Head-Fi.org › Forums › Equipment Forums › Sound Science › What causes this (amp related)?
New Posts  All Forums:Forum Nav:

What causes this (amp related)? - Page 2  

post #16 of 178

My professional engineering opinion, having done both amplifier and speaker development, is this is potentially harmful and a serious flaw with the Asgard. The Schiit amps are clearly not designed for good objective performance. And when designers have other goals in mind there are often some serious side effects. NuForce doesn't seem to care the uDac-2 has serious channel balance problems for example. They claim that bad channel balance was necessary for the best sound at the price. The Asgard may have been designed with similarly misguided, or sloppy, priorities.

 

I hate to say this but your AKGs may already have been damaged and it's just not obvious by listening to them.The material properties of drivers are critical to their performance. For example, when I did speaker development, we had a high-end tweeter dome that was accidentally "dented" in. It was a replaceable dome/voice coil assembly so I took the tweeter apart and carefully popped the "dent" out. The dome looked perfect. But the tweeter then measured completely differently and sounded worse. It had a new peak and a big dip in the response that were not there before the dent. We ordered a new dome/voice coil and it again matched the original measurements.

 

The manufacture (Seas) told us denting the dome creates concentric weak areas in the material. And at some frequencies it no longer behaves as much like a piston. Some call this "breakup". The force from the voice coil is supposed to move most of the diaphragm as a piston with the outer suspension being the part that's supposed to flex. But when the diaphragm has weak areas, it effectively "breaks up" into a smaller diaphragm within the bigger one. The weak area acts as an unintended suspension.

 

That was a fabric dome treated with some sort of polymer. The AKG drivers are obviously made out of a different material--I suspect a polyester film like mylar or something similar. My guess is polyester is prone to a similar sort of weakening but I'm not a materials expert so I can't say for sure. But anything that alters the rigidity and flexibility of the diaphragm material *will* change the performance of the headphones. That much you can be sure of.

post #17 of 178
Quote:
Originally Posted by nwavguy View Post

My professional engineering opinion, having done both amplifier and speaker development, is this is potentially harmful and a serious flaw with the Asgard. The Schiit amps are clearly not designed for good objective performance. And when designers have other goals in mind there are often some serious side effects. NuForce doesn't seem to care the uDac-2 has serious channel balance problems for example. They claim that bad channel balance was necessary for the best sound at the price. The Asgard may have been designed with similarly misguided, or sloppy, priorities.

 

I hate to say this but your AKGs may already have been damaged and it's just not obvious by listening to them.The material properties of drivers are critical to their performance. For example, when I did speaker development, we had a high-end tweeter dome that was accidentally "dented" in. It was a replaceable dome/voice coil assembly so I took the tweeter apart and carefully popped the "dent" out. The dome looked perfect. But the tweeter then measured completely differently and sounded worse. It had a new peak and a big dip in the response that were not there before the dent. We ordered a new dome/voice coil and it again matched the original measurements.

 

The manufacture (Seas) told us denting the dome creates concentric weak areas in the material. And at some frequencies it no longer behaves as much like a piston. Some call this "breakup". The force from the voice coil is supposed to move most of the diaphragm as a piston with the outer suspension being the part that's supposed to flex. But when the diaphragm has weak areas, it effectively "breaks up" into a smaller diaphragm within the bigger one. The weak area acts as an unintended suspension.

 

That was a fabric dome treated with some sort of polymer. The AKG drivers are obviously made out of a different material--I suspect a polyester film like mylar or something similar. My guess is polyester is prone to a similar sort of weakening but I'm not a materials expert so I can't say for sure. But anything that alters the rigidity and flexibility of the diaphragm material *will* change the performance of the headphones. That much you can be sure of.


Does this mean that the Schiit amps can damage every brand of headphone or is it impedance related?

 

post #18 of 178

Well, I think that 16 ohm in-ears will like this a lot less than some 600 ohm studio cans.

post #19 of 178
Thread Starter 

Well, I've decided I'm just going to return it at this point.  Jason offered to accept it as a return so I'll just go that route.  It's a shame, if it wasn't for this I'd definitely have been ecstatic with it.

post #20 of 178

Hey all,

 

I just saw this thread here. Shike, sorry you feel like you have to send the amp back. All I can say is this:

 

1. On Asgard, a mild turn-off transient is normal, but we'll check your amp when it comes back to make sure it isn't excessive. 

2. We're getting near 1000 Asgards shipped between 115V and 230V models, and we've never heard of headphone failure caused by Asgard, including IEMs.

3. Our standard final listening-test HD650s have gone through thousands of on-off cycles on all models, and they've never had a problem.

