Schiit Happened: The Story of the World's Most Improbable Start-Up
post-15432720
Post #56,131 of 58,195

Pietro Cozzi Tinin

Headphoneus Supremus
Joined
Jan 26, 2017
Messages
3,372
Reaction score
7,033
Location
Delft, The Netherlands
Joined
Jan 26, 2017
Location
Delft, The Netherlands
Posts
3,372
Likes
7,033
Age
56
Finally found someone happy with those bleeding Grados.

IMG_20200123_132123.jpg

Now please @CAPT Deadpool, read your PM's.
 
post-15432728
Post #56,132 of 58,195

Pietro Cozzi Tinin

Headphoneus Supremus
Joined
Jan 26, 2017
Messages
3,372
Reaction score
7,033
Location
Delft, The Netherlands
Joined
Jan 26, 2017
Location
Delft, The Netherlands
Posts
3,372
Likes
7,033
Age
56
Appears to be a Tablet Mark I.
Portable model. Secure erase technology. Day/night display with AlwaysOn capability.
And still in operating condition.
Impressive.
You're correct except that Mark wasn't born yet.
 
     Share This Post       
  • Like
Reactions: yonson and bcowen
post-15432758
Post #56,133 of 58,195

audio philestine

100+ Head-Fier
Joined
Feb 16, 2018
Messages
273
Reaction score
821
Location
Western MA, USA
Joined
Feb 16, 2018
Location
Western MA, USA
Posts
273
Likes
821
Oh kids, it goes much further back... Did a lot of my early programming in Algol-60 on an Elliot 4130. Graduated to PDP-10 (various languages/OSs) Then C and PDP-11 Unix V7, then VAX 11/750 with 4.1 BSD. Then Symbolics 3600 LISP machines. Some VAX 11/780 VMS as well. Then Sun-2 Motorola 68010. Side-trip to Mac 512 and pretty much every major Mac hardware and software issue since. Then SGI Irix. Some WindowsNT C++. Then lots of Java on Sun and Intel-Linux desktops and servers. And it goes on...
That's some nice computer pr0n there.... I'll be in my bunk :wink:
 
post-15432761
Post #56,134 of 58,195

bcowen

Headphoneus Supremus
Joined
Jan 3, 2018
Messages
3,960
Reaction score
6,559
Location
North Carolina
Joined
Jan 3, 2018
Location
North Carolina
Posts
3,960
Likes
6,559
     Share This Post       
post-15432762
Post #56,135 of 58,195

bcowen

Headphoneus Supremus
Joined
Jan 3, 2018
Messages
3,960
Reaction score
6,559
Location
North Carolina
Joined
Jan 3, 2018
Location
North Carolina
Posts
3,960
Likes
6,559
Ya know, that's the entry level tamper.

You should go big or go home, and go for the unobtanium option (only available on special order, please consult the factory for lead dates).

JJ :ksc75smile:
You have an unobtanium one?!? Totally not fair. I've had one on order for what seems like forever, but the lead time is still posted as infinity. :confused:
 
     Share This Post       
  • Like
Reactions: johnjen
post-15432795
Post #56,136 of 58,195

Brubacca

100+ Head-Fier
Joined
Jun 5, 2014
Messages
396
Reaction score
494
Joined
Jun 5, 2014
Posts
396
Likes
494
Whichever power amp you choose will sound better with a proper pre-amp.
JC
Right now I am using the pre section of the Naim. The SYS has the volume all the way up so it can just be a switchbox.

Do you think the SYS is degrading the sound in this configuration?
(Also yes my intent would be to use the volume of either the SYS or Freya in the new configs, the Naim has a AV Bypass function)
 
     Share This Post       
post-15432893
Post #56,137 of 58,195

sam6550a

500+ Head-Fier
Joined
May 23, 2016
Messages
758
Reaction score
1,236
Location
Rural South Georgia
Joined
May 23, 2016
Location
Rural South Georgia
Posts
758
Likes
1,236
Looks like decision time is here.... I sold off a pair of emotiva amps that I never liked all that much.

I now have some funds burning a hole in my pocket...

My sources are Gungnir MB and Rega P2 w/ HanaEL. These are switched with a SYS into a Naim UnitiQute...

Really looking at two choices here:
Get Freya (S or +) and use Naim Amp.
Get Aegir and use Sys as pre-amp.

Which do you think would come out sounding better?
Gungnir MB-->SYS-->Aegir is my recommendation.
 
post-15432895
Post #56,138 of 58,195

macdonjh

100+ Head-Fier
Joined
Sep 16, 2015
Messages
474
Reaction score
334
Joined
Sep 16, 2015
Posts
474
Likes
334
At one of the earliest Stereophile shows I heard the then recently introduced Martin-Logan CLS electrostatic loudspeakers.

