2019, Chapter 12:
Passing the Torch
Some people are gonna hate us.
“Oh no you didn’t! No way you changed Bifrost to Bifrost 2! Heresy! The whole point of Bifrost is that it’s upgradable! You’ve betrayed us! What’s the point of upgradability if you go and throw it down the toilet? Why, just why? You said there would never, ever, ever, ever be a Bifrost 2!”
Deep breath.
Believe it or not, the world will still turn, the sun will still shine, and our tiny, tiny corner of audiophiledom will go on pretty much as it always has, even as we replace Bifrost with Bifrost 2.
(And, sorry guys, I never said there would never be a Bifrost 2. At least I don’t think so. If I did, sorry, I ain’t perfect. And Apple was once going to do a wireless charger. Life isn’t always 100% predictable.)
“But I still don’t get it,” you say. “If Bifrost’s upgradable, why is there a need for Bifrost 2 at all? What’s so much better about Bifrost 2 that it had to be done? And what does this mean for all of us who already own a Bifrost?”
Great questions. Let’s attack them in order.
Time for a New Generation
It’s funny. There’s more and more evidence piling up that humans (and many animals) are simply programmed to die. As in, eat well, work out, live the best and safest and most perfect life with 120 different supplements every day, and an internal clock is still gonna say, “Ya know, bub, it’s time to move aside, time for the kids to have their chance.”
Why do I mention this? Because it’s a bit like that with Bifrost. Bifrost was conceived, quite literally, in the Garage Era. Bifrost was also conceived in (largely) a vacuum. Other companies weren’t really focused on upgradability, especially at Bifrost’s price. So, we guessed at what features it would need to be upgradable:
- USB board (because, at the time, USB was more than a bit sucky, and also because it was changing fast). .
- Analog board (mainly because Mike was planning a Bifrost Multibit, even back then.
And, yeah, so the boards were internal, and you had to take the whole chassis apart to change them. Not a huge deal, right? I mean, we were a tiny company. Hardly a company at all. And, I mean, nobody was really offering upgradability, at least nowhere near Bifrost’s price. So that was fine.
And, yeah, the clamshell chassis meant that the backside I/O was fixed. We wouldn’t be able to add an input that was different than USB, or additional outputs like XLRs, without throwing away the entire inner chassis. But how much would that really change? Again, not a big deal, when you’re entering a market where there was literally no upgradability—at all—anywhere near the price point.
The result?
Almost universal accolades. Stereophile even called it “the highest value product we have ever reviewed, period.”
And, with Bifrost’s upgradability, we were able to move to a much better USB input in just over a year after introduction (Gen 2), then move forward again with Gen 5. The original analog board brought the introduction of the improved Uber Analog board, and then the Bifrost Multibit and 4490 variants. In 8 years, Bifrost has evolved a lot. It’s a fundamentally different product.
And we accomplished this without requiring you to buy 5 new DACs along the way, like you would with a non-upgradable DAC.
“So why change it now?” you ask. “Come out with a new analog board, a new USB board, and call it a day!”
Well, here’s the thing: you didn’t see all the pain in Bifrost’s evolution.
In the process of getting from the original Bifrost with Gen 1 USB input to the Bifrost Multibit with Gen 5 today, a whole lot of stuff happened under the hood.
The most significant hidden change was a new motherboard. When Bifrost Multibit was being developed, it turned out that the old motherboard didn’t really cut the mustard when Multibit’s additional power requirements hit. So, every Bifrost Multibit had to come back to Schiit for upgrade.
Yes, that’s right. No self-install. Cue the howling.
And yes,
I get it. it’s fundamentally a crap situation to have to send your DAC back to our shop for an upgrade.
But there was more.
Bifrost Multibit had to have the DSP added on to the DAC card itself, because it was never intended to have DSP on the motherboard. Bifrost also stuck with the older SPDIF receiver, because it was on the motherboard. And every time we considered analog changes, the lack of holes for XLR outputs grated. Especially with the introduction of Jotunheim, a balanced amp that could sit happily atop Bifrost…
if it had balanced output.
And, as we considered inputs other than USB, we knew—Bifrost had a fixed hole that was USB sized, and that was it. We’d have to start replacing metal if we wanted a different input. No bueno.
Ironically, our frustrations came to a head not during a Bifrost upgrade, but during the introduction of Yggdrasil’s Analog 2 boards. It was that experience that profoundly shaped Bifrost.
And by “profound,” I mean profound.
As in, Bifrost 2 almost became completely
non-upgradable.
