Schiit Happened: The Story of the World's Most Improbable Start-Up
post-14115949
Post #30,376 of 58,324

Ripper2860

Headphoneus Supremus
Joined
Feb 24, 2018
Messages
3,176
Reaction score
7,290
Location
Texas, USA
Joined
Feb 24, 2018
Location
Texas, USA
Posts
3,176
Likes
7,290
I purchased the Eitr after I had the Mimby a few weeks. I initially purchased the Eitr to address an intermittent USB 'glitch' that would cause sound to stop and then restart with audible distortion. The Eitr did not resolve the issue (turns out it was my PC and a fresh reinstall of Win10 resolved the issue), however, the sonic benefit of having the Eitr was noticeable to me. I'll try to describe as best I can: it seemed to deliver music that was more 'musical'. Music had less glare with a bit more detail and more precise with crystalline highs (cymbals sounded like real cymbals). Music emanated from a deeper silence with less (or no) underlying noise. The instruments had more 'air' around them and more texture with better rendering of complex musical passages.

Now some may say BS and expectation bias, but remember, I originally purchased the Eitr to resolve what I thought was a USB problem and despite it not solving my issue (again, it was my Windows install), I kept it because of the improvements I heard. Of course YMMV, depending on your ears (or imagination), and equipment (headphones, amp).

Note: Some may ask why I went with Eitr to solve potential USB issue and not a Wyrd. Well -- I wanted a brushed aluminum case (Eitr) as opposed to steel panted case (Wyrd) - it was all about the looks. I wanted it to match my Magni 3 and Mimby. And Eitr gives me a essentially a Mimby w/ Gen5 equivalent USB -- isolation and re-clocking. :wink:

Hope this helps.
 
Last edited:
post-14115950
Post #30,377 of 58,324

OldRoadToad

500+ Head-Fier
Joined
Jan 30, 2007
Messages
863
Reaction score
1,175
Joined
Jan 30, 2007
Posts
863
Likes
1,175
OTOH, there's a hint in Mike's thread just now that something they would/could never do, in fact...

VU Meters, DAC/AMP/BBQ, QuadTube Powered® and RedLeds® with Toslink, Coaxial and USB inputs as well as a nice, power full phono pre-amp. They all ready use TOAD® - Toggle Only Actuated Device.

As for financing. I can not do that. I must have the money prior to purchase else wise, I can not truly afford it. This is but one reason I think so highly of the company. It is with in reach, even if your reach is limited. Yes, I use a credit card but must pay it off every month. My cars are 2011 and my motor bikes are from 2006 and 2007 respectively. I tend to keep things for a while. I have been looking at SUVs but the pain of purchase ($$$$$) has to date, kept me from buying.

I hear some say...“Ah, but a toad's reach should exceed his grasp, or what's a heaven for?” But Browning was not talking about credit card debt (or toads). A toad's reach should never exceed his income. No matter how much I may want (and that is all this is to me), I must first meet the needs of family then my self. After which, come my wants.

Schiit is affordable. If you wish to pay on "time", might I suggest all one need to is to pay themselves every month. I have a Harley piggy jar (can not find a toadish jar, dammit) and that is where I put all of my change and even some ones or may be fives. Until I have what I need to be able to pay for what I want.

ORT
 
post-14115964
Post #30,378 of 58,324

artur9

1000+ Head-Fier
Joined
Apr 25, 2015
Messages
1,185
Reaction score
830
Joined
Apr 25, 2015
Posts
1,185
Likes
830
Someone just accused me in a PM of being condescending. I don’t know how to respond to that except to sound condescending. But I refuse to apologize for my experience, my education, or my attitude. All are hard fought for and earned by years of struggle, and I didn’t have any daddy’s money to help me. :)

I guess bottom line is if you don’t like my posts I can do nothing about that and you are free to ignore me.
I think you need to offer up even more experience, education and attitude. And maybe find a new daddy with some money.

Other than that, I find your posts refreshingly straightforward and non-mystical.
 
post-14116016
Post #30,379 of 58,324

Letmebefrank

500+ Head-Fier
Joined
Apr 18, 2015
Messages
970
Reaction score
641
Joined
Apr 18, 2015
Posts
970
Likes
641
VU Meters, DAC/AMP/BBQ, QuadTube Powered® and RedLeds® with Toslink, Coaxial and USB inputs as well as a nice, power full phono pre-amp. They all ready use TOAD® - Toggle Only Actuated Device.

As for financing. I can not do that. I must have the money prior to purchase else wise, I can not truly afford it. This is but one reason I think so highly of the company. It is with in reach, even if your reach is limited. Yes, I use a credit card but must pay it off every month. My cars are 2011 and my motor bikes are from 2006 and 2007 respectively. I tend to keep things for a while. I have been looking at SUVs but the pain of purchase ($$$$$) has to date, kept me from buying.