Regarding CG: I grew up watching the original Star Trek series. In black and white, mostly. But, it was in reruns almost constantly, so it is in my bones.
In 2006 remastered versions with new CG were made. I'm sure they were thinking they were improving on the cheesy 1960's special effects. Now when I watch the new versions what I see is a mix of charming 1960's era effects (achieved with a limited budget and vintage tech), and cheesy 2006 era CG.
I guess "cheesy" is subjective.

Oh well. 30-40 years ago they colorized B&W movies, because few under 30 or something wouldn't watch them. In Suddenly (1954) staring "old blue eyes," Frank Sinatra, they gave him, you guessed it, brown eyes.
Martin-Logan CLS, oooh, I've always wanted a pair. My college graduation present to myself (after a couple of years' savings) was either Apogee or Martin-Logan: Apogee won.

I'm not a big fan of redoing other people's work. Leave Star Trek alone. Don't colorize classic movies. And, please Hollywood, no more remakes. Somebody, somewhere, has to be writing original stories that are worth bringing to the screen.

I hope Tammy Wynette doesn't suffer the same fate as Frank. The lyric just wouldn't work in reverse.
 
     Share This Post       
  • Like
Reactions: bcowen and adydula
post-15432904
Post #56,139 of 58,195

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
Looks like decision time is here.... I sold off a pair of emotiva amps that I never liked all that much.

I now have some funds burning a hole in my pocket...

My sources are Gungnir MB and Rega P2 w/ HanaEL. These are switched with a SYS into a Naim UnitiQute...

Really looking at two choices here:
Get Freya (S or +) and use Naim Amp.
Get Aegir and use Sys as pre-amp.

Which do you think would come out sounding better?
I'd get the Aegir and the Sys, then you can get a Freya (or a saga) at a later date.

Sys measures well, is transparent, and highly rated..... Stereophile actually lists it on their recommended section. It won't flavor your sound, which in this case is a good thing because it would allow you to appreciate the sound of Gungnir MB and Aegir for awhile until you decide you want a more feature rich preamp.
 
     Share This Post       
  • Like
Reactions: thebmc and wout31
post-15432929
Post #56,140 of 58,195

JohnnyCanuck

500+ Head-Fier
Joined
Jan 29, 2014
Messages
733
Reaction score
1,296
Location
Ottawa, Canada
Joined
Jan 29, 2014
Location
Ottawa, Canada
Posts
733
Likes
1,296
Right now I am using the pre section of the Naim. The SYS has the volume all the way up so it can just be a switchbox.

Do you think the SYS is degrading the sound in this configuration?
(Also yes my intent would be to use the volume of either the SYS or Freya in the new configs, the Naim has a AV Bypass function)
With the SYS used as a switchbox in your system as you describe you probably won't notice any degradation to your SQ.

If you intend to control your volume with the SYS then you will be dramatically increasing the source impedance that is seen by your power amplifier. This is what will degrade your SQ. It may be subtle enough to be acceptable to your ears but that's something that you'll have to determine for yourself. Much will depend on your source and amp.

JC
 
     Share This Post       
post-15432980
Post #56,141 of 58,195

Zojokkeli

1000+ Head-Fier
Joined
Aug 9, 2013
Messages
1,443
Reaction score
612
Location
Finland
Joined
Aug 9, 2013
Location
Finland
Posts
1,443
Likes
612
I'm not a big fan of redoing other people's work. Leave Star Trek alone. Don't colorize classic movies. And, please Hollywood, no more remakes. Somebody, somewhere, has to be writing original stories that are worth bringing to the screen.
A24 studios have brought pretty interesting movies out there.
 
     Share This Post       
post-15433143
Post #56,142 of 58,195

wout31

100+ Head-Fier
Joined
Jan 29, 2016
Messages
324
Reaction score
668
Location
The Netherlands
Joined
Jan 29, 2016
Location
The Netherlands
Posts
324
Likes
668
Website
quadesl.nl
Oh kids, it goes much further back... Did a lot of my early programming in Algol-60 on an Elliot 4130. Graduated to PDP-10 (various languages/OSs) Then C and PDP-11 Unix V7, then VAX 11/750 with 4.1 BSD. Then Symbolics 3600 LISP machines. Some VAX 11/780 VMS as well. Then Sun-2 Motorola 68010. Side-trip to Mac 512 and pretty much every major Mac hardware and software issue since. Then SGI Irix. Some WindowsNT C++. Then lots of Java on Sun and Intel-Linux desktops and servers. And it goes on...
Very nice line-up of programming languages and machines.
In those days you were 3/\I2|\|/\/\J7(_)I2^/25 right?
 
     Share This Post       
post-15433287
Post #56,143 of 58,195

quimbo

100+ Head-Fier
Joined
Feb 11, 2011
Messages
240
Reaction score
252
Joined
Feb 11, 2011
Posts
240
Likes
252
You're not counting the punchcards?
I have been programming professionally for over 40 years and started out with punch cards, used to get 1 compile a day. Took a few days just to get syntax errors corrected, never mind logic. Now i can compile 100s of times a day
 
post-15433496
Post #56,145 of 58,195

W S L

New Head-Fier
Joined
May 5, 2019
Messages
16
Reaction score
15
Location
Singapore
Joined
May 5, 2019
Location
Singapore
Posts
16
Likes
15
2020, Chapter 1:
The Weldenheim Chronicles

“Weldenheim?” you might ask. “What the heck is that?”

