Chapter 15:
DAC in a Toilet Paper Roll
Shortly after we started shipping Bifrosts, Mike brought me something that would change the company.
“Take a look at this,” Mike said, handing me a Bifrost USB card.
I didn’t think anything of it. We were making Bifrosts, we had tons of Bifrost USB cards, they worked, and that was that. And I knew we weren’t going to be suddenly introducing a new USB card, only a few weeks after we announced the Bifrost itself.
“It’s a USB card. So?” I asked, not taking it.
Mike waggled the card at me. “Just look at it.”
I took the card and sighed. And that’s when it sunk in:
this USB card had RCA jacks on it.
I looked closer. The USB card also had a few more parts added onto it. I recognized one: an AKM4396 D/A converter.
“Is this a USB DAC?” I asked, incredulous.
Mike nodded his head vigorously and cackled. “Yeah, it is!”
“USB powered?”
“Yep! And it sounds really good!”
“Wait a sec,” I said. “I thought you were Mr. Anti-USB?”
“Yeah, but sometimes you just gotta say, ‘what the heck!’” (Except without the h and e, replace with f and u.)
An aside: this was really the beginning of Schiit’s ongoing “What” phase, where we’ll try a lot of different things—and, if we like them, make them into products. This is what got us Mjolnir, Vali, the upcoming Yggdrasil analog stage…as well as a shelf full of experiments that may never get turned into products.
Anyway, back to the frankensteined USB card. As I held it, three thoughts immediately came to mind:
- How inexpensive could this thing be?
- How do we keep the cost of the chassis from dominating the cost of the product?
- This is a whole lot smaller than anything we currently make, what would the chassis look like?
“What’s the BOM look like, cost-wise?” I asked Mike, using corp-speak for the Bill of Materials, or, in English,
All that stuffs ya gotta put together to makes it.
Mike cackled again and told me.
My head exploded. If we could get a chassis cheap—I mean, really cheap—we could sell that little DAC for $99.
Ninety-nine bones. That’s a whole different part of the market. I didn’t know how many we could sell, but I knew, even then, it would be a hell of a lot more than Bifrost.
“And we could put it in a toilet paper roll,” Mike said.
“What?” I asked, coming back to reality.
“Well, you know, for those guys who want something cheaper than a Bifrost. Here you go. DAC in a toilet paper roll. You don’t want to upgrade, you want cheap and disposable, we have cheap and disposable.”
I had a brief mental flash of a (thankfully) alternate future in which we’d be selling cardboard tubes with electronics inside of them.
“A toilet paper roll wouldn’t pass FCC,” I said.
“So wrap it in foil,” Mike suggested.
“No.”
“Or machine it out of aluminum tube.”
“No.”
“We can cut the little spiral on the outside so it looks like a toilet paper roll.”
“No!”
Mike frowned, looking offended. “You don’t like my new DAC,” he whined.
“Oh no, I like it just fine,” I told him. “Assuming it sounds good.”
Mike laughed and grinned. “Just plug it in.”
So I did. And it sounded good. Really good. I knew right there that this would be our next product. I knew we had to make it. It wasn’t a case of ‘why,’ it was a case of ‘why not.’
But not in a toilet paper roll.
The Challenges of a Changed Game
That first modified USB card completely changed Schiit as a company. Arguably, it’s the biggest factor in us moving from a “hobby business” to a “real company.”
But note when this happened: say, November 2011. Modi (and Magni) didn’t show up on the scene until late in December 2012—over a year later.
Huh? Our simplest products took over a year to develop?
What’s up with that, you ask. (And some are snickering in the background about how long it took to get a sellable Ragnarok. Hey, bite me. I almost did an April Fools announcement that the Ragnarok and Yggdrasil were cancelled.)
Bottom line: yes, developing even simple products can take a good long time—that is, if you want to get them right. Especially if they’re clean-sheet designs that move you into entirely new spaces as a company.
Modi itself was a challenge, on several fronts:
- It required an entirely new, and hella cheap, chassis design. When we started the development, I didn’t know if we could meet the very aggressive price point I set—especially without going to China.
- It had to be as simple as possible, which meant surface-mount and very easy assembly. We were already doing surface mount with Bifrost, but it was brand new to us at the time. Chassis-wise, even our insanely simple chassis for Bifrost, Asgard, etc were clunky and slow to put together—we needed something simpler and easier.
- It required huge production runs—in the thousands—much more than we’d ever done before. Huge production runs meant (comparatively) large investment—we had to be ready for it.
So what did we do? Well, let’s start with the one thing we
didn’t do: we didn’t sit back on our asses and say, “You know, we have a pretty good thing going here. Why take chances? Iterate the current products, keep milking the products for all they’re worth, and take the money.”
Because, like it or not, that’s what most companies do.