I hear some say...“Ah, but a toad's reach should exceed his grasp, or what's a heaven for?” But Browning was not talking about credit card debt (or toads). A toad's reach should never exceed his income. No matter how much I may want (and that is all this is to me), I must first meet the needs of family then my self. After which, come my wants.

Schiit is affordable. If you wish to pay on "time", might I suggest all one need to is to pay themselves every month. I have a Harley piggy jar (can not find a toadish jar, dammit) and that is where I put all of my change and even some ones or may be fives. Until I have what I need to be able to pay for what I want.

ORT
Here ya go bud!

 
Last edited:
     Share This Post       
  • Like
Reactions: OldRoadToad
post-14116100
Post #30,380 of 58,324

wink

His amps are made out of recycled beer cans
and his source from tomatos.
Joined
Apr 13, 2009
Messages
20,298
Reaction score
3,509
Joined
Apr 13, 2009
Posts
20,298
Likes
3,509
We've been solicited by companies who could make financing an easy option for buyers (like Affirm.) It wouldn't expose us to any risk. I'm sure it would cost us some percentage, but we haven't engaged with the potential financing companies enough to know how much.

So, if we have some relatively easy and turn-key options to offer financing, why aren't we interested?

In short, one word: philosophy.

Or maybe two words: philosophy and cussedness.

Because, you see, I'm a bit uncomfortable in providing an option that makes it easier for our customers to get into debt. I don't think holding debt is a good thing, and I don't really want to encourage it. Especially not for audio gear. If our customers want to get into debt, there are tons of credit cards out there that would be more than happy to help. But Schiit didn't dangle the card in front of their face. Schiit didn't help them go into debt, or go deeper into debt.

Crazy? Old-fashioned? Possibly. But I sleep better at night.
And the very worst of financing is that it is inherently inflationary and causes untimely demand. How many folks buy cars thinking oh, it is only $xxx.xx per month. Fits right into your budget, and allows the dealer to ramp down discounts and add unwanted accessories. Trouble is, the $xxx.xx per month is now getting up to 84 months. You really want to keep your car 7 years? Warps the market for buyers and sellers. Credit card companies can allow the user years and years to give them $10,000 or more for an Yggy at up to 30% interest. Really? The only advantage is that it makes the buyer subject to living beyond his means. Real estate is another example. What we do not realize here is that much of the world has no mortgages which tends to hold the price quite low. In the case of New Zealand or Argentina, it can be less than half. Besides all of that, we are schitty bankers with no experience. Financing, bad for buyers, bad for us. Naaah, we won’t be doing it.
...
 
     Share This Post       
  • Like
Reactions: Ripper2860
post-14116112
Post #30,381 of 58,324

AudioBear

Headphoneus Supremus
Joined
Mar 3, 2007
Messages
1,656
Reaction score
1,020
Joined
Mar 3, 2007
Posts
1,656
Likes
1,020
Someone just accused me in a PM of being condescending. I don’t know how to respond to that except to sound condescending. But I refuse to apologize for my experience, my education, or my attitude. All are hard fought for and earned by years of struggle, and I didn’t have any daddy’s money to help me. :)

I guess bottom line is if you don’t like my posts I can do nothing about that and you are free to ignore me.
Geez, I liked your original post so I gotta like this one I guess. You do not sound condescending to me but if the person who PM-ed you and said that feels that way they should just ignore you. It's how they feel, can't argue with that. Keep posting as you always have, a straight shooter.
 
Last edited:
     Share This Post       
  • Like
Reactions: RCBinTN
post-14116115
Post #30,382 of 58,324

wink

His amps are made out of recycled beer cans
and his source from tomatos.
Joined
Apr 13, 2009
Messages
20,298
Reaction score
3,509
Joined
Apr 13, 2009
Posts
20,298
Likes
3,509
VU Meters, DAC/AMP/BBQ, QuadTube Powered® and RedLeds® with Toslink, Coaxial and USB inputs as well as a nice, power full phono pre-amp. They all ready use TOAD® - Toggle Only Actuated Device.

As for financing. I can not do that. I must have the money prior to purchase else wise, I can not truly afford it. This is but one reason I think so highly of the company. It is with in reach, even if your reach is limited. Yes, I use a credit card but must pay it off every month. My cars are 2011 and my motor bikes are from 2006 and 2007 respectively. I tend to keep things for a while. I have been looking at SUVs but the pain of purchase ($$$$$) has to date, kept me from buying.

I hear some say...“Ah, but a toad's reach should exceed his grasp, or what's a heaven for?” But Browning was not talking about credit card debt (or toads). A toad's reach should never exceed his income. No matter how much I may want (and that is all this is to me), I must first meet the needs of family then my self. After which, come my wants.

