moves in a fluid (air) that exerts resistance to movement (friction and fluid dynamics) and a driver which has mass and when moving it has momentum.
 
The same way your car doesn't go from 0 to 30mph in the blink of an eye, and stop in the blink of an eye, neither does a driver but it is still pretty fast.
 
The lighter you make your driver, the faster it can respond and the more like a perfect square wave the output will be.
 
 
What I see when I look at that graph is a the driver moving out (peak) but moving out a little too much because it has momentum and it takes it time to stop and change direction, then because it's overshot, it warbles back and forth like a vibrating string trying to move to the position the voice coil is being directed to. Imagine a car driving up onto a curb, the level changes suddenly and the car moves with the level, but not instantly, and it kind of bounces because of it's suspension, trying to find equilibrium again.   Anyway, after the peak then it moves in (valley) and does the same thing. There is a slight amount of time where the driver is accelerating, and decelerating each time it moves and that comes across in the graph as a swinging saw tooth type of peak at the leading and trailing edge of each part of the waveform.
 
Am I incorrect?
 
 
As you see in this graph, the 600 ohms have the smallest peaks, the 250 in the middle, and the 32's the largest.
 

 

[obobskivich]

Am I incorrect?

You're (kind of) describing IRF - square wave response doesn't show the information you're ascribing to it. In order to reproduce a square-wave "perfectly" the system would have to have DC-to-light perfect FR; otherwise the square wave will be non-square (due to phase inaccuracies). It doesn't matter how clean or not the driver is either (because none of them have perfect phase). The differences you're seeing between those plots show up as FR differences when you view the same data differently (one lets you visualize time, one lets you visualize frequency - both show the same system). One doesn't imply "better" or worse than the other, it's just a different way of visualizing roughly the same data. Depending on your personal preferences, one may align more with what you want to hear though. As I explained earlier (with links too), the 600R version looks to have a slightly better IRF plot, but without seeing the computed CSDs and/or hearing them, it's tough to say if that's worth phoning home about (remember - it doesn't matter if you can show it to be different (or to exist or not) to a scope, it matters if that difference >JND). Finally the HeadRoom comparator is worthless for viewing those together, because it isn't overlaying them accurately.

Lighter by itself doesn't mean cleaner/faster either - lighter and better damped/controlled (motor + enclosure) is the requirement there. Larger surface area (which means less excursion (and therefore less THD)) also helps. That will reduce ringing and distortion in principle - but it doesn't by itself mean "this will sound better" or "be better" (if you can even quantify that one) - you may prefer something entirely different, which is ultimately what matters.

Impedance has roughly nothing to do with this though, beyond the Beyerdynamic parts (which use variations of the same driver); higher impedance does not imply better IRF or FR, nor does it imply less ringing or better damping/control. It just means higher impedance.