The Thermonuclear Option
In the midst of the furore about the Yggdrasil Analog 2 upgrade (do I have it, do I deserve it, why does it have to come back to you, why do you want my analog boards, what’s the real difference, it looks the same, etc...), I was pissed.
“I don’t know why we even bother doing upgradable DACs,” I told Mike. “If we’d just introduced an Yggdrasil 2, we’d probably have less complaining. They’d just have to buy a new DAC.”
“But upgradable is the
rightway to go,” Mike said.
“Tell that to the customer who has to be without their Yggdrasil for three weeks, because they have to ship it here for an upgrade.”
Mike opened his mouth to respond, but I cut him off.
“Look at the forums. There are a ton of people buying new Yggys and selling their old ones, because it’s easier and faster than waiting for the upgrade.”
Mike nodded. “Yeah, that’s bad.”
“What happens if we do this to Gungnir? Or Bifrost?” I asked.
Mike went a little pale. Because, like me, he knew that changes like the Analog 2 boards were hard to explain. It wasn’t like the old days when they were introducing new DACs every year. These upgrades might have the same exact DACs. That looked weird. Couple that with the cost of the DACs themselves, and you had an untenable situation.
“Someone who just spent $500 to upgrade their Gungnir to Gungnir Multibit might not be thrilled about paying $500 again to upgrade to Gungnir Multibit Analog 2,” Mike said, reading my mind.
“Bingo.”
Mike sat and thought for a while. Finally, he asked, “We have a really old Bifrost that needs to be updated. What would you do?”
I frowned. “Full thermonuclear option,” I told him. “Not upgradable.”
Mike shook his head. “But...it’s
wrong.”
I sighed and pushed on, full of vitriol and certainty. “Maybe not for the dude buying a $500 DAC. Maybe the RIGHT thing to do is just make the best product we can, and not worry about upgradability.”
“Even USB?”
“Even USB,” I said, not listening to the little voice in the back of my head that whispered,
That’s a really ****ing dumb idea, boss.
Mike sat silent for a long time, to the point where I thought he might get up and leave the room. We have our arguments, mostly about delta-sigma (I think that we should have a reasonable number of delta-sigma products, Mike would rather not have any), but usually the arguments end up being productive.
“And put XLRs on it,” I added, into the silence.
“Cheap-ass balanced,” Mike said.
“Or real balanced, the AD5547s aren’t all that expensive.”
Mike sat and thought. Finally he said, “Okay.”
And that was that. I dropped it. I didn’t know if Mike’s “OK” meant “OK, let’s do it,” or “OK, you’re an idiot.” I didn’t know if he would make a Bifrost 2 or not. I didn’t know if he would make it upgradable or not.
Some weeks passed.
Eventually, Mike came in with a board. “Here’s your cheap-ass balanced, wrong-way not-upgradable Bifrost 2,” he said.
And, holy Schiit. There you go. A single Jotunheim-sized board, with no risers to be seen. No plug-in boards. Gen 5 USB. XLR outputs. And...
“Wait a sec, that’s not an AD5547,” I said, referring to the Bifrost’s DAC chip. “That’s a AD5781, like Gungnir.”
“Fifty-seven-eighty-one-A,” Mike said, referring to the lower grade of the 18-bit DACs we use in Gungnir. “Just two of them, plus cheap-ass balanced.”
“What does it sound like?” I asked.
“Have a listen,” Mike said, grinning.
To cut a long story down to size, it sounded damn good. Much, much better than you’d expect. Really a mini Yggdrasil.
What’s more, it measured...ah, better than any True Multibit product we’d ever done. In fact, balanced and single-ended measured so similarly I thought Mike had connected the two outputs together. But he hadn’t.
(Imagine—
same performance from balanced and SE. Yes, Bifrost 2 works like this. Of course, the measurements still won’t win any awards for THD when compared to delta-sigmal solutions, as I’m sure some reviewers will remind us. But they are impressively linear and noise-free.)
“Holy Schiit,” I said. “Price?”
“Same as Bifrost.”
“Holy holy Schiit!”
And for a moment, all was right in the world. Non-upgradable was the way to go. We’d introduce a Bifrost 2, and after some blowback about how we abandoned our upgradable platform, we’d be free to move on to a Bifrost 3 and 4 and 37 and whatever, whenever there were actually meaningful upgrades. The back panel would no longer be a sticker-fest! The sun would shine on a dog’s ass!
But still, something bugged me. I didn’t say anything about it at the time, but I knew it would come back and bite us in the butt:
USB.
Even then, Mike and team were working on the new USB that was to become Unison. What happened when that showed up? A running change? Oh hell that’s a whole ‘nother circle of doom. A Bifrost 3?