Schiit is affordable. If you wish to pay on "time", might I suggest all one need to is to pay themselves every month. I have a Harley piggy jar (can not find a toadish jar, dammit) and that is where I put all of my change and even some ones or may be fives. Until I have what I need to be able to pay for what I want.

ORT
..
.
 
     Share This Post       
  • Like
Reactions: OldRoadToad
post-14116350
Post #30,383 of 58,324

Pietro Cozzi Tinin

Headphoneus Supremus
Joined
Jan 26, 2017
Messages
3,375
Reaction score
7,041
Location
Delft, The Netherlands
Joined
Jan 26, 2017
Location
Delft, The Netherlands
Posts
3,375
Likes
7,041
Age
56
I've never found you supercilious or imperious.:darthsmile:

You are the go to technical voice who dutifully answers some of the same questions over and over again....and you probably don't even own a cat.:)
Since he is a member of our secret sub community "the outhouse" Constanza is family too.
 
     Share This Post       
post-14116504
Post #30,384 of 58,324

JohnnyCanuck

500+ Head-Fier
Joined
Jan 29, 2014
Messages
735
Reaction score
1,303
Location
Ottawa, Canada
Joined
Jan 29, 2014
Location
Ottawa, Canada
Posts
735
Likes
1,303
Schiit is affordable. If you wish to pay on "time", might I suggest all one need to is to pay themselves every month. I have a Harley piggy jar (can not find a toadish jar, dammit) and that is where I put all of my change and even some ones or may be fives. Until I have what I need to be able to pay for what I want.

ORT
You can find your toad 'piggy bank' here.

You're welcome. :)

JC
 
     Share This Post       
  • Like
Reactions: RCBinTN
post-14116519
Post #30,385 of 58,324

KoshNaranek

1000+ Head-Fier
Joined
Jun 10, 2016
Messages
1,128
Reaction score
2,292
Location
Off Planet
Joined
Jun 10, 2016
Location
Off Planet
Posts
1,128
Likes
2,292
Someone just accused me in a PM of being condescending. I don’t know how to respond to that except to sound condescending. But I refuse to apologize for my experience, my education, or my attitude. All are hard fought for and earned by years of struggle, and I didn’t have any daddy’s money to help me. :)

I guess bottom line is if you don’t like my posts I can do nothing about that and you are free to ignore me.
I always interally voiced you as the curmudgeonly, experienced engineer (my father). This has worked well for me so far. Since I get along quite well with my father, I have no problem.

Just be yourself, you are loved and respected
 
Last edited:
post-14116662
Post #30,386 of 58,324

Rensek

500+ Head-Fier
Joined
Jan 29, 2018
Messages
961
Reaction score
2,139
Location
US Midwest
Joined
Jan 29, 2018
Location
US Midwest
Posts
961
Likes
2,139
A question for Jason, a week or three ago you mentioned we should see increased production on stuff as you were implementing changes to speed things up. You mentioned overseas distributors we're going to get precedence over Amazon (which I take no issue with). What are your plans for Amazon? Would love to be able spend my rewards on Schiit.
 
Last edited:
     Share This Post       
post-14116732
Post #30,387 of 58,324

rlow

100+ Head-Fier
Joined
May 3, 2017
Messages
318
Reaction score
472
Location
Canada, 'eh
Joined
May 3, 2017
Location
Canada, 'eh
Posts
318
Likes
472
New products on the Schiit site this morning.
Lyr 3, multibit DAC module for Jotunheim and Lyr 3.
 
     Share This Post       
post-14116733
Post #30,388 of 58,324
Joined
Jun 9, 2010
Messages
2,652
Reaction score
17,586
Location
Newhall, CA
Joined
Jun 9, 2010
Location
Newhall, CA
Posts
2,652
Likes
17,586
Website
www.schiit.com
2018, Chapter 5:
Where Do I Start?


Hoo boy. This is gonna be interesting. Because the first thing that comes to mind is the title above. Where do I start? How do I explain it? How do I then go on to enumerate just how different, and just how crazy, this whole new thing is?

For those of you who missed the press release, I’m talking about Lyr 3.

And, yeah, Lyr 3 is different. We basically took the Lyr and Lyr 2 design and nuked them from orbit. Not a single shred of the original Lyr remains. Dynamically adaptive output stage? Nope. Two noval triodes? Nope. A single-purpose headphone amp? Nope. A U-shaped chassis? Nope. Gone, gone, gone, and gone.

What’s the same? Both the old Lyrs and Lyr 3 have the same power rating. Both designs have two transformers (but radically different types in Lyr 3, C-core transformers with 50% higher VA rating). And yeah, both designs have power switches on the back.

But that’s about it. On first glance, Lyr 3 looks more like a hybrid of Saga and Jotunheim. It takes a single 6SN7-type tube, and uses the same modular architecture as Jotunheim.

“Oh holy hell, I bought like $4,742 in tubes for my Lyr! Why the hell’d you pull the rug out from under me now?”

