2024, Chapter 14
Do Epic Schiit
Late in 2023, I emailed Martin something like this:
“We just need to have it do one thing. Press one button on a phone and have it do one thing on a preamp. That’s it.”
“It will work,” Martin responded.
Exasperated, I sent back: “Just one button. It’s
time for it to do something.”
“It” was Forkbeard.
After months of diddling, it still wasn’t working. At least not in the way I defined “working.”
Yeah, we’d made progress. We’d gone through a couple of revs of modules. We had a chassis drawn up. And we had Kara boards with connectors to plug the modules into. Stephan and Martin had established connection to the modules, I’d worked through some power management issues, and they’d pronounced them all Bluetooth-y and communicate-y and schiit, but I actually
hadn’t seen anyone press a button on a phone and have a product do something. Not an input change, not a mode change, not a mute, not a volume change. Nothing.
So I was getting a bit anxious. This was supposed to be the biggest announcement of 2024. An all-new technology. A game-changer.
Anyone, at any tech company, would be biting their nails over this.
But that wasn’t the end of our stress.
Oh no, not by a long shot.
“There’s nothing for sale in San Antonio,” Tyler told me, looking morose. “At least not at the kind of prices we want to pay. And interest rates are stupid.”
Tyler was referring to our search for a building in Texas, a place to move California operations.
“So what do we do, lease again?” I asked. That didn’t thrill me. There’s a big difference between leasing and owning your space.
“We’ll still save by leasing in Texas,” Tyler told me.
I didn’t look happy.
“We have to find an answer by the end of the year,” Tyler reminded me. That was when our California leases were up. Which meant huge increases in rent. Which had already grown to 6x what we were paying in Texas on our mortgage on our building in Corpus Christi.
Sigh. “Go ahead and look at leases. But I’d rather buy.”
Tyler nodded. “I know that. But it’s looking like it’ll take a miracle.”
So, no Forkbeard, no building.
What else could go wrong?
Oh ho ho.
Singular™, our own delta-sigma modulator based on our own math, is implemented on an FPGA. While we were getting our FPGA knowledge up in-house, we contracted with a firm to have it implemented. The work was done.
Then they kept working on it.
Yeah. Without a contract. Without an agreement.
On our technology.
And then they wanted paid.
Yeah. These are truly the ball-shrinking moments in business, when you realize that what you’ve been working on for 30 months may end up not happening at all, or, worse, stuck in a morass of lawyers and suits.
Recap:
- Forkbeard not working
- No building to move to
- Singular maybe dead completely
Is that all?
Har har.
This is also the time that the first Wotan prototypes went up like fireworks. I mean, literal explosions of flashes of light and smoke. These were the high-voltage versions we ended up not going with at all. We had to find another approach.
So that must be it, right?
Oh no. This is also the time that Lisa and I finally got started on building a house on land we’d bought in Boerne a year ago, after a fairly disastrous stint with a fancy architecture firm that designed something that looked very cool, but would have blown the budget by 2x. We didn’t do that. We ended up going with the house I designed on Chief Architect over a weekend as a model for what we were looking for (and the architectural firm admitted they’d never looked at).
One little problem: we didn’t have the free cash to build the house, and I was loathe to try to get a construction loan, for all the pain they cause, and also because I knew we were going for a loan on a business property if we found one. Building the new house meant selling the Cali house. And the Cali house was a big weird thing that wasn’t for everyone, so it might be, er,
interesting to sell.
And no, for the wags out there, you can’t just take a bunch of cash out of the company, especially a company facing big moving costs, plus a down payment on a building (hopefully) and all of the other $ that leaves the building during relocation.
Summary of late 2023/early 2024:
- No Forkbeard
- No building
- No Singularity
- No Wotan
- No (personal) cash
A sane person, and a sane company, would have probably decided:
**** all this, ain’t gonna do it. Forego the move, kill Forkbeard, build the house later or sell the land and buy something already done, save your energy for Singularity and Wotan, stuff you know is in your wheelhouse.
But we are insane.
So what did we do?
All of it.