Patience. All will be revealed. Sit back, relax, and enjoy a read about one of the most interesting projects I’ve had in a long while—the world’s first and only ribbon headphone amplifier, Jotunheim R.

What’s most fun, perhaps, is that this story starts long before I was aware of the Raal/Requisite headphone. You could almost say the Jotunheim R is a summation of my entire career in audio.

The story starts back in the Sumo days, now almost three decades past.

Sumo was my first real job. Yeah, I had a job working for a government contractor doing communications system stuff, but that’s not a job. It’s a placeholder, a thing that dispenses enough money for you to swerve away from asking the hard questions: why am I here, will this ever mean anything, am I ever going to have any fun?

And Sumo made power amplifiers. Big honkin power amplifiers, amps that could deliver 200 amps of current per channel (on a short term basis, of course). Sumo’s amps used a error correction scheme that allowed us to turn up output current to 11. Really neat stuff. I’ve referred you before to Hawksford and Cordell if you want the theory behind it.

It’s funny how ideas have a way of coming around, a few years or decades later. “Error correction” is now known as “feedforward” and is the pimp du jour.

I may have even been King Pimp of Feedforward Mountain if I’d decided to include it on the original Jotunheim, when I started playing around with it back in 2014 or so. The first prototypes included the Hawksford/Cordell/Sumo error correction scheme. I pulled it out of the final design at least in part because it could deliver enough current to weld the TRS jack to the plug if you got it into your head to yank the headphones out at the wrong time when it was playing loudly. I could have detuned it a bit to keep the jack and plugs from becoming one, but the complexity, and my own concerns about compensation, kept it out of the final design.

“So what does this all mean, Stoddard?” someone shouts from the audience.

Well, if by “this all,” you mean “all this blathering about old amps and current,” it’s to set the stage. To provide the ambience. To hint at where this is all going.

Because, if you’re talking ribbon headphones, you’re basically talking about driving a near-zero-ohm load. A dead short. A screwdriver.

And if you want to drive something like that, you need a ton of current.


A Bit of Foreshadowing

Now, here’s the thing: long before the Raal/Requisite ribbon headphone was a reality—think, just a couple of years after Schiit started up—I had a meeting with another headphone company, and they mentioned they were working on a ribbon headphone.

“But of course there are a lot of technical challenges,” they told me. “The main one being that the ribbon is almost a dead short.”

I knew ribbon tweeters frequently used transformers to bring their impedance in-line with that of a typical speaker driver. “I expect transformers are too heavy for headphones.”

They nodded. “So we’d have to develop an amplifier specifically for driving these headphones.”

“Which makes it like an electrostatic,” I said.

“Yeah, so now it’s a system. So we’d have to sell a headphone and an amp. And the amp design itself is a challenge.”

I frowned. “Not really, I’ve done some really, really high current amps in the past. I don’t think it’d be too hard.”

They looked at me doubtfully. “It’s nearly a dead short.”

“And Sumo amps used to do 50-200A into 0.1 ohms on a burst basis.”

Now I had their attention. But from all the side-glances and sighs, I quickly figured this was something they were just playing with, maybe to keep the interest of a fickle designer-type in their company. Still, I let them know that if they wanted to do a wacky high-current amp, I was up for it.

Because, after all, it was a great challenge. It’s not often you get an audio engineering job that’s so far outside the norm…and so different than anything else. Electrostats were outside the norm, sure, but there were plenty of stat amps out there. Super hard-to-drive headphones like the HE-6 were outside the norm, too, but there were a growing number of super-stout headphone amps at the time, and plenty of speaker amps to use. Doing something for a wacky ribbon headphone would be a ton of fun.

But, after we parted, the weeks rolled past, and nobody contacted me about a ribbon amp. I wasn’t particularly surprised.

Ah well, no worries, plenty of other stuff to do…


Enter Someone Crazy Enough to Do It

I first became aware of the Raal/Requisite headphones the way I find out about many new things these days—I read about them on a forum.

At first, I didn’t really know what they were. They appeared about the same time the MySphere products showed up, so I remembered thinking, “Ah yeah, more K1000 type stuff,” lumping them together, and forgetting about them.

For a while.

I think someone told me more about the Raals at some point, and it sank in that they were ribbon headphones. But this was the middle of a crazy engineering time at Schiit, so, for some reason, it didn’t really register that these were ribbon headphones, and it didn’t occur to me to ask what the heck these guys were doing to drive them.

Until Danny from Requisite visited the Schiitr sometime in early June, 2019.

Danny was looking for an amp to run his headphones—or, more specifically, an amp to run the impedance converter box that ran the Raal/Requisite ribbon headphones. Eddie took care of him, and Danny was impressed by the capabilities of Vidar.