Well, in short, the world is changing. You may have noticed that those expensive tubes for your Lyr keep getting more and more expensive. You may have noticed that really good tubes are getting harder to find. And you may have noticed that matching a pair of tubes can be a monumental pain in the rear end.

So yeah, the world changed, and so we changed with it. That’s why Lyr 3 has only a single tube. Because it’s a lot easier to get one great tube. Because there are good new-production 6SN7s out there. And because octals sound better.

Aside: at least they sound better to us. Some people think we’re crazy. Or evil. Or both. YMMV.

But that’s really only the start of the story. Because Lyr 3 is now not just a headphone amp. And it introduces two new technologies that may be as important as multibit DACs, time- and frequency-optimized digital filters, and Pivot Point.

But before we talk about modularity and new tech, let’s talk a bit more about the rationale that led down, well, a completely new path.


The Fork in the Road

As soon as we introduced Jotunheim, the requests started pouring in:

“I’d really really like Jotunheim with tubes!”

“Jotunheim would be perfect with tube gain!”

“Can you do a version of Jotunheim with tubes, remote control, resistor attenuator like Ragnarok, and a multibit DAC, for $499?”

Okay, so maybe I’m exaggerating a bit on the last one, but you get the picture. People wanted a Jotunheim with tubes. A Lyrenheim, you might call it.

But, for one simple reason, that wasn’t gonna happen. Put two tubes in a Jotunheim, like Lyr, and you’re in instant meltdown. The extra power supply to run the heaters and tubes would put it into thermal overload.

I mean, hell, Lyr 2 runs plenty hot, even with single-ended output. And Mjolnir 2 is plenty, plenty toasty even with a much larger chassis.

And Jotunheim was already pretty packed. Even if we solved the thermal problem, there wasn’t a ton of room for the second transformer we’d need for the tubes.

But still, we got inquiries.

“Oh please oh please make a tube Jotunheim!”

“My wallet is just waiting for a Tubenheim!”

Heck, some people even started using Saga in front of Jotunheim to get a little tubey-ness in there. But that was a clumsy, two-box solution. Three if you added a DAC. And that still didn’t solve the problem of making a balanced tube hybrid in a Jotunheim-sized box that didn’t melt down.

At the same time, though, a couple of things were converging.

One was slowing sales of the Lyr 2. The original Lyr had ridden the wave of hard-to-drive orthos and sold very well. Lyr 2, while a more flexible amp, was still more in ortho-world than in IEM-world…and the ortho world was moving on to easier-to-drive options. The time for the original Lyr architecture seemed to be, well, ending.

The second was our experience with octal tubes, and single-tube designs. Saga proved that we could get really nice performance out of a single octal tube (albeit as a buffer), and Vali 2’s odd combobulated tube/transistor architecture made a really nice hybrid (with gain), even if it had some compromises like capacitor-coupled output. Both of those facts kept rolling around in my mind.

And we had the Jotunheim modules. I knew we even had a multibit DAC coming. Jotunheim with the multibit DAC prototype was very nice. But a tube hybrid with a multibit DAC, in one chassis…

Maybe we could do something with a single tube. Something like Vali 2, but with proper bipolar supplies and no coupling capacitors.

And maybe if we didn’t worry about balanced. Just go single-ended, like the original Lyr and Lyr 2. Hell, it made it easier for the owner anyway. No need for custom cables to get full power out of the design.

Aside: there are many days I regret doing anything with balanced. It’s confusing for the customer, and can lead to intense audio nervosa. If there weren’t real advantages to having hardware-balanced DACs, I’d seriously consider not doing balanced at all. No, really. I’ve seen far too many people go crazy trying to root out all single-endedness in their system. I also see far too many products out there that simply use balanced connectors, but aren’t balanced internally at all. Balanced done right can be spectacular, but I don’t consider balanced to be a requirement for a truly high-end system. But then again, I usually listen to Modi Multibit and a Vali 2, so maybe I’m a philistine. Yeah, I have a Gungnir Multibit and Mjolnir 2, but those are big, big products and I’ve been spending less time on the couch and more time at the desk.

Aside to the aside: although it may sound like a sales pitch, I will probably be switching to Lyr 3 with a Multibit DAC card for primary desktop listening.

But, you know what, those two thoughts above (single tube, no balanced) were enough to get me going. Lyr 2 had rail voltages that would be usable to run Jotunheim cards, so the modular thing was feasible.

The only thing was, would the Multibit DAC card fit? A single octal tube punches a pretty big hole in the chassis.

Luckily I had a Saga and a Multibit DAC prototype. Laying the DAC inside the chassis, it just barely fit (about 1/10 of an inch between it and the tube socket.)

I grinned. This could work. This could really work!

It was only later that the nightmares began.