Let’s Talk Business
Now, I know there are people out there saying, “Why’d you move from California in the first place, I love California, why couldn’t you make it work out?”
And there are others saying, “What happened to the California employees? Did you leave all of them behind?”
And still others saying, “What happened with the building?”
So let’s start with those questions.
First, the California move. The move was long on the table. Moving the small products to Corpus Christi in 2021 allowed us to forego large price increases and streamline production. The space we were at in California was a horrible morass of stuck-together industrial units, with huge amounts of unusable space and inefficient workflow. We weren’t able to make large changes in the building because we didn’t own it, and the owners didn’t want to sell (why would you, when rent literally went up 2.5x during the time we were there)?
Nor was the move a secret. Key personnel knew about it from 2022 onwards. And when I finally announced it to everyone in January 2024, it was a 9-month notice. Also, everyone was invited to come to Texas and keep their job and their rate, or take (very) generous severance if they stayed on in California to the end of August.
So what happened? Many people who wanted to come with us didn’t, mainly due to friends and family. I get it; you can’t fight those kind of bonds.
We did keep the QC department entirely intact: Asa, Zacchias, and Amita. Shaun stayed on for complex technical support, while still living in California. Alex moved up to San Antonio to run the show there, while Tyler stayed in Corpus, so everything worked out fairly well.
Was it smooth? Was it easy? No.
But we don’t have to do it again. Because, after looking at a ton of buildings in San Antonio, a miracle happened.
And, I mean, boy it looked bleak for a while. We went through a dozen leasable spaces, any of which might have worked, but they were leases.
Leases that we wouldn’t want to spend to build out.
Leases that weren’t ours.
Leases that change price every year.
Yuck.
The time came when the three of us—Alex, Tyler, and myself—went up to San Antonio to do a final survey of what was for sale.
“There are several that are at the price you want to stay below,” Tyler said. “But I put a couple in there that are nice.”
“What do you mean? The others aren’t nice?”
Tyler frowned. “You’ll see.”
“But they have enough space. We make things. It doesn’t have to be fancy.”
Tyler didn’t say anything.
And, on that day, I saw what he meant. The first place was in a run-down section of town, but it was big, maybe even bigger than we needed. The high-tension lines looming over the building weren’t confidence inspiring, nor was the rust-colored spray-on foam insulation they’d used on the 1960s-era roof.
The next was similar, and the third was worse. The fourth you had to drive through a bombed-out looking neighborhood to get to, but boy was it cheap. Unfortunately it had very little parking and I swear we saw an actual drug deal go down at the no-tell motel 100 yards down the street.
We did a couple more after that, which were nicer, but blindingly expensive. I started to lose hope. I started wondering if we’d be stuck leasing.
“One more,” Tyler said.
This one was on the other side of town, south of 1604 and west of I-10, near UTSA. As we got closer to it, I slumped.
“This is a nice neighborhood,” I said.
“Why’s that bad?” Tyler asked.
“We can’t afford it.”
“It’s more than you want to spend, but a lot less than you think,” he told me.
Uh-huh. I said nothing.
When we got there, my hopes dropped. It was a nice building, much nicer than we’d ever had in California. It looked more like an office than manufacturing, but it was zoned right. And it was sitting on over 2 acres, so we could expand if we needed to. Tons of parking, too.
“There’s no way we can afford this,” I said.
“They just dropped the price another 15%,” said our realtor, who’d met us there.
“What?” Tyler asked. “Why?”
“It has a…little foundation problem,” he told us.
Oof. That didn’t sound good.
“They really want to get rid of it,” our realtor said. “They bought the company, got the building, now just want to get rid of the building now that they’ve consolidated.”
Hmm.
We went in.
It was a very very nice space. What’s more, it was already pretty much set up exactly as we would have done it. Hot and cool warehouse, separate area for production, separate area for QC, even an old-skool server room with a ton of AC we could use for burning in products.
“What’s wrong with the foundation?” Tyler asked.
Alex laughed. “Just look at it.”
At this point we were inside the building, which was devoid of furniture. You could see for almost 100 feet down a hallway.
“Don’t drop your basketball in the back,” I said. “Or you’ll be picking it up in the front.”