A few days later, Danny sent an introduction email to Mike and I, asking if he could get Vidars, in black finish, possibly to sell with his headphones. At the time, I was in the middle of engineering work beyond imagining (all the normal stuff that goes sideways around a product launch, plus CanJam LA coming up, plus stuff we were still working on and was late…I mean, heck, remember, we introduced 13 products last year), and I sent him an email back that said something like:

Holy crap busy!
Talk to Denise she will help!
Ten thousand too many things!
(But yeah let’s talk when it gets a little less crazy.)


Jason

Denise got Danny his Vidars for the LA CanJam, so that might have been that. Heck, I could have stepped back and said, “Sure, please buy Vidars,” and call it done. I mean, that’s what Danny wanted, and he would be perfectly happy doing that.

But it had sunk in: this was a ribbon headphone.

And I also remembered: people didn’t ask for cars, they asked for faster horses.

As in, until the automobile all they could imagine was a faster horse. Before the personal computer, all they could imagine was a larger TV. Before the iPhone, all they could imagine was a better phone keyboard.

So I made it a point to go to the LA CanJam, and to stop by the Raal/Requisite booth. Danny was there, with a Vidar, as well as a couple of other amps that sell for much more than a Vidar. He was playing the headphones, and allowing people to A/B the amps.

He was also telling people how much he liked the Vidar, and what a great value it was. He may even have mentioned it sounded better than at least one of the other, more expensive amps (or maybe I hallucinated that…we’ll never know).

But…he clearly liked the amp, and he was happily telling people how good it was. At that moment, I turned to Denise and said, “Get him as many Vidars as he wants.”

But I still hadn’t heard the headphones.

Now, I’ve heard a ton of headphones. Tons. Dynamics and planars and stats and open and closed and semi open and high efficiency and low efficiency and high impedance and low impedance and bright and dark and brash and slow…no, I’m not bragging, this is my job. The point is, I’d heard so many headphones that I thought I had a pretty good grasp of the breadth and depth of how they could sound.

When Danny handed me the Raal/Requisite headphones, I looked at them a bit skeptically. The ribbons are protected by sharp post-modern crescents of stainless steel, and the whole assembly is bolted to angular, articulating carbon fiber wings that look like pieces of Darth Vader’s TIE fighter. I was kinda-sorta afraid I was going to deli-slice my ears when I put them on.

Danny saw my hesitation. “They’re actually really comfortable. Just put the pads near your ears, and adjust the flaps in for more bass, out for less. Leaving them open about like they are is the most natural sound.”

Well, heck, it couldn’t hurt too much, could it?

I put them on. They didn’t fit very tight, but, as Danny said, they were comfortable. I pushed them back on my head as instructed and Danny lit up some tunes.

Ho.

Lee.

Schiit.


Seriously, this was completely unlike anything I’d ever heard before, and completely different than what I expected. Even on the loud-ass show floor, the Raal/Requisite headphones were the most dynamic, transparent headphones I’d ever heard. Furthermore, they didn’t sound anemic or thin, like I’d expected from the open-baffle design. They had bass. Maybe not a ton of sub-bass, but it sounded, well…shockingly realistic.

Danny saw the look on my face and grinned. “So…yeah?”

“Yeah,” I said.

Eventually I took them off. “We need a pair of these.”

“Can do,” Danny said. “In fact, can I come back to the Schiitr and have a listen to some other gear when I drop them off? I want to try some DACs and preamps to go with the Vidar.”

Oh you want even more gear? I almost laughed.

“Of course. I’ll see you there.

And that’s what started it all. Just one email, a quick listen, and an agreement to meet at the Schiitr. I still wasn’t thinking about direct-drive amps. I was just happy to hear a great new headphone, and that the guy behind them liked our gear. Win-win.


The Dumb Protagonist

I could have made it easy.

I could have shut up.

But no. When I have an idea, and it’s the right idea, I’m pretty damn stubborn.

But I’m getting ahead of myself. Let’s rewind to the day when I met Danny at the Schiitr, and he delivered our Raal/Requisite headphones. I figured it was just another meet and greet, but Danny actually had a specific goal in mind: putting together a complete Schiit system for his headphones—not just an amp, but the DAC and preamp needed to drive it.

Danny had been using a different DAC with volume control. I told him that if he wanted the same functionality from our products, he needed a DAC and a preamp, or we needed to go to a modular headphone amp like Jotunheim, with an internal DAC card and preamp outs.

Danny tried the DAC-and-preamp route, and thought it was OK, but that it was a lot of boxes.

“What I really want,” he told me, “Is something convenient for recording engineers doing mobile mastering. They won’t want to mess with a lot of boxes. One box for D/A and volume, plus one speaker amp, especially a small one like Vidar, is ideal.”

“Let’s try the Jotunheim,” I suggested.

We had one with a Multibit DAC card in it, so it was easy enough to swap out. One Jotunheim, one Vidar, done.

Danny listened for a while and seemed very pleased. “This sounds really good! And this can run two Vidars as well?” Danny asked.