There Are Doubts

Even before I sent the first Lyr 3 board out—which never fully worked, but we’ll get to that—I started having doubts. Big doubts. Bigger than I’ve ever had when working on a next generation product.

I mean, yeah, kinda makes sense when you’re considering such a huge departure. But honestly, I was worried less about Lyr 3 than the fallout that might come after it.

I mean, what would happen when people immediately started asking about Asgard 3? I had no idea if we’d be able to dissipate enough heat to do a real Class A amp in a Jotunheim-style chassis. And I had no idea if there was still a place for a low-power Class A amp in a new modular lineup, in any case.

Or, worse, what would happen when they started asking about Valhalla 3? Because that was even worse than Asgard 3. Valhalla 3 was a complete non-starter. With 4 tubes taking up most of the space in the chassis, a Multibit DAC card wouldn’t fit. And the fact that it didn’t have the voltage rails to run any card at all is, well, kinda-sorta problematic.

Which meant that if we did a Lyr 3, we might end up with two amps stuck permanently in the past. Asgard 2 and Valhalla 2 might just have to soldier along, unchanged, into a brave new world where we were gettin’ real modular.

Or, worse, Asgard could be updated, but Valhalla 2 would be stuck in the past. Eeek.

Or, even more horrible, Valhalla 3 could be cobbled up…but it would only take the 4490 DAC and phono cards, not the Multibit DAC. Oh holy hell, how do you explain that one?

So yeah, you could say I had some dyspepsia.

And that was before I got into the actual design of Lyr 3. It turns out that a simple idea—improving the topology we first experimented with in Vali 2—wasn’t a simple idea at all. Vali 2 gets away with a lot of shortcuts, because it has a very simple power supply, and it uses output coupling capacitors. Getting rid of those capacitors—and dramatically improving the tube/transistor hybrid gain stage—ended up requiring changes of epic proportions. In fact, the whole hybrid gain stage had to be re-architected completely, around a wacky operational point servo that kept everything in balance between the tube and transistor. From there, I used a fairly conventional darlington BJT output stage…which would work, but didn’t sit right with me. It was, well, maybe a bit too conventional. Too much of a sidestep from the old Lyr’s sliding bias “dynamically adaptive” output stage. Maybe. But I needed something that would work, and a conventional output stage would definitely work.

There was also the problem of heat. I intended on running fairly high bias through the output stage, and I needed to reject the heat. Previous designs used a U-shaped aluminum chassis as a heatsink, with parts bolted directly to the chassis. This design was going to use a Jotunheim style “sled and L” chassis, with over half of the chassis being steel (and therefore not a good heat conductor.)

I needed a way to move the output stage heat to the top of the chassis. At first, I envisioned doing an L-bracket and bolting it to the top chassis, which would work, but would also prove tricky in assembly, and would complicate service and repair. Still, I thought that was really our only option, so that’s what I went with.

When the first prototype boards showed up, I didn’t have any chassis parts, much less the L-bracket needed to connect to the top chassis. I did, however, have junk Magni tops. I drilled a few holes in one of them and bolted the outputs through it. It would provide plenty of heatsinking for short runs of the prototype design, I figured.

Alex, of course, saw the Magni chassis sticking up off of a Jotunheim-sized board, and was amused. “Find another use for our B-stock?” he asked.

“Hopefully not,” I told him.

“Still, you could screen the name on it, sell it open-frame…” he teased.

I shuddered. Not with 200V running around on the board.

But when I got around to powering up the boards, I realized I had real problems. Real problems, like smoke and flames. As in, parts literally catching on fire.


And There Is Fire

Yes. Smoke and flames.

Now, my prototypes are sometimes bad, but they usually aren’t that bad. Smoke and flames means something is VERY wrong. I went back over the schematic, checked that I’d used the right parts, and shook my head. Everything seemed OK. The design should work. Or, well, it shouldn’t catch on fire.

I replaced the parts, made sure they were correct, and brought up the power realllllll slow…and as soon as the tube started glowing, poof, smoke. Again. Sigh.

The next rebuild, I changed some resistors to limit the amount of current that could flow through the mirror. This time it didn’t smoke, but it drew 200V of power supply down to nothing.

Yeah. That’s a problem.

I eventually traced it to a bonehead mistake that I made, miscalculating the operating point of the 6SN7 and thinking I needed a more complicated power supply with small negative rails to get it in the right region. What happened is that I got larger negative rails than I expected, and a tube trying to run 100mA. Which meant the current mirror tried to run 100mA. Which is about 20W on a 200V rail. Through very small devices. Yeah. No go.

I eliminated the negative supply and the tube ran at 5mA per side, like expected. Much better. But by then the board was junk. It was time to order a second prototype.

The second one ran, but had problems with noise. Noise isn’t a surprise in early prototypes, but the amount of noise was somewhat suprising. It also, oddly, didn’t really sound like I thought it should. There was something missing.