So yeah. Except for the 10” lean, the building was perfect.
I started looking at it harder. It was a concrete tilt-up, which is fairly weird for Texas. Built in 1986, it clearly wasn’t moving very fast. No roof leaks. No plumbing problems. Really perfectly fine except for the tilt.
“There are companies that specialize in fixing stuff like this,” said the realtor.
That’s all I needed to know.
“Get a quote. And then let me know how low you think we can go.”
Condensing the rest of the building story: we got a great deal, Tyler walked through the fire to work with our bank to get an excellent loan, and then we put 6 figures into 130 caissons and structural foam to stop the lean. We even managed to correct it a bit. But in the end, we got a building for about ½ off, a building we have no right affording. And we have room to build.
And now, after a horrifying summer of moving—I mean, packing up and moving a business that’s been in operation over 14 years is nuts—we’re now largely operational. There are still too many things in boxes, but Schiit San Antonio is now added to Schiit Corpus Christi.
Still no Schiitr, but we’ll get to that in 2025.
Tech Follies 101
“Forget the building,” says someone. “That’s just real estate. What about all this forked-up tech? Forkbeard not working and Singularity in limbo? Tell us about that.”
My message to Martin about not seeing Forkbeard working? Wanting to see one button do one thing?
Yeah, that’s me. I work iteratively.
I get something kinda-sorta working and then add on. Sometimes that leads to insanely stupid stuff like the first version of Stjarna not having standoffs to mount the board to the chassis. A lot of times it helps iron out complex analog problems that may otherwise be insurmountable.
Martin works differently. Martin basically does the whole thing and hands it to you, done. So he doesn’t really think in terms of “let’s have one button do one thing and show it’ll work,” he thinks, “let’s do an app that does it all and hand it to everyone to use.”
Long story short, in early 2024 we had a fully working (but very ugly) dev app that did everything. Everything for Kara. But also everything for everything else we were developing. Including Saga 2, Vidar 2F, Aegir 2F, Kraken, and hell, some things I still can’t name. Every time I think, “There’s no possible way it could be set up for this product,” it is.
But that’s just the way he works. Like in late October, shortly before we were to announce Forkbeard at CanJam,
I still hadn’t seen a cosmetic app.
As in, yeah, we had the ugly development app. But that was nothing like the real thing. It was paged and parceled. You pick a product and control it. It had tons of info like IDs and communications status and temperature to 5 digits just hanging out there.
You get it. OOOOOOOgggglllly.
What’s more, I’d never seen a working version of any of the advanced features we wanted to show. No Visual Volume. No Real-Time Power. Nothing.
So I started getting nervous. “Hey Martin, when’s that cosmetic app going to be ready.”
“Soon,” he told me.
“In time for the show?”
“In time for the show.”
I pushed. “Can I get a preview?”
“No.”
Cool. Not cool. I didn’t know what to write for the chapter and the press. I didn’t know what advanced features we’d have, if any. I wrote a version that was full of all the details, then pulled them all out, then rewrote it kinda in the middle.
One day before the show, Martin releases the cosmetic app.
“And it has Visual Volume,” he said.
Because that’s how Martin does things.
Now, of course, we were still tuning it literally the night before the show, but the main thing is it got there and worked flawlessly. And, it’s a template for our future. More thought. Less iteration. More solid launches.
I get it.
But it’s still going to take me a while to get used to the way Martin works.
Now, with Singular, that was a whole different thing. I don’t need to get into all the gory details, but it was, at best, a severe miscommunication.
The short version:
What happened was that we contracted a firm to help us get our filter, and our own delta-sigma modulator based on our math, implemented on an FPGA. The work was done. They delivered a final product. We paid for it.
Done, right?
Well yeah. Except I wasn’t really super-involved in all the details. So when our internal staff said the firm was working with the code again, I figured it was some kind of cleanup thing, or we’d defined a new scope of work.
No.
What happened was they decided to work on the code by themselves, without a contract, and without money being exchanged. Which led to some rather testy exchanges once I figured out they were working with our IP without authorization, on something that was supposed to be done.