Eventually, I broached the subject. “Mike, I think we need to have USB upgradability.”
Mike nodded. “I figured.”
“Then why didn’t you say something?”
“You were all-fired on killing all upgradability, what am I gonna say?” Mike asked. “I knew you’d come around eventually. I’ll add the USB card back in, and we’re good.”
“Yep,” I said.
And again, for a while the birds chirped and the unicorns farted, and all was well.
But eventually, doubts crept in. If we stuck the USB board into the chassis like Bifrost, you’d still have to open it up to change the USB board. That wasn’t ideal, because we couldn’t advise someone to simply open up an AC-powered product and poke around inside of it. We’d have to wrap it in 7000 layers of legalese so we wouldn’t be sued into oblivion when someone grabbed onto a live wire.
Still, it was better than what we had. I was OK with it. At least until the fateful day when Alex was up in my office, and we were commiserating about the (ongoing) pain with the Yggdrasil Analog 2 upgrade. I told him that he could look forward to a future where Bifrost wasn’t really upgradable, and that would simplify things a lot.
That’s when he said something that would change the course of the company.
He said:
“Upgrades aren’t the problem. The problem is that
we have to do them.”
I rocked physically back in my chair. Because that was
exactlythe problem.
If we didn’t have to do the upgrades—if the product was
designed to be upgraded in the field, without opening it—then we could do all the friggin upgrades we wanted!
“What’s wrong?” Alex asked.
“Not wrong,” I said. “Right. Ask me what’s right. Or at least ask me in a few hours, after I do some drawings.”
Alex looked puzzled, but accepted my explanation and moved on.
I, however, went into high gear. I’d looked into doing slide-in cards for Jotunheim’s modules back in the day, but I’d abandoned the design. It was costly, it used custom connectors, and required brackets and daughterboards and would have made the available space on the card even smaller. Not a great solution.
But if I thought horizontally rather than vertically, and if we stamped the brackets so they were inexpensive, and if I could find some connectors that didn’t cost a trillion dollars...then maybe the cards could slide in from the back.
Which would then mean the chassis
wouldn’t have to be opened.
Which would then mean
we had our upgradability back!
Which would also then mean we’d have
true, in-the-field hardware upgradability in an inexpensive DAC!
And, here’s the thing: suddenly I realized we had all the tools to do this. I found the connectors in the PCI-E parts bin, for literally 1/15 the cost of the custom connectors we’d looked at before. I used the 3D printer to prototype the brackets, and sent out for full 3D printed chassis so we could see how they all fit together. I flipped the add-in cards upside-down, so there would be plenty of room on the motherboard for DSP and I/O. Dave did a layout with all the modules in place, and worked out all the control system between them. He even did a Unison module in the new form factor.
In a month or so, we had a working prototype.
Best of all, it retained all the best qualities of the original, non-upgradable version, except that it was fully modular. And it added a couple of new features—namely remote control and phase inversion.
Aside: yes, I said remote control. Yes. Bifrost 2 has remote control. It also has balanced outputs and phase inversion. No, the remote doesn’t control volume. Just input select, mute, and phase.
Mike went to work on a BOM and I went to work finalizing the chassis. This was gonna be a killer product!
All done, right?
Well, not yet. As we were finishing up the design, one thing bugged me: the 8-pin DIP we used to contain the firmware for the DSP. That was a crappy part, a relic of a forgotten era when pins went through boards, and it was obsolete to boot (we bought several tens of thousands of them, and were investigating doing a daughterboard for a surface-mount part). Barf. Time to kill that.
I sent an email to Mike and Dave, letting them know we were missing one key component to a fully autonomous DAC—the firmware DIP. Other companies allowed you to upgrade it with USB, and some even made great hay about how their products were upgradable for free. To really make sure that we never had to bring Bifrost 2 back ever again, we really needed some other way to upgrade firmware.
Mike and Dave agreed. Dave schooled me a bit on the logistics of upgrading a multi-board DAC—in short, USB might not be the best way to do it, and there was the real question of what needed to be upgraded—DSP firmware, operational firmware, USB firmware, all of them?
In the end, we settled for a microSD card slot, plus a warning that it wasn’t for playing your music files.
Now, if we have firmware updates, we can just send you a microSD with the new card, or you can download the firmware onto your own card. Then just plug it into the back of your Bifrost 2, and it’ll determine what needs updated and get ‘er done.
And there you go. Full hardware and firmware upgradability...in a $699 DAC. That’s what we’re calling Autonomy™. As in, once we ship it, it’s fully autonomous. You’ll never have to send it back to the mothership for upgrade.