They’re scared. Risk-averse, they call it, in typical corporate doublespeak. Call it what it is. Scared. You’re scared to jeopardize your accomplishments to date. You’re scared of burning your profits on something that might not work out. You’re scared to step out of the mold you’ve made.
Hint: the mold you’ve made is a coffin. Break out, or die.
“Okay, fine, I get it,” you tell me. “Go ahead and write what you did, already.”
Fine.
Chassis: Set A Target, Stick To It. On the chassis side, the first thing we did was to actually set a price target. If we could bring it in at or under the target, we had a product. If it came in more than the target, we’d have to think again. This is the first time we’d set a price target on a chassis, instead of just sitting back and saying, “Well, let’s see where it lands.”
To maximize our chance of hitting the target, we put the chassis design up for bid at MFG.com, as well as with our two local chassis vendors at the time. Note we had two for a time. Note we’re now down to one. Corollary: cut your losses quick if things don’t work out.
It ended up that several vendors undershot our target price, some significantly. One of the local guys was under the target, and one was over. The local guy who was under the target price wasn’t the least expensive bid, but I knew already not to gamble with quality. We went with the one local guy, even though the chassis could be cheaper. Corollary 2: Don’t be frigging cheap. If you set a target price and your most preferred guy comes in under it, don’t grind them for the couple of bucks you might save if the long-shot new guy across the country works out. Just place the PO.
Assembly: Simple At All Costs. I’m doing this out of order, because the chassis design comes before the chassis. But you get the picture here. Assembly time is a function of chassis design. The simpler the chassis, the lower the assembly time.
We first experimented with variations on our current chassis—a U-shaped piece of metal wrapped around an inner sled. But, in that case, you were still talking 16 screws or so, between the ones that connected the chassis, the ones that connected the boards, and the ones that, well, connected the connectors. That’s a lot of screws.
That’s why we decided to have a “sled” design quoted—a new concept that used only a front “L” shape, rather than a U-shape. The main advantage of this is that it eliminated the bottom screws. The end result? 7 screws, not 16. I know, it doesn’t sound like much, but it makes a huge difference in production.
At the same time, we asked our vendor their opinion on aluminum versus steel—we were smart enough to know that painted steel was better, but we needed the vendor to say, “Yes, steel will be less expensive—and it can be repainted if it’s damaged.”
Refinishing a product was a dramatic new concept for us. As I’ve mentioned before, our other chassis, at least the aluminum parts, are junk when they’re damaged. Being able to refinish the chassis was a big deal.
And that’s how Modi ended up in a steel box, rather than in aluminum and steel: simple economy.
Production: Bite the Bullet. Yep, big runs are pricey. There’s no way around that. And that took us to another point that could change the company.
When you’re talking big production runs, you really have two choices:
- Save your own money. Funny, this is the way that businesses used to do it all the time. Seems it’s gone out of fashion today.
- Borrow the money. Go to the bank, get a line of credit, or get a loan against your inventory or receivables. This is what our accountant advised, citing all the normal reasons for getting in bed with a bank:
- You’re growing fast, this allows you to grow faster
- Keep your own money for other stuff, like building spaceship-styled campuses, Porsche GT3s, and vacation homes
- At current interest rates, it doesn’t cost that much.
Guess what we chose?
Yep, right in one. As in, #1. We chose to save our own money for this, because either (a) we’re a little stupid/touched in the head/out of touch with thems modern ways of doin things, or (b) we don’t think a bank should pay for our own dice-rolls.
Okay, I’m being flippant. But I have seen what happens when a company gets addicted to bank financing. Once you’re on the take, you don’t get off. And then the bank comes in and starts dictating what you should sell, and when.
No, thanks.
So, that’s my long-winded way of saying that there’s no easy way around the capital outlay necessary for large production runs. We bit the bullet and made the investment—an investment much larger than we ever could have made when the company started.
We were growing up.
Ego Talking
But, in my mind, there was something even bigger: Modi needed to make sense…as part of a whole line. Mike’s original idea was a tiny chassis (much smaller than today’s Modi) that could be used with any of our larger amps. But I had noticed—already—that people were stacking Asgard, Valhalla, and Lyr atop Bifrost…and they looked very good together. Which meant only one thing:
“We need a matching amp,” I told Mike.
“A matching amp for what?”
“The little DAC. We need a small, cheap amp to match it. $99 as well. A sub-$200 stack of an amp and DAC.”
Mike looked thoughtful. “How are you going to do any good headphone amp that cheap?”
I waved a hand. “I have a bunch of ideas.”
It turns out I shouldn’t have dismissed the “amp problem,” because that was arguably what set the Modi back a good 4-6 months. It took a long, long time, down many dead-end paths, before we had an amp we could pair with the Modi.
But that’s a story for another chapter.