4. We absolutely do design for excellent objective performance, as well as subjective performance. We've had many, many rave reviews on Asgard, including from engineers who have measured the performance of the amp. 

5. Between Mike and I, we have designed and shipped dozens of audio products, ranging from tube preamps to surround sound decoders, and shipped well over 100,000 units of said products over the last 30 years, to both critical and objective acclaim. This ain't our first rodeo. If you have any engineering questions about our products, I invite you to discuss them directly with myself and Mike.

 

All the best,

Jason

 

post #21 of 178
Quote:
Originally Posted by upstateguy View Post


Does this mean that the Schiit amps can damage every brand of headphone or is it impedance related?

 

 

I would say it could, in theory, damage most any headhone but higher impedance ones are likely less at risk. The other issue with DC, even if it doesn't physically trash the diaphragm, is causing permanent magnetic changes (damage). DC currents tend to magnetize ferrous metals, and I'm not an expert on the subject, but some smart people have said certain headphone designs are susceptible to that sort of damage.

 

I'm guessing here, but the Asgard seems like essentially a Nelson Pass-ish single ended design. If it really draws even close to 35 watts of power it has a big power supply, big heatsink(s), and big output devices. Just those things, combined with the complex to manufacture enclosure, cost some serious cash--especially to build it in the USA. That probably didn't leave much money left over for say an output relay and associated circuitry to protect the headphones. So, sort of like with NuForce and the uDAC-2, Schiit may have cut some corners to keep it under $250.

 

But, in my opinion, potentially damaging customer's headphones is not a good corner to cut. And, to go a step further, all that money spent on the relatively massive single ended design is also a waste unless you're after tube-like euphonic distortion. I do give them points, however, for great looking products and a very slick website (even if it is 95% marketing hype).

 

It would be interesting to get the Asgard on the test bench. I'm fairly sure it would measure much worse than a $39 Cmoy in most aspects. That would probably just turn into yet another objective vs subjective debate. The Asgard has enough distortion, and possibly a high enough output impedance, that it may well be easy enough to pick out in a blind test against a more accurate amp. So it would come down to which one you like better--accurate or distorted.
 

 

post #22 of 178
Quote:
Originally Posted by Shike View Post

Hey guys,

 

I got a new Asgard amp.  I found something odd when I shut off the power, the dome on my K702 pulls in to where the surround crumples a bit.  I contacted Jason and he noted this is normal operation.  I was wondering: what causes this?  Any thoughts?  Here's a video:

 




one of the most scariest video's Ive ever seen.

Im looking for an amp for my K601, that is how I got here. 

 

So. is this normal in every argard or was this one defective?

post #23 of 178

 

Quote:
Originally Posted by nwavguy View Post

I'm guessing here, but the Asgard seems like essentially a Nelson Pass-ish single ended design. If it really draws even close to 35 watts of power it has a big power supply, big heatsink(s), and big output devices. Just those things, combined with the complex to manufacture enclosure, cost some serious cash--especially to build it in the USA. That probably didn't leave much money left over for say an output relay and associated circuitry to protect the headphones. So, sort of like with NuForce and the uDAC-2, Schiit may have cut some corners to keep it under $250.


 

You're making a lot of assumptions. To help clear things up:

 

1. Nelson Pass? Nope. But it is single-ended.

 

2. Complex to manufacture enclosure? No. This is one of the ways we can offer such high performance at a very low price and manufacture in the USA. The enclosure is extremely simple, only two pieces of bent sheetmetal. It's actually much simpler and cheaper than most of our competition.

 

3. Big heatsinks? Again, no. The enclosure itself is actually used as the heat sink for the output devices--which eliminates even more cost. That allows us to put it into the power supply.

 

4. In our testing, we didn't believe that an output relay was necessary for Asgard, with such a relatively minor transient. Especially since it would be best to have the relay lift the output, rather than short it, and thus add another contact to the signal path. Our customer experiences tend to bear this out. Many amps in this price range, or even above, do not have output relays. 


Edited by Jason Stoddard - 7/17/11 at 1:09pm
post #24 of 178


 

Quote:
Originally Posted by Jason Stoddard View Post

 


 

You're making a lot of assumptions. To help clear things up:

 

1. Nelson Pass? Nope. But it is single-ended.

 

2. Complex to manufacture enclosure? No. This is one of the ways we can offer such high performance at a very low price and manufacture in the USA. The enclosure is extremely simple, only two pieces of bent sheetmetal. It's actually much simpler and cheaper than most of our competition.

 

3. Big heatsinks? Again, no. The enclosure itself is actually used as the heat sink for the output devices--which eliminates even more cost. That allows us to put it into the power supply.

