[Stevko]

Still waiting for Schiit news
Miss my Hel 2.
Considering buying the «new» 2E

 

[RCBinTN]

That's what hammers were invented for, right? And for bigger problems, sledgehammers.

Beat to fit ... paint to suit

 

[Jason Stoddard]

2022, Chapter 3
Turd Polishing?

Ah, come on, you’ve heard the saying: “Ya cain’t polish a turd!”

Graphic. Visceral. Precise.

You know exactly what it means, don’t you?

Aaaaaannnnnnnndddddd…that’s nice and topical. Well-targeted. Precise. Because today we’re going to talk about something that most people think is stone-age, ancient, flawed, beyond redemption, and surpassed by everything from ear tampons to five-figure streamers.

Yeah, we’re talking vinyl.

As in, stuff you press a sound wave into, scratch a rock across, and boost the output to hear. Or, more specifically, we’re gonna talk about the boost side, because we’re done with the scratching-a-rock-across side (sorry).

Because the boost side is important. I mean, you’re talking 100x to 1000x that tiny signal coming off that rock and cotton-candy wire apparatus, lovingly wound by crazy Japanese or equally insane Brooklynites. And getting that 100-1000x with low noise and low distortion is one of the most interesting challenges in audio.

Now, in the past, I didn’t think this was the case.

But then there was Mani.

And I wondered what I could do better. I took a hack at a simple version on the Ragnarok 2/Jotunheim 2 phono card, but it took me, what, 6 full years to really conquer the original Mani.

Because, let’s face it: Mani was good. Damn good. Crazy good for an affordable phono preamp. It earned tons of accolades. It was solid, reliable, and pretty darn high performance. It was clearly one of the best of the inexpensive phono preamps. And when you have one of the best, there’s only one path:

To create clearly the best affordable phono preamp…period.

That’s Mani 2.

The Marketing Spiel

You can skip this if you think you are super-susceptible to marketing. If you don’t think so, read on.

Here’s the spiel: Mani 2 is simply the highest-performing, most-flexible affordable phono preamp. Period.

Yeah. I know. Your arms are crossed. You look doubtful. That’s wayyy too simple, you say. It can’t be that easy, you say.

But let’s break it down:

Highest-performing: have you seen near 100dB SNR and -100db THD+N for a phono preamp running typical MM gains—in this case, about 44dB? Yeah, there you go. Oh, and by the way, we’re doing this with a passive RIAA network—not feedback. Plus, it’s +/-0.2dB from RIAA on all settings, verified on an APx (525 in this case, too lazy to take it to CA to measure on the APx555, sorry.)

Most flexible: have you seen a $149 phono preamp with 4 resistive and 4 capacitive loads, both selectable individually? Plus selectable 6dB/octave or 12dB/octave passive low cut to help with rumble? Yeah. Didn’t think so.

Affordable: it’s $149, in case you missed it. Which is less than the inflation from 2014 to today, by the way.

So there you go. Spiel is facts-based. What a shock.

Hint: all of our spiels are facts-based, and we never talk about how products sound. Doubt us? Read the site. Be surprised. It’s OK. We know you’ve been lied to.


Path, Part 1: The Card

So how did we get to this world-beating phono preamp?

Well, of course it started with Mike, saying “hey, we need to do an affordable phono preamp,” and then doing it, based on a topology he used to use back in the tube phono preamp days, but with op-amps. That was Mani. Mani was damn good. Mani was hard to beat.

Then, me, looking at Mani kinda cross-eyed and saying, “Well, this thing is kinda complicated,” and doing a simpler version for the MM-only phono preamp card that went into Jotunheim (and later, Ragnarok 2).

Aside: decoder ring for people not super-into vinyl:​

​

MM: Moving Magnet. AKA “most cartridges out there.” Fairly high output. Only needs a gain of about 100 or so.​

​

MC: Moving Coil. AKA “audiophile alert.” Typically lower output. Needs maybe a gain of 1000, plus may need specific loading.​

​

MI: Moving Iron. AKA “Grado.” Can be high or low output. Can be high or low cost. Talk to Grado. Ain’t our bag as far as ‘splainin. Though we do like a lot of them.​

Then, me, saying, “we should do this Mani 2 thing serious-like, and do a fully discrete version, and all that.”

Now, before you get all bonerific and excited, fully discrete phono preamps can absolutely be better than ones made with IC gain stages, but they are typically (a) very complex, and (b) somewhat variable. The first Mani 2 prototype, a fully discrete one, was both, and had the added disadvantage of not having enough space for the massively paralleled first gain stage I wanted to do. I learned a lot doing it—not least of which is that I didn’t feel like doing a discrete phono stage in the Mani form factor.

Aside: the reason discrete can absolutely kill here is due to Johnson noise—no, I’m not making that up—which you can minimize more effectively in a discrete design using lots and lots of parts than in an IC.​

Plus, discrete gain stages can only take you so far. One of the interesting results of the early prototype tweaking was that the passive RIAA network had a bigger effect on sound quality…as in, the resistors and capacitors in-between the gain stages were arguably more important!

So, after the first not-so-hot discrete prototype, I decided to get serious. As in:

1. Take a look at the best of the obtainable op-amps out there.
2. Play with the best resistors and capacitors for the passive RIAA I could find.
3. Experiment with how the power supply interacted with the gain stages, etc.
4. See how feasible it was to add additional loading.
5. Look into other features people might want.

Path, Part 2: The Prototypes

I started with two prototypes—one very much akin to Mani 2, but without the LF filter the production one featured, and one balanced version, in a Magnius/Modius sized chassis.

“Balanced!” some are saying. “Oh yeah serve that up, that’s what I want!”

Welllllll…or not.

Here’s the thing. When I first mentioned balanced phono preamps to Mike, he looked confused. “Are they doing balanced cartridges now?” he asked. “Like, 6 wires from the headshell? I’d have to re-wire.”

Because, contrary to the “all cartridges are balanced” “wisdom” of the current