So, I did what I usually do when faced with a prototype impasse: I put it on the shelf and moved on to other things.


The Epiphany: Continuity

Even as I worked on other projects, though, the problems with Lyr 3 bugged me. The noise could be conquered; it was certainly a problem with the current sources driving the bias servo, or a problem with the tube heater supplies, or a bit of both.

But the sound...

That was the problem. It should sound better.

And, going from the innovative-but-flawed “dynamically adaptive” output stage of Lyr 2 to something totally conventional, well, kinda irked me. It was like giving up. Yeah, the old stage had its problems, but the new one, was, well...a little too vanilla.

I don’t know when the epiphany came, but when it did, it seemed totally obvious: do a constant transconductance output stage.

Constant transconductance was different.

Constant transconductance solved a problem baked into every Class AB output stage.

Constant transconductance made sense.

Now, I can’t claim that it was my brilliance that led to this realization, however. Anyone who has Bob Cordell’s Designing Audio Power Amplifiers book, like me, knows that nonlinear transconductance in the crossover region of a Class AB amp is a real problem, one that’s simply there, like a stain on a wine-glass, or a feral cat prowling on the edges of a fancy restaurant’s patio. Most Class AB amps simply resort to optimum biasing to keep the nonlinearity to a minimum (but it is still there), or rely on tons of feedback to squash it flat (but it is still there.)

So I knew about the problem because of Bob Cordell. But I was aware of it even before that, during my time at Sumo. Sumo used a complex Hawksford error-correction scheme (also refined and implemented by Cordell) to compensate for transconductance nonlinearity in its MOSFET-output power amps. In the Sumo amps, the error feedback system was hand-adjusted on every amp for maximum null. It worked pretty well...but it was complicated and sometimes unstable.

But it was another designer, John Broskie, of TubeCad.com, who brought the issue of transconductance doubling to the fore: https://www.tubecad.com/2015/10/blog0331.htm

Broskie, who I’m convinced has more interesting ideas before breakfast than, well, all other audio designers put together, took Cordell’s ideas and ran with them. Ran, as in, he proposed a good half-dozen solutions to the problems of transconductance doubling. Some used diodes to decrease the emitter resistance once the device turned on, others used complementary devices, some used Class A and Class C stages in conjunction with Class AB to get linear transconductance.

I remember trying a couple of his designs (on the Magni 3 prototype, of all things) and finding that they didn’t measure all that different, so I decided that maybe the math was a little wonky. I put that one aside as well.

But, when I came back to Lyr 3, everything collided. And that’s how I knew that we needed to do a constant transconductance output stage. And I knew there were many ways to skin that cat. I just didn’t know which one.

As usual, when confronted with something like this, I went to the breadboard. Eventually, I had something that wasn’t quite Broskie, and wasn’t quite Cordell, but achieved the goal: a meaningful linearization of transconductance. What’s more, it was measurable. The stage measured 10-15db better distortion above the crossover point than the standard output stage.

And, even more interesting, it used NPN and PNP parts on both rails, and both were on at any given time. This is important, because NPN and PNP parts don’t exactly match, so if you can use both at all times, you’re also reducing potential nonlinearities.

With that new output stage in hand, I went to a third prototype.

But...

But hell, a constant-transconductance output stage uses a lot of parts. And a lot of big resistors. If I put them on the bracket I intended to use...well, hell, the bracket wouldn’t be big enough.

I needed a different solution. In fact, it kinda sucked that I couldn’t just bolt the devices to the bottom chassis, like the old U-shaped chassis we used in the past. Even if I did the bottom out of aluminum, bolting the parts down would require a weird-ass board, and the mechanical aspects of the output stage would interfere with the cards.

Unless...unless I could put all the parts on the bottom of the board.

But how would I get the heat out of them? I remembered something I’d read about—and dismissed—a long time ago. Gap pads. Gap pads are like sil-pad thermal pads, the same thing we use to thermally couple transistors to heatsinks al over the place, but thick. Like, thick enough to go under a board and draw heat away from a whole lot of devices.

But would it work? No clue. But hey, this was still a prototype for a product that might never exist, so why not try it? I ordered some cut-to-fit gap pad material, put all the devices on the back of the board, and crossed my fingers. I didn’t really have the right kind of chassis for this, but when the prototype came in, I just used a C-clamp to hold the board down onto a hunk of aluminum.

And it worked.

And this one sounded right.

I mean, really, really right.

One of our early “dog-eared” listeners, when confronted with a number of prototypes, including Lyr 3, immediately keyed on it. “Holy hell, this thing is great,” he said. “Forget the rest, just do this one.”

And, even though he was a bit hyperbolic about it, I got where he was coming from. He’d heard much more expensive (proposed) stuff. But Lyr 3 hit a sweet spot. It was an exceptional combination of tube and transistor capability—the quintessential hybrid.