 

4. In our testing, we didn't believe that an output relay was necessary for Asgard, with such a relatively minor transient. Especially since it would be best to have the relay lift the output, rather than short it, and thus add another contact to the signal path. Our customer experiences tend to bear this out. Many amps in this price range, or even above, do not have output relays. 

 

Thanks for the added info. I agree output relays are not all that common (although the $129 FiiO E9 has one). But the reason most amps don't need them is most amps don't try to destroy your headphone diaphragms when you turn them off like the Asgard apparently does. That's not what I'd call a "relatively minor transient".

 

As for contacts in the signal path that's not a problem in a headphone amp. Period. In audio power amps, where the relays switch under load with currents of up to several amps, they can be a problem. But in a headphone amp the currents are trivial and the contacts entirely transparent. Doug Self has researched the topic fairly extensively. Signals of similar levels are often sent through relays in all sorts of ultra expensive rave reviewed audiophile gear (input selection and even precision stepped volume attenuators).

 

The chassis really is a work of art. You're leaving out the vent "grate", big sexy knob, silkscreening, etc. from your description above. But, regardless, it's a great look for $249.


 

 

post #25 of 178


 

Quote:
Originally Posted by nwavguy View Post

Thanks for the added info. I agree output relays are not all that common (although the $129 FiiO E9 has one). But the reason most amps don't need them is most amps don't try to destroy your headphone diaphragms when you turn them off like the Asgard apparently does. That's not what I'd call a "relatively minor transient".

 

As for contacts in the signal path that's not a problem in a headphone amp. Period. In audio power amps, where the relays switch under load with currents of up to several amps, they can be a problem. But in a headphone amp the currents are trivial and the contacts entirely transparent. Doug Self has researched the topic fairly extensively. Signals of similar levels are often sent through relays in all sorts of ultra expensive rave reviewed audiophile gear (input selection and even precision stepped volume attenuators).

 

The chassis really is a work of art. You're leaving out the vent "grate", big sexy knob, silkscreening, etc. from your description above. But, regardless, it's a great look for $249.


 

Thanks for responding. Our measurements show a 100-150mV transient into a 32 ohm load. This is 0.0007W of transient--or 7/10000 of a watt. In our testing, this was at or below a half-dozen similarly priced amps. 

 

We can argue all day about contacts--but even the relay manufacturers talk about degradation of the contacts caused by outgassing of the coil when left on for long periods of time. This is a real phenomenon, documented on the datasheets. It's something we have to consider when adding a relay.

 

And one last thing: the vent "grate?" That's part of the inner chassis. Still only two parts. Silkscreening? Does anyone not screen their chassis? Big knob? I'm almost embarrassed to tell you what we pay for them (and yes, from a US supplier.) Now, if we were a boutique operation making amps 25 at a time, yep, the chassis is going to dominate the cost. But we're literally making thousands of very similar chassis per year, shared over many products. Chassis cost is not as big of a concern as you think.

post #26 of 178


 

Quote:
Originally Posted by Jason Stoddard View Post

Thanks for responding. Our measurements show a 100-150mV transient into a 32 ohm load. This is 0.0007W of transient--or 7/10000 of a watt. In our testing, this was at or below a half-dozen similarly priced amps. 

 

We can argue all day about contacts--but even the relay manufacturers talk about degradation of the contacts caused by outgassing of the coil when left on for long periods of time. This is a real phenomenon, documented on the datasheets. It's something we have to consider when adding a relay.

 


I suspect what we're seeing in the video is more than 150 mV into the 62 ohm load of the AKGs. The power isn't the issue here as I'm not suggesting the Asgard is causing thermal damage. The issue is the driver is being pushed well past its design limits for excursion by a significant amount of essentially DC. How much voltage that requires on the AKGs, I don't know. But, regardless, it's not good if it's distorting the driver diaphragm. And I think most here will agree it's not typical. So I don't know what your other half dozen amps were, or if your testing method is flawed, but I've never seen anywhere near that serious of a transient in anything I've tested. Most are much more brief and much less violent.

 

The relay outgassing issue is well understood, and from what I understand, rarely a real world problem these days. The datasheets specify the contact resistance over the life of the relay with a million+ contact cycles. And it remains negligible even with a $2 relay. The headphone jack is a much more significant concern. Douglas Self calls relays "blameless" and considers them the ultimate way to switch audio signals. He's used relays extensively in commercial designs (like the Cambridge Azure 840) without any reliability issues. Test instrumentation, including audio analyzers, are loaded with relays in the signal paths.