Then they said, yeah, and by the way, the amount of work we put into this was (insert a number that is much bigger than our typical yearly profit), and hey wouldn’t it be nice to get paid.
I told them that maybe they thought we were rich or dumb, but I’d run consulting firms before, and there was no way I would have ever had my team put that kind of work into something on a speculative basis, and oh by the way that’s our IP.
And that’s when lawyers got involved, and we considered—hell, more than considered, it was a done deal—killing Singular entirely, since a big weird dark cloud like that wouldn’t do anyone any favors, and we could concentrate on new directions, like Multiform.
So what happened? Typical lawyery stuff. A deal was made, money was exchanged, and neither party was happy. A good definition of a compromise. But Singular is now 100% free and clear and nobody else’s and if we do any more work on it it will be done in-house.
Aside: think business is fun and easy? Think you can be the “nice guy?” Sometimes it isn’t. This was one of them.
But in this crapfest, it caused enough uncertainty to make Gungnir 2 start as Multiform™ rather than Singular. Not a bad change, because we had prototyped both, and both were strong in their own way.
And yes, you will see at least one implementation of Singular in 2025. We’ve already started production planning for that.
Because We Don’t Have Enough to Do
In all of this craziness, one thing proceeded along pretty much as expected: our house. We even managed to sell the California place in enough time that everything worked out, though oh boy there were some finger-biting moments.
But in the middle of all of this, Lisa decided she wanted handmade medieval filled tile for the backsplash for the kitchen, like she had in the California house.
One problem: the company we used before ghosted us.
Two problems: the other companies we found that did something similar either didn’t want to do the custom colors we wanted, or weren’t interested in custom at all.
So what did we do?
The least sensible thing ever. Lisa decided to make her own tile.
Now, let me make this clear: this is medieval style filled and scraped two-color plus glaze tile. Not single color. Not painted. Filled and scraped. Which means you need a pattern to stamp into the clay, then you fill it with colored slip, then you scrape it flat, then you bisque fire it, then you put glaze on it and fire again.
Oh yeah and did I mention that neither of us had ever done any ceramic work before?
Yeah.
We.
Are.
Dumb.
But Lisa got herself a kiln and a slab roller and started experimenting, and I, being stupid, agreed to provide the 3D-printed patterns to stamp into the tile. After a day’s worth of cussing, I came up with a frankenmethod for tracing medieval tiles using Photoshop and Illustrator, and then managed to bring the vector file into Alibre to extrude from a flat stamp. After some more cussing and experimentation, I figured out the best depth and material, and made a bunch of stamps.
A bunch. As in like 35 separate designs. Most were what you might expect to see in a 13th-century monastery, but of course we had to throw some silly ones in like a dalek and a surfing knight holding a beer.
And, somehow, in the middle of everything, Lisa was able to produce some 300 tiles, all by herself, using a 7x7 kiln running pretty much 24/7 on our back patio in Corpus Christi.
You know. Because we don’t have enough to do.
The Actual Move
Unless you’re some kind of European landed gentry, you’ve moved. You know how much it sucks. You know nobody sane wants to do it.
Well, now consider what a move looks like for a company that has grown organically into almost 18,000 square feet of space for the last dozen years. A company that has severe problems throwing things away when it doesn’t need them. A company that, for many years, was so disorganized that we could lose over a half a million dollars in DACs on a shelf.
Yeah. Total madness.
So how did we do this in a way that caused the least amount of disruption for our customers? Simple: everyone stepped up.
Alex, early on, shook his head and said, “Well, you had me replicate California operations here in Corpus in 8 weeks in 2021. I guess I get to do it again.”
Elvis, early on, said, “We better build up a crapton of stock in Valencia and ship it, finished, to San Antonio, so you guys have something to sell over the holidays. And, by the way, I need another couple of guys to do it.”
Aside: and that’s why you still see things with California tape on them. They were built in Valencia.
And Alex planned and Elvis built.