Plus, as usual, our own unique True Multibit™. Yes, now with a TM as well. Because it’s time to differentiate our 16-20 bit DACs with unique time-and-frequency-domain optimized digital filter from the rest of the options out there. And also, because unlike the old Bifrost, there’s no delta-sigma option for Bifrost 2.
As Mike would say, “That’s the
right way to do it.”
What About Grandma?
“Well, that’s great and all, but I have an old Bifrost. What are you gonna do for it? Shuffle it off to a hospice?”
In short, no.
Here’s what’s gonna happen:
- We’re going to continue supporting Bifrost for the foreseeable future. That means “for the full term of your warranty, plus for as long as parts are available.” So, technically, the last Bifrosts will fall out of warranty in 2024. But that’s not the end. Heck, we still fix 9-year-old Asgards today.
- Bifrost will continue to get upgrades whenever feasible. For example, when we announce the Unison USB upgrade for existing DACs, you’ll be able to upgrade your Bifrost with Unison USB. Furthermore, the Unison upgrade on Bifrost does not require the Bifrost to return to the mothership for the upgrade.
Now, that doesn’t mean that Bifrost will get all the upgrades that Bifrost 2 gets. The AD5781 analog card, for example, simply isn’t feasible. But as long as we can do upgrades,
we will.
“So I have a Bifrost, what do I do?” you ask. It’s simple:
as much or as little as you’d like.
Option one: you can continue on with Bifrost. It’s a great DAC in all of its incarnations, and has won many well-deserved accolades. There’s no need to upgrade. Plus, it supports all current versions of Windows, including Windows 7. Bifrost 2 is for Windows 10 only (and MacOS, and Linux, and iOS, and Android). Enjoy your Bifrost for many years! You also will have some upgrade options that will continue well into the future:
- Upgrade to USB Gen 5 now, if you have an older USB card or no USB input
- Upgrade to Unison USB when it’s released (most likely after Windows 7 is officially dead, since there are no Windows 7 drivers).
- Upgrade to Bifrost Multibit, if you have a Bifrost, Bifrost Uber, or Bifrost 4490
Option 2: You can move on to Bifrost 2. This is probably most useful if you want the balanced outputs, but it also may be worth it for the remote control, depending on your situation. But you should keep a couple of things in mind:
- Bifrost 2 is the first product with Unison USB, which does not support any Windows OS other than Windows 10. As in, the OS has to be fully UAC2 compliant. This means you’re fine on Windows 10, MacOS, Linux, iOS, and Android. Just not Windows 7. Windows 7 is officially dead in January 2020, but it’s not officially dead yet, so if you’re dead-set on using an older version of Windows, Bifrost 2 is not for you.
- It’s not like we have upgrades planned for Bifrost 2 for next month, or even next year. The intention is not to upgrade unless there are meaningful improvements in the technology. And, beyond USB (or maybe another input in the future), advancements in True Multibit technology are few and far between. So it’s entirely possible that you won’t see much change in Bifrost 2 for a while.
- If you bought a Bifrost or Bifrost Upgrade from us, or from an authorized reseller, your price on Bifrost 2 is $599, not $699.
“Oh hell there’s this new Autonomy thing now in Bifrost, what about Gungnir and Yggdrasil? What if I’m considering those DACs now?” You ask.
Okay, fair question.
Fair answer:
we don’t know. Not entirely.
Depending on how Autonomy is received, it may or may not find its way up the line. However, it almost certainly
won’t result in an Yggdrasil 2. Yggdrasil is a significant investment. We want to protect your investment. So, while we will be working to ensure that as many Yggdrasil upgrades as possible don’t require a factory visit, we’re not a fan of burning the current architecture to the ground and starting over. So, I’d expect that if you’d like an Yggdrasil with remote control, you won’t have to sell your current DAC in order to get it. Eventually.
And yes, all of our upgradable DACs will get Unison USB as an upgrade when Windows 7 officially dies. Again, remember: no drivers for Windows 7, so that upgrade may not be for you.
But, again, a warning: if anyone thinks we operate with a Communist-style 5-Year Plan carved into stone tablets and completely inviolable, you’re, well, exceedingly misguided. The future may change. And we may choose to change with it. As always, there are no guarantees, other than the current products on our site (that are in stock) and the prices in the cart (plus applicable local taxes and customs.)
So, the time of Bifrost is over. It’s time to step aside for Bifrost 2. A much more capable platform—a much more upgradable platform—a better platform for today, and for the
foreseeable future.