2023 Chapter 14
More is Better
Let’s make this real simple: Yggdrasil More Is Better is the best-measuring multibit DAC in the history of the world. Period.
Yes, even at 20kHz.
And it even sounds good.
How did we achieve this feat? With a new TI precision DAC IC…and the prodding of some early listeners.
Here’s how it happened.
“We Have An Improved DAC11001”, Says TI
The original Yggdrasil More is Less used the TI DAC11001A. This is a great-measuring chip on paper, especially if you concentrate on the One Number to Rule Them metrics (full scale, 1kHz, 20k bandwidth). However, its sonics never impressed Mike or Dave, and as soon as GoldenSound tested it at higher frequencies, it quickly fell apart.
In addition, most reviewers thought it was the least best of the Yggdrasil flavors, including the least-expensive Less is More.
Sales were, well, less than stellar.
“Well, that’s that,” I figured. The high-measuring multibit DAC experiment was over.
Except TI ain’t dumb.
So, several months after launch, they came to us and said, “Hey, we have this improved version of the DAC11001A, the DAC11001B. Would you like to try it?”
Now, I knew the difference between A and B. Or at least I thought I did. Because Analog Devices does A and B grade DACs too.
Aside: confusingly, the B grade is higher in the precision DAC world. Yeah, I know. Talk to them; I’m just the messenger.
And here’s what I thought I knew: A and B grades were no different for what we were doing. Because we’d used both A and B grades of an Analog Devices DAC, and they measured the same in terms of THD, noise, etc in our products.
Needless to say, we weren’t very excited about TI’s new “B” grade parts. Because I expected the same results on the APx.
But TI was persistent.
First, they sent me a handful of pre-production parts to try out. Which is cool and all, but I was busy, and I was going back and forth between Texas and California, and they were in California, so they were only on my desk half the time, so I just, well, kinda ignored them.
But TI eventually reminded me that we had some of their parts, and we should really try them out, and that they had improved the performance at high frequencies.
Now I paid a bit more attention—because performance at high frequencies was the DAC11001A’s bugaboo. But how much better could it be? I mean, the DAC11001A was baaaaaad.
How bad? This bad:
As in, the DAC that measured -118dB THD+N at 1k/20k bandwidth was -48dB at 20k/90k.
Aside: to be fair, it’s a -96dB part at 90k bandwidth, because that’s the way measurements work—the wider the band, the lower the number. -96dB is ****ing fantastic at 90k. It’s just that losing nearly 50dB of performance from 1k to 20k is a problem.
And so I went into this whole thing thinking, well I’ll just swap the parts and see what they mean by “better” and when it’s not better I’ll tell them sorry guys we’re not interested and that will be that.
So I swapped the parts.
And put it on the analyzer.
And saw this:
Yeah. Right. Better. Like 47dB better.
FORTY SEVEN dB.
47!!!
That’s…pure bonkers.
TI saying “yeah this is better,” all low-key like, is kinda like saying “yeah a hydrogen bomb is better than a firecracker when you’re looking to blow things up.”
I mean, when I shoot for improvements, I’m pretty happy with 6-10dB.
6-10dB is significant. Especially if it’s in noise floor, that’s stuff you can hear.
20dB? Overjoyed.
Forty-friggin-seven dB? I didn’t know what to think. That’s like, higher than the gain on a MM phono preamp. That’s, like, 30,000% better. That’s…just hard to take in.
Now, some people are saying, "Well, that doesn't look like One Number to Rule Them performance!"
Right. Yep. Because One Number to Rule Them is done at 1kHz, not 20Khz, and bandwidth limited to 20kHz. When we do apples to apples, here's what we get:
So there you go.
But the real question was: what did it sound like?
The Acid Test
So I took it home and had a listen. And, oh boy, was it different.
In comparison, the old MIL with DAC11001A chips sounded, quite literally, broken. The highs were strange and strained, which was perhaps not surprising considering that their measured results were getting scarily close to transducer distortion levels, while also exhibiting some noise floor modulation.
Was it the best Yggy I’d heard? In some ways yes. In terms of having a dark background, in terms of purity of highs, in terms of overall cohesion, it was better. Dynamics were better on OG and LIM. But OG was a bit gray, and LIM was a bit oversimplified.
The surprise was that it was certainly playing in the same field as those other DACs now!
Aside: when we first introduced the Yggdrasil flavors, we did a real, no-kidding, double-blind listening setup with selected reviewers. No-kidding as in, the guy who set it up wasn’t on site at all, and nobody—not even me—knew what was what. The results were in a sealed, obscured envelope. Before the test, Mike and Dave were vehement about their dislike for MIL, and I thought it would be very funny if they chose that as the best when listening blind. In fact, I almost expected it. What happened was that they went in after the reviewers, sat down, stood up 90 seconds later, and said “X is this, Y is that, and Z is the other thing.” And they were 100% right.
Aside to the aside: and no, the guy who set it up wasn’t Mike or Dave’s friend.
So I brought the Yggy DAC11001B prototype back into the office, where our early panel had a listen and pronounced it good, if a little “twee,” in one person’s words. From there, it went to a couple of other early listeners.
One early listener was shocked how good it was, and agitated for me to bring it out as another Yggdrasil variant. Another was Dan Clark, who loved it and has used it at shows over all other Yggdrasil flavors.
And so, over a fairly short period—weeks, really—we decided to do a short run of Yggdrasil “More is Better,” with the DAC11001B.
Which, as I spend more time with it, is really a fantastic DAC chip. TI is really going for the crown in the precision DAC arena, and this chip really delivers. Even better, it’s less expensive than the Analog Devices AD5791B (though still fairly insanely priced at $62ish in 1000 piece quantities—and we need 4 per Yggdrasil MIB).
Your Chance at MIB
And so now you have another choice in the Yggy flavor wars: Yggdrasil More is Better. It’s not an inexpensive DAC, clocking in at $300 more than Less is More. And, to start, we’re only selling complete DACs, but if you give us a few more weeks, we should have enough available for upgrade boards as well.
Is it for you? Maybe. I consider it the best all-around Yggdrasil.
Mike still likes LIM. Dave still likes OG. But OG is going away. The availability of the Analog Devices AD5791 isn’t confidence inspiring, and its price/performance is now very poor. So the future will be LIM or MIB. Both fantastic DACs. Both TI based. Your choice.
“Will there ever be a ‘More is More?’” someone will ask.
Of course, we’re still going to try to do an ultimate Yggdrasil. But there’s nothing ready for prime time yet. And, consider that if it’s ultimate, it may cost significantly more.
In the meantime, I hope you enjoy MIB, the latest iteration in the evergreen Yggdrasil.