 

Anyway, it's mostly just another audiophile myth with a few grains of truth behind it (speaker relays in power amps can be a problem). Audiophiles are incredibly adept at taking a specific problem related only to specific circumstances and assuming it also applies everywhere else--like vibration reduction. Then, of course, they use biased, sighted, listening tests to confirm what they expect to hear. And, despite their brains deceiving them in well established ways, a new myth is born and propagated.

 

post #27 of 178

Thanks again for the reply. I suppose we'll have to differ about transients, each based on our own experience. 

 

But, as far as relays go, here's a note from a current Omron datasheet--so, even if it's "rarely a real-world problem," the manufacturers are still warning about it:

 

 

Long-term Continuously ON Contacts

Using the Relay in a circuit where the Relay will be ON continuously for long periods (without switching) can lead to unstable contacts because the heat generated by the coil itself will affect the insulation, causing a film to develop on the contact surfaces. 

 

 

post #28 of 178
I highly doubt that the level of DC offset we're dealing with here will damage any headphones that are made, well, correctly. Dynamic drivers are far from as fragile as many people think. Has anyone actually measured the transient voltage?
Edited by revolink24 - 7/17/11 at 4:34pm
post #29 of 178
Quote:
Originally Posted by revolink24 View Post

I highly doubt that the level of DC offset we're dealing with here will damage any headphones that are made, well, correctly. Dynamic drivers are far from as fragile as many people think. Has anyone actually measured the transient voltage?


I've never seen a diaphragm crinkle like that under normal operation.

 

Shike wrote "I'm hitting 150-250mv DC offset".

 


Edited by xnor - 7/17/11 at 4:39pm
post #30 of 178
Quote:
Originally Posted by nwavguy View Post

My professional engineering opinion, having done both amplifier and speaker development, is this is potentially harmful and a serious flaw with the Asgard. The Schiit amps are clearly not designed for good objective performance. And when designers have other goals in mind there are often some serious side effects. NuForce doesn't seem to care the uDac-2 has serious channel balance problems for example. They claim that bad channel balance was necessary for the best sound at the price. The Asgard may have been designed with similarly misguided, or sloppy, priorities.

 

I hate to say this but your AKGs may already have been damaged and it's just not obvious by listening to them.The material properties of drivers are critical to their performance. For example, when I did speaker development, we had a high-end tweeter dome that was accidentally "dented" in. It was a replaceable dome/voice coil assembly so I took the tweeter apart and carefully popped the "dent" out. The dome looked perfect. But the tweeter then measured completely differently and sounded worse. It had a new peak and a big dip in the response that were not there before the dent. We ordered a new dome/voice coil and it again matched the original measurements.

 

The manufacture (Seas) told us denting the dome creates concentric weak areas in the material. And at some frequencies it no longer behaves as much like a piston. Some call this "breakup". The force from the voice coil is supposed to move most of the diaphragm as a piston with the outer suspension being the part that's supposed to flex. But when the diaphragm has weak areas, it effectively "breaks up" into a smaller diaphragm within the bigger one. The weak area acts as an unintended suspension.

 

That was a fabric dome treated with some sort of polymer. The AKG drivers are obviously made out of a different material--I suspect a polyester film like mylar or something similar. My guess is polyester is prone to a similar sort of weakening but I'm not a materials expert so I can't say for sure. But anything that alters the rigidity and flexibility of the diaphragm material *will* change the performance of the headphones. 

Hey, what kind of claptrap is this?  This is not news.  

 

I would not expect the dome to deform as shown in the video. That is obviously a design problem with the AKG driver.  The voice coil former should bottom out before the cone surface distorts as it did in the video.

 

It is recommended practice to turn down the volume all the way and then unplug the headphones before turning on/off the power on any amp.  This procedure has been recommended for the Schiit amps, including Asgard and Lyr.  The FACT is, the amps were designed for the best combination of objective and subjective performance.  To meet cost constraints, lines had to be drawn somewhere.  Turn on/off protection in the form of high quality relays could have been added for a cost.  I'm guessing that they could easily have added $20-$30 or more of retail cost to the amplifiers.  Any design is a set of carefully considered tradeoffs balanced against design goals and cost.  All three amps from Schiit have been designed for maximum sound quality at the most reasonable cost.  Jason please correct me if any of my assumptions are wrong.  Personally, I would rather have sound quality be prime design goal, with reliability bringing up a close 2nd.  I am very pleased with both my Asgard and Lyr.  Seems to me, Schiit have met and exceeded the aforementioned goals in their products so far.
 

 

New Posts  All Forums:Forum Nav:
  Return Home
  Back to Forum: Sound Science
This thread is locked  
Head-Fi.org › Forums › Equipment Forums › Sound Science › What causes this (amp related)?