“If you want, yes, it has balanced outputs.”

“Then that’s the way I’d want to go,” Danny said. “Jotunheim and Vidar.”

And that was it. That should have been it. That would have been it, if it wasn’t for me. If it wasn’t for me thinking that a Jotunheim R, a Vidar, and the Raal/Requisite resistor box was still a heck of a lot of boxes. If it wasn’t for the idea that came crashing down on me like an oxcart of sledgehammers: what if we just made a direct-drive amp instead?

“You know, we could probably do a direct-drive amp for your headphones,” I told Danny.

Danny blinked at me, as if I’d just asked him if he wanted to join Amway or start a Popeyes franchise together. You know, like I was a bit nuts. Or had Tourettes. Or was just prone to blurting stupid ideas at people I'd just met.

Eventually he found his voice. “I don’t think it would be that easy,” he said, speaking very slowly, as if to someone very stupid.

“I’ve done a lot of high-current amps.”

“It’s a 0.2 ohm load,” Danny continued.

“I know.”

“It would have to deliver seven amps of current,” Danny tried again.

But there was no dissuading me. In the grips of engineering frenzy...catching a glimpse of a new, weird, totally bizarre, completely different, and awesomely fun project, I had no doubt in my mind that this was totally doable. Some mental math gave me the rough idea…it was basically a question of super low rails and super high current, which meant huge efficiency…but nothing insurmountable!

“In fact, I think I can probably hack up a Jotunheim for proof of concept,” I told him. “It shouldn’t have to be any larger physically. We could just try it out with a lab supply and see how it goes.”

Danny laughed and gave me a doubtful smile that said, do whatever you want…but I think you’re a little nuts.

Eddie had been there, watching this whole exchange in silence.

Finally he piped up. “You’ll see,” he told Danny. “This guy’s a genius.”


A Different Opinion

“Genius?” Tyler said, when I told him what happened. “More like idiot. Someone wants to buy stuff you have, stuff you make…stuff we have in stock, and you say, ‘Nah bro, here let me do this pie-in-the-sky thing instead.’”

“Yeah, what he said,” David added. David is our JPL-by-day, hangs-out-at-Schiit-too-much guy who will someday figure out the JPL job is still a time-marking, day-wasting thing to get out of, but yeah, I get it, hanging your ass over the abyss and doing your own company isn’t for everyone.

I laughed. When they said it like that, yeah, it sounded stupid, but this was a really neat engineering challenge, it was something completely different, it was New Idea Day, and they weren’t gonna bring me down.

Heck, I could probably get a proof of concept high-current amp running that afternoon, if I really wanted to.

I had an old Jotunheim board that I could sacrifice. I had a lab supply to run the outputs off of. I could kludge in the 4-pin male connector somehow and get it hooked up, if it looked like it wasn’t gonna blow up or thermally run away.

So what’s holding you back? I asked myself.

Tons of things. There was still an overload of engineering work for real, actual products, products that were already slated for production. This was a silly lark, a distraction, a vanity project for driving exactly one headphone.

But it was a great headphone.

And sometimes you have to follow your whims.

I found the Jotunheim and got to hacking.

Big aside: the engineering behind what I was doing was pretty straightforward. Basically, what I had to do was dramatically increase the Jotunheim’s current transfer efficiency. Jotunheim, as it is, has:

1. Fairly large emitter resistors
2. Dropping resistors in the power supply
3. A transformer with a decent amount of DC resistance in its output (because it was designed for higher voltage rails
4. Output devices designed more for linearity at low output current, rather than brute force.


And there’s nothing wrong with all of this. When used with its target design loads—namely, headphones in the 16-600 ohm region—it works perfectly fine, resulting in a solid, reliable, powerful headphone amp.

But when loaded with a 0.2 ohm headphone, well, it goes to its knees. It’s like running a 16 ohm headphone on Valhalla 2. It just can’t do very much.

So, to maximize current output, I needed to kill as much internal resistance as possible. At the same time, the output voltage rails would have to get a lot smaller, to ensure the amp didn’t go into thermal runaway when the emitter resistors got smaller.

Which really opens a can of worms. Would I need a Vbe multiplier, rather than diode bias? Would it still go into thermal runaway if it got too hot? Would the output devices be a limiting factor?

But that’s engineering. You can obsess about what-ifs, or you can build something and see if it catches on fire. With Jotunheim R, I chose the latter approach.

So that’s what I did. I got hacking. 4.7 ohm emitter resistors got 50 times smaller. Dropping resistors in the power supply got dropped. I didn’t have a transformer designed for lower rails, so I started with a lab power supply, which would deliver up to 5 amps at any voltage I wanted to set it at, under 30V or so.

In under an hour, I had a Jotunheim R “first hack” version running on a lab supply with +/-4V rails on the output stage. It seemed thermally stable, and it passed signal.

“This is too easy,” I said.

“Except the Raals don’t plug into the female 4-pin output,” David reminded me.