But that really undersells what happened, because everyone went above and beyond. Alex found the most efficient way to get full truckloads to San Antonio, and decided what was best purchased new and what was better brought from California. Elvis oversaw production, pack-out, and cleanup of our sprawling office. Tyler and Alex went up and personally painted the new building. Everyone in Valencia worked hard until the end, from Cameron in tech to Will at the Schiitr.
In the end, no, not everyone came. Like I said, you can’t break the bonds of family and friends. And it shouldn’t be our goal to do so.
And twenty-some-odd full truck loads later, everything is in San Antonio.
Now, it’s not yet completely organized.
And we’ve probably lost a few things in the move.
(But not $500K in ICs.)
In short: it’s still a bit chaotic, but all is well. I’ll do a video up there shortly so you can see the new digs.
Productucopia
Okay, let’s shift gears to more practical things, like lists.
Despite all this upheaval in 2024, we introduced something between 7 and 11 products this year, depending on how you count.
Because we are crazy.
Let’s recap:
- Aegir 2/Aegir 2F. Early this year, we introduced Aegir 2, an improved Aegir with more power and lower operating temperature, thanks to a revised Continuity™ scheme like we use in Tyr. It was just a bit too early to get Forkbearded at the beginning, so it was replaced with Aegir 2F in November—the exact same product, now with Forkbeard.
- Vali 3. Finally Vali gets the 100V rail a tube deserves, and a beautiful new chassis thanks to our stamping wizard partners. It’s also an insane sounding amp, and a perfect introduction to tubes, at only $149.
- Kraken. Completing the Syn Stack, Kraken won’t be around for long. You can argue I only did it so I could have my Syn and one of our own amps for my systems. I never expected it to be super-popular, so it’ll go away when they’re sold out, while Syn finds new life in gaming. Fun fact: Kraken was the first product with a functional Forkbeard port.
- Saga 2. Flexing our manufacturing muscle in Corpus Christi a bit, we delivered the least-expensive, most-flexible, best-sounding Saga ever. At $279, it’s the highest value preamp in the world…and it’s made in Texas. Add Forkbeard, and at $329 total, it has no match anywhere.
- Gunnr. Taking Width and Presence off of Syn and moving them to the most powerful gaming device we’ve made, ever…and coupling it to a new Linear Override power supply makes Gunnr the choice for serious gaming. At $229, another insane value from Corpus Christi.
- Gungnir 2. The rebirth of Gungnir brings Forkbeard, Multiform, Autonomy 2, over-the-air updating, Unison 384, two USB inputs, and a completely new architecture for our penultimate DAC.
- Wotan. The return of the big-power, drive-anything amp adds Forkbeard and feedback-free mode switching to our repertoire.
- Kara F, Freya+ F, and Vidar 2F. Forkbeard comes to our popular preamps and compact amp.
So, by the numbers:
- 7 new products if you’re being strict and leaving out “F” updates
- 11 new products if you count everything
- 8 products Forkbearded to start
What’s next? At least 1 more Forkbearded product soon (Stjarna). Two more shortly after. And no, not everything will be Forkbearded. Expect to see both Forkbearded and “Blood and Guts” products in 2025.
Aaaaannnd…expect to see
a lot of products.
What We Did Right
While we’re on a more functional tear, let’s do the two favorite lists of the year: what we did right and what we did wrong.
We’re start with the one you’re less interested in, because, let’s face it, seeing people pat themselves on the back isn’t really a big spectator sport, and could actually be off-putting.
But to get it over with, what did we do well?
- Obvious victory is obvious. We managed a ton of complex change very well. Not flawlessly, but good enough that the seismic upheaval of 2024 was fairly invisible to our customers. Orders were shipped, customers were supported, we didn’t have too many out-of-stock problems. Simply maintaining that was a huge win.
- Delivered the most sweeping product/tech introduction ever. Forkbeard was a complete change of direction, involved 8 products at launch, and leapfrogged us over the entire audio world with patent-pending technology…and it works flawlessly.