That third prototype, though, still had a problem.

It was (still) too noisy, which was pretty amazing considering the amount of filtering and bypassing I’d thrown at it. What’s worse was that the solution to the noise wasn’t exactly obvious. The main rails were quiet. The current sources running the bias servo were filtered with 5 poles (!)—one of the main reasons this amp has a nearly 1-minute-long start-up sequence.

Which left one thing: tube heaters.

I really, really didn’t want it to be the heaters. I was using AC heaters. AC is usually fine (when done right)(with elevated supplies). But AC was a step back from Lyr 2. Lyr 2 used DC. But Lyr 2 also ran insanely hot...and in large part because it used DC. The heat we had to throw away in regulation was large and painful.

Aside: One of the goals for Lyr 3 was to have it, ah, be a lot less like a barbecue than Lyr 2. Lyr 2 draws 38W from the AC socket, all the time, every time. That’s 38 watts it has to dissipate through the chassis. That’s a hot chassis. In contrast, Jotunheim pulls 16 watts, all the time. That’s a whole lot less. Lyr 3 does 25W. Not as good as Jotunheim, but a hell of a lot better than Lyr 2.

I spliced a DC supply into the heaters, and presto: 20dB less noise.

Aside: 20dB is “a metric ****ton.”

crap.

What I needed was DC heaters, without the regulation. Almost everyone in tube-land says “don’t use unregulated DC, it might be worse than AC.”

But, you know what? What “everyone” says may be wrong.

I tried a simple unregulated DC supply. Boom. 18dB less noise.

2dB higher than DC? For, what, 6W less dissipation? Sold. Experts be damned. Lesson learned: try stuff in your own application, before instantly dismissing it. If I’d done this in Lyr 2 and Mjolnir 2, they’d run a lot cooler. Dummy. I believed the “prevailing wisdom.” Don’t. Trust, but verify.

So, did we have a product?

In short, not really.

Not until we went through three different bottom chassis, anyway. I had three chassis made up to test thermal qualities:
  1. One in steel
  2. One in 0.050” aluminum
  3. One in 0.063” aluminum
I hoped the steel would work, but I didn’t expect it to. And it didn’t. Using a gap-pad to draw heat out of devices onto a steel chassis doesn’t get you far, because steel is a poopie conductor of heat. Sad, because all of the engineering on Jotunheim, including the slide-lock top hardware, assumed 0.048” steel.

I wanted the 0.050” aluminum to work, because the slide-lock top hardware would probably work with it. But, while it worked OK on the heat side, it was really flimsy. It felt cheap. It wasn’t acceptable.

The 0.063” aluminum meant we needed to mill out pockets so we could use the slide-lock hardware for the sleek no-fasteners look of Jotunheim, but it worked the best, so that’s the way we went. Sigh. Expensive. But sometimes expensive is right, as in this case.

Now, we had a product.


Introducing Coherence and Continuity

As I lived with the prototype for a while, I began realizing how unique this product was.

It used a tube hybrid stage that required no interstage coupling capacitors. Hell, it was DC in and DC out. Try that with tube hybrids. It’s not easy.

It had a constant-transconductance output stage, which (I believe) is unique.

And that’s when I started thinking: we really should name these things. Not silly names, not inconsequential names, but names with some weight and meaning. Call this my marketing background catching up with me, but I think it’s important.

Aside: as I think that going back and naming our true Multibit and time- and frequency-domain optimized digital filters is important. But I’ll get to that.

And that’s why I spent a weekend day thinking of what these technologies were, and what they represented. And writing down dozens of words that came to mind.

And here’s where I ended up:
  • Coherence™. Because in this topology, tube and bipolar junction transistor work seamlessly together. One does not work without the other. And neither work without the unique bias servo that also sets DC offset.
  • Continuity™. Because, like it or not, every other Class AB output stage is discontinuous. Period. Our unique output stage maintains the transconductance doubling in the crossover region as far out as the device can go. Not only that, it blends both NPN and PNP output so that both devices are on at any given time. It’s our best shot at a truly linear and continuous output stage.
Both of these were really, really key new technologies.

Coherence, the tube hybrid stage, yeah, while we began the first exploration of it with Vali 2, really didn’t come into its own until we eliminated the compromises in the circuit with Lyr 3. Vali 2 maybe isn’t really a Coherence stage at all, or maybe Coherence Lite. What’s sure is that you’ll be seeing a whole lot more Coherence through the Schiit line. Hell, maybe it’s time to give up on non-hybrid technologies completely, since we’re clearly better at putting tubes and transistors together, than doing pure tube amps.