“Crap.” I’d hoped to get it running that afternoon. I would have to order some male 4-pin XLRs and wait for them.

“I can make an adapter,” David said.

He took apart a couple of male XLRs and shorted the internal pins together, making a quick-and-dirty way to connect to the prototype amp.

“Give it a try,” he said.

Gingerly, I plugged in the Raals. Gingerly, because an amp that’s stable without a load might not be stable with a load.

Nothing caught on fire. There was no noise.

That was a good sign!

I turned up the volume a bit. And we had music! Holy schiit!

I mean, boom, there it was, working.

Or so I thought.

I did notice the readouts on the power supply, though. When we started cranking up the music, the output power supply quickly reached readings of 1.5-2A. That’s 1500-2000mA. And. That. Is. A. Ton. Of Current. And it still wasn’t that loud. Danny’s comment about needing 7A to run them suddenly was a lot more credible.

Aside: for example, if I wanted to do Jotunheim R with op-amps, it would take something like 100+ OPA1688 output stages to equal it…and maybe not even then, since the OPA1688 is not so hot at low voltage. That’s a TON of current.

But it played music!

I put the headphones on and had a quick listen. They sounded…a bit different than I was used to. Bright, and with less bass. But heck, the amp might be oscillating. It might have other problems. The main thing was that it was working—and on the same afternoon!

I took a quick video of it playing and sent it to Danny, with a short note saying, “Hey, this actually works, pretty cool huh?”

A still from the original video:



Danny didn’t reply immediately. Which was probably for the best. Because “working,” is one thing. “Right” is a whole ‘nother thing entirely.


Details, Details

I won’t bore you with the ten thousand design details of Jotunheim R, because, unless you’re an engineer, you probably think I’m a bit nuts anyway.

But there was plenty of hilarity in the design process, starting with Danny’s response to the original video.

Where he asked something like, “That’s cool, did you get the right pinout, and how did you do the baffle compensation?”

Pinout?

Baffle compensation?

LOL. It turns out that one reason the prototype sounded like ass was that it was out of phase. Running pin-to-pin like David did with the adapters wasn’t correct—Raal uses a different pinout than other 4-pin headphones. So it was out of phase.

Bright? No bass? Check and check.

Aside: but it worked! Clutching real hard to my engineering win!

Oh, and bright? Yeah, the Raal resistor box isn’t just full of resistors, it also has a couple of big chokes in it for frequency shaping. Since the Raal/Requisite headphones are open baffle, they use the resistor box to compensate for the bass fall-off.

Say it with me: durrrrrrrrrrrrrrrrrr….

Aside: again, it worked! Kinda.

So those were my first two engineering challenges.
  1. Fix the pinout. This was pretty easy, since the male 4-pin connectors could be bent to give the right connections on the board. It’s not something you would do in production, but for a prototype, it was perfectly fine.
  2. Get the baffle compensation in there. This was more interesting, because there were a ton of ways to do it, including just using output inductors like Raal, or using a passive network on the input, or doing active compensation in the feedback.
I decided to use active step baffle compensation to start, and that’s what Jotunheim R ended up with, since it worked so well. After measuring the output of the resistor box, I came up with a network that mimicked its effects.

Aside: well, I mimicked its effects as closely as I could do it with an active versus a passive network.

And, since the custom transformer I’d ordered came in, I also spent some time listening to it.

Holy moly. My notes at the time were simple: “Scary scary transparent, but distortion in bass is evident.”

The bass distortion turned out to be oscillation. I cured it with different compensation, but I wasn’t happy—the compensation I was using would effect performance. Still, it was worth measuring current output. And it clipped at 7.1A per channel! That was a ton of current…and right where Danny said it should be!

That’s where it got its first development name: Weldenheim, because “it had enough current you could weld with it.” A bit of hyperbole, yes, but still a very stout amp.



But I was disappointed. I wanted to get well above where Danny wanted the current output. Double it, if possible. And Weldenheim should be able to do it. What was the bottleneck?

A few things, really.

One problem was the bridge rectifier. At low voltages, you can lose a ton through the bridge. I changed it to a Schottky.

Another problem was the filter caps were being drawn down. They were a lot higher voltage than they needed to be, so I upsized them by 3X for something that was more congruent with the rails.

But the big thing was the output devices. They really were meant for like 1-2A output each. So the beta was collapsing at high current. I changed them out for new parts—rated at a massive 15A—and the difference was immediately clear.

With the bridge, caps, and output devices, we were now clipping at 10A!

That was more like it.

But could I do better?

I played with a lot of dumb ideas, including trying to keep the rails solid with ultracapacitors. It sounds like a solid idea, but in reality it’s not very bright. Ultracapacitors have a ton of equivalent resistance. Adding them didn’t change the measurements one bit. Factor in the uncertainty of running ultracapacitors on a supply with significant ripple current, and it was best not to go there.

Around this time, we allowed a few people to listen to the prototype, which was still an old Jotunheim board, with a bunch of parts flying in the air. We had to warn them not to turn it up for long (I knew I would need a heatsink, but I didn’t have one at the time).