- Introduced several standout products, including Vali 3, Gungnir 2, Wotan, Saga 2, and Gunnr. Vali 3 is a standout hybrid amp, insanely cheap in these inflated days, and also lauded by critics for incredible sound. Gungnir 2 introduces so many new ideas in our DACs it sounds like a joke—from dual USB to Unison 384 to Autonomy 2 to Multiform to over-the-air updates. Wotan is a huge step away from what we’ve done in amps to date, emphasizing power and capability, and prioritizing sound over measurements with its switchable overall feedback, as well as being the focal point for our first Giga-Stack and Visual Volume. Saga 2 completely nukes the price-performance equation in preamps, delivering features not found in preamps at 100x the price. And Gunnr redefines mission-critical gaming performance with technology taken from Syn.
- Started the process of verticalization. The only way we’ll win in an increasingly costly environment is by going vertical. Doing our own boards, doing some of our own chassis, doing more internal components with 3D printing, designing more custom parts for more efficient production. We made great strides in 3D printing last year, and started the process of getting PCBs in house.
- Introduced important new communications. We’re now producing our videos internally, and are rolling out both my own “blathers” and helpful videos, like the one we just did about tubes. There will be lots more, on more tech issues, on specific products, and more. Oh yeah, and we added a previous product listing to our guides, with all old manuals and helpful details, like can we repair a 14-year-old Asgard out of warranty. Hint: yes, we can. Let us know what else you want to see.
What We Did Wrong
And now for the part that everyone probably reads first: what we did wrong. And no, I’m not gonna do like last year and give you platitudes, because we did plenty of crap wrong, and it’s good to reflect on that.
So here we go:
- Under-qualification led to too many stupid problems. Yes, we were busy, and yes, it was a crazy year, but it doesn’t excuse all the small-but-showstopping problems we had with new product introductions. Saga 2 had a true horror of a bug that put its introduction back several months. We threw away 2 runs of boards to get it right, in addition to dozens of firmware changes. And pretty much the same thing happened with Gunnr. Both of those are examples of things that should have been worked out before production. Hell, Valhalla 3 got taken completely off the table due to intractable thermal issues we only discovered after chassis prototypes. And that’s why we’re focusing on qualifying products in a lot more detail this coming year.
- Overconfident about the year. I said that Gungnir 2 would be coming in summer with Singular. Har. I also said we’d have Valhalla 3 about that time. It almost didn’t happen. We simply got too cocky. And that’s why there’ll be less prognostication about 2025.
- Surround wasn’t as big as expected. Kraken is our shortest-lived product. Syn is gaining traction as a gaming device. Still, I’d like to take a run at it, but I bet it’ll be in a much more limited, higher-end, lower-production device.
- Not enough resources to put into a big formal Forkbeard launch. As in, we didn’t have an Apple-like intro event and we didn’t do a bunch of slick produced videos. But that’s OK. This is a marathon, not a sprint. The videos are coming, and we’ll do an event or two in 2025, as we add some really cool features.
The Future
“So what’s coming?” everyone asks.
This is where you expect the big list of products, the list I never really give you. And this year the list will be even shorter, because I learned my lesson by being too cocky about Gungnir 2 and Valhalla 3.
I mean, there is one gimme: Stjarna, which should be along shortly.
I also mentioned Skoll F in the wrapup video, which is close enough to talk about (as in, we have metal and first article boards).
Aaaah, and Gjallarhorn F. Also soon. And Valhalla 3.
But those are all just appetizers.
What I can say is that you should expect unprecedented new introductions. Because we’re not moving. It’s not all upheaval. There’s not a ton of change to deal with this year. So we’re not bogged down. We’re not distracted. We can do a lot more.
And we’re already sitting on Forkbeard, the best interface tech in the business, as well as truly unique digital approaches like Multiform, as well as advanced analog topologies like Nexus and Continuity. The foundation is rock-solid, and we’ve already built farrrrrr past what you have seen.
We're unleashed.
In 2025, we’re gonna
do epic Schiit.
![20250517_125204[1].jpg](https://cdn.head-fi.org/a/12739349.jpg)
I try to leave expectation bias at the door and just listen, some expect a sound and might defend it because they just dropped thousands on expensive gear when they might have picked Schiit in a blind study.
Gjallerhorn is a terrific budget amp too! I’m still shook by how good they sound.
. Probably my best spend is on the room but shared space....