Continuity, the constant transconductance output stage, you’re not going to find anywhere else. John Broskie actually brought Bob Cordell to Schiit’s room at RMAF 2017, where we talked about constant transconductance output stages (Bob will probably have lots more to say about these in the near future, though it’s not my place to say what). When I mentioned “Oh yeah, we’re introducing something like that, in early 2018,” he was visibly startled. Nobody expects anything like Continuity out there. Again, you’ll probably be seeing more of this in our line.

And, you’ll probably be seeing more of us naming our technologies that we consider important...that is, unless you guys think this is stupid. And, don’t worry—we won’t go overboard into icon-and-pseudoscience-hell.


Sanely Modular

Some of the sharper-eyed readers will have noticed that, in our press releases about Lyr 3 and Jotunheim Multibit, we’ve mentioned that the same modules are used in both Lyr 3 and Jotunheim. In short, yes they are.

Go back and read that again.

To highlight: both Lyr 3 and Jotunheim use the same optional modules.
  • Same Multibit DAC.
  • Same 4490 DAC.
  • Same phono.
Same across the board. This isn’t an accident. This is by design. We intended both products to use the same cards, because that’s what makes sense. It’s efficient. It reduces the number of product variations we have to deal with. It makes our life easier, which means you benefit because there’s less possibility for mistakes (like, the wrong card in the wrong product.)

Sounds simple? Not really.

In fact, it took an entire re-engineering of the 4490 DAC card to make sure it worked in both products. Hell, we even had “4490 single-ended” and “4490 balanced” cards before we came to our senses and consolidated them.

Aside: Lyr 3 is a single-ended amp, so it simply ignores half of the balanced signal coming off of the cards; all cards have inherently balanced output.

In terms of the 4490 card, this changed everything. Initially intended only for Jotunheim with balanced outputs and passive summing, it didn’t work for Lyr 3. Like, at all. Which meant that if we wanted it to work for both products, it would need a complete redesign.

So that’s what we did. We went from the original’s passive summing to active summing using the OPA1662, and went from the original’s single-pole passive filter to an active two-pole filter.

Now, some people may think this is a less “pure” approach, but the proof of the pudding is in the eating, or, in this case, hearing. And the new 4490 card sounds significantly better. It’s now standard in both Jotunheim and Lyr 3.

In the case of the Multibit card, you won’t be surprised to know that it uses the same AD5547 DAC as the Bifrost Multibit and Modi Multibit, together with the same Analog Devices SHARC implementing our own frequency- and time-domain optimized digital filter, as with all of our other Multibit DACs.

What’s more, these three cards are a great foundation to build on. Don’t be surprised if future products—even more expensive products—use these cards.


About The Future...

Okay. So we now have a Lyr 3. And a Jotunheim. Both are great choices for the desktop. They’re a sane size, and they’re each a one-box solution. But when you’re staring at $599 for a Jotunheim with Multibit card, or $699 for Lyr 3 with Multibit, do you start wondering “is there a more affordable option?”

Specifically...something like a $249 Asgard 3, which could take the same modules and be a $449 product with the Multibit card.

Wait. Before you answer, consider:
  1. If we do an Asgard 3 like this, it won’t be a 100% class A amp like the original Asgard and Asgard 2.
  2. However, it will be a lot more powerful—think 3.5W RMS or so per channel.
  3. It will also have the Continuity output stage, so Class A is less, ah, meaningful than before. That is, if you believe our marketing blather.
And also consider: there will never be a Valhalla 3 like this.

So there’s no Valhalla 3 option. Just Asgard 3. At $249. A lot more powerful than the old Asgard. Not Class A (but pretty damn high bias, like Class A for 1/2W or so, and with the Continuity output stage on top of it.)

So, I’m asking you: does a product like this proposed Asgard 3 make sense?

Or not?

Why?

Your answers will help shape the future.
 
post-14116751
Post #30,389 of 58,324
Joined
Jun 9, 2010
Messages
2,652
Reaction score
17,586
Location
Newhall, CA
Joined
Jun 9, 2010
Location
Newhall, CA
Posts
2,652
Likes
17,586
Website
www.schiit.com
Oh, and two other things:

1. If you have an order in for a Jotunheim, don't panic--they're back in stock as of today, and we'll be contacting you to see if you want to change your order to the Multibit DAC module or move to Lyr 3.
2. Mike and I will be on video at 2PM PST this afternoon, talking about Lyr 3 and Jotunheim Multibit. It'll be on our Facebook page (and later, probably elsewhere, if you don't want to use Facebook.)

sometimes.jpg
 
Last edited:
post-14116767
Post #30,390 of 58,324

decodm

Head-Fier
Joined
Sep 19, 2017
Messages
79
Reaction score
55
Location
Brazil
Joined
Sep 19, 2017
Location
Brazil
Posts
79
Likes
55
That's some nice news, @Jason Stoddard !!

I assume both are gen 5 USB?
 
     Share This Post       

Users Who Are Viewing This Thread (Users: 31, Guests: 70)

Top