And their response?

“Wow, this is cool that it works…but hmm, it is helladangheckadoodle edgy, ain’t it?”

Well, those weren’t their exact words, but I could see it. They were impressed it worked at all, but they weren’t happy with the sound. Dissappointing.

Of course, we could just go back to selling Vidars and Jotunheims, and be done with it.

But. No.

I knew what I really needed to do: I needed to move Jotunheim R over to Nexus™.


Into the Nexus

Why Nexus? Mainly because it sounds better.

And yeah, I know I’m gonna get crucified for this, especially after my comments about how small the differences were on the Magni 3+/Heresy blind listening.

Go ahead. Crucify me. That’s fine. Sometimes things may have a bigger effect. Nexus is a much more tonally rich, “wet”-sounding topology than Pivot Point. It’s also easier to compensate. I suspected it would be much better in the real world.

So I started the hard work: putting Nexus on a Jotunheim-sized board, while also doing the changes necessary for the Raals—adding the baffle step compensation, changing the output connector, etc. The end result was a 4-layer board like Jotunheim, but one that was significantly more complex, with the need for many matched parts. I also added something that I thought many pros would find useful: a way to disable the preamp outputs. It also got a host of other little tech changes like a revised high-voltage supply and proper Vbe multipliers for biasing.

Aside: unlike Jotunheim, Jotunheim R’s preamp outputs are passive. I got a little paranoid about someone maybe shorting a preamp output to ground, so I went this way instead of an active preamp output. Equal numbers will hate me or love me for this decision; it’s the way things work.

On first start-up, with a slightly tweaked transformer, the Weldenheim, Nexus Edition started up and worked…and it didn’t oscillate like the original. Which meant that compensation was easier.

But, when cranked up, it still oscillated when it hit the rails.

Argh.

Then I remembered: wait a sec, this is a super high bandwidth amplifier, running tons of current into a low impedance load. It might need a Zobel network on the output!

Aside: a Zobel network is common to most speaker power amplifiers, and helps ensure that the speaker amp sees a defined load impedance at high frequencies. It’s not typically used in headphone amps…but Weldenheim had grown far beyond most headphone amps!

I pasted on a Zobel network (just the parallel C+R part, scaled for the Jotunheim R’s expected impedance load)…and BOOM, there you go. Fully working, no oscillation.

And, most importantly, it delivered over 13A, both channels driven, into the Raals. This was when I started worrying about smoking the headphones, they were getting so loud.

Okay. It wasn’t double what Danny said, but it was enough.

It was time that I had Danny and some other listeners back to hear it. And this time, they were all smiles. No comments about the brightness or edge. In fact, Danny thought it was a little rolled. So we tweaked the baffle step compensation a bit until he was happy.



So was I done?

No. Not even.

It wasn’t even a product yet.


Connecting to the Real World

So why wasn’t Weldenheim (er, Jotunheim R) a product yet? If it performed well, and sounded good, you’re done, right?

Ah. No.

The problem with reality is, there is this place called The Real World, where unexpected things can happen.

Unexpected things like someone running Jotunheim R all the way up until it got so hot it cooked itself (we’re talking gross distortion, pure crazytown stuff, things nobody would want to listen to—except someone DOES), or making an adapter to run Jotunheim R with TRS headphones (which would short the outputs, and, in the case of the prototype, would let all the magic smoke out.)

To make it a real product, we would have to add:
  1. Over-temperature protection.
  2. Over-current protection.
But there were a ton of gotchas here as well. Over-temperature, in our speaker power amps, is monitored by a microprocessor connected to a linear temperature sensor. Over-current, in the speaker power amps, is also overseen by the microprocessor, in conjunction with a complex ground-current-sensing scheme that wouldn’t work with Weldenheim’s transformer.

Aside: the plain-Jane Jotunheim doesn’t need this level of protection—it’s designed to gracefully fade as current output increases, thanks to its internal levels of resistance. When you start going for maximum current transfer, though, you lose this internal protection.

I really didn’t want to throw a microprocessor in Jotunheim R, because it was already complex and expensive, thanks to the exotic parts and short run sizes. Developing firmware for a short-run product wasn’t all that appealing, either.

It took a while to figure a way around it, but I eventually ended up with Hall-Effect current sensors (which are technically not in the signal path at all, they simply sense the current passing through an internal wire) and a thermostat (rather than linear current sensor). With some simple old-school analog logic, I was able to create a protection system that will mute the output if Jotunheim R tries to source over 31A from both channels, or if it goes over 95 degrees C on the heatsink (plenty safe for devices rated for 175 degrees C continuous.)

Aside: yes, 31A. That’s thirty-one AMPS. And yes, you can trigger this protection on the APx when testing into the Raal headphones and running a signal of more than 2VRMS into the amp.

The next prototype included the Hall Effect sensors, the thermostat, screws for a heatsink (custom from AAVID, designed to fit under the True Multibit DAC card), and, most importantly, switching for the baffle step compensation. Because, who knows, maybe Raal/Requisite will come out with ribbon headphones that don’t need baffle compensation, and then it would suck massive donkey balls if you had to buy a whole new amp for that. Or maybe other companies will come out with ribbon headphones. Or maybe you want to run a dozen pair of Magneplanar LRS in parallel off this amp (only partially kidding, more on that later.)

And that was that. I figured we were ready for production. I had our PCB assembly house do a half-dozen prototypes of this version, figuring we’d send them out to various listeners and see how they did.

When the boards came in, they worked…with a couple of hilarious exceptions:
  1. The baffle compensation and preamp enable switches worked in reverse. Yes, I am an idiot. Maybe you already knew this.
  2. The protection system worked, but thanks to me leaving a couple of key capacitors off, it made a weird whining noise as it came up to operation (this was a physical whine from the chassis itself, not through the transducers.)
Still, it was good enough to put a “Production Qualifier” badge on, and see what people thought. The next rev could clean up the switches and protection oddities.



This is where we could start getting serious. We sent all but one of the production qualifiers out to listeners and reviewers who were familiar with the SR1a headphones, and asked them if they noticed any glitches or operational oddities, as well as what they thought of the sound.

Everyone was thrilled. Some said it even sounded better than many speaker amps. The amps worked fine, and the protection system worked as expected.

Time to start up production!

I made a couple of board revs, sent the board out, and commenced to speculate about possibly having the Jotunheim R out at the end of 2019.

Aside: I’ve said I’m an idiot, right?

Aside to the aside: but we didn’t miss by that much, right?

And that’s what we’re launching today: a production Jotunheim R. As of this writing, it is the first and only direct-drive ribbon headphone amp in the world. It substantially simplifies using the Raal/Requisite SR1a headphones, eliminating the need for the resistor box, speaker amp, DAC and volume control. It’s a one-box solution for one of the best headphones in the world, and it’s ready for any future ribbon headphones that don’t need baffle compensation.

“Well, that’s all great, Stoddard,” someone will say sarcastically. “But if you hadn’t been doing your pet project for one single headphone, what else could you have been doing?”

Great question.

Because sometimes you have to look around corners.

Ask the Apollo program guys.


Technological Fallup

Fallup. Not fallout. Because this is a good thing.

As with the Apollo program, which arguably significantly accelerated, or directly birthed, many of the technologies that have us glued to phone screens for 18 hours a day, the Jotunheim R project has had a big impact on us.

How?

Let’s start with this interesting observation: Jotunheim R is more than just a ribbon headphone amp. In fact, it’s a damn nice speaker amp. It laughs off 8 ohm loads like Ragnarok 2 laughs off 32 ohm headphones. Hell, it’s under 0.001% THD at 1W into 8 ohms.

The catch? It also maxes out at like 3.2W into 8 ohms. That’s the low rail voltages working against you. I didn’t test it, but it would probably do almost 6W into 4 ohms and 10W into 2 ohms. Maybe. Don’t quote me. However, here’s the thing: you could wire up an adapter, and run speakers all day off the Jotunheim R.

But there’s more. Remember, it can also drive insanely low impedances.

So, what about hooking up, say, 10 pairs of Magneplanar LRSes in parallel? It should drive that just fine. How would it sound? I don’t know. Just blue-skying here. But Jotunheim R opens up new things we can play with. Which is pretty cool.

“Sounds pretty niche, Stoddard,” someone says. “I don’t want a fleapower amp that I have to make adapters for, and I don’t want to buy a metric asston of Maggies.”

Cool. Because there’s a lot more fallup. It’ll just take a while longer to appear.

In fact, one of the key technologies I mentioned in last year’s wrap-up chapter, OG tech, is a direct result of Jotunheim R development. OG is already working in a proof-of-concept prototype, and, with more refinement, could end up being a foundational technology that leads to whole new classes of products. I really can’t say more, because there may still be something I missed, but this looks like a very, very good thing.

But that’s only one new piece of technology. While developing Jotunheim R, I was excited by the possibilities so much, that I came up with other new ideas. I still have yet to prove some of them, but they are also things we’ll be working on this year.

So yeah, Jotunheim R may have seemed like a distraction…but in reality, it was something that led to entirely new vistas. Sometimes you have to get weird, chase the bizarre, and have some fun. Because you don’t know where it might go.
Hi Jason. I've listened to one of the Jotunheim R preproduction units, and I have one question: can you make a more powerful version? Maybe something equivalent to two vidars in monoblock mode?
I'm asking because just as I got to an enjoyable volume level, the unit I was listening to started clipping.. So a Jotunheim R with more power will offer more peace of mind, at least for me that is.
This was what I was listening to when the clipping hapened by the way. Lots of electronica and bass.
Kitaro_enchanted_evening.jpg
 
Last edited:
     Share This Post       

Users Who Are Viewing This Thread (Users: 23, Guests: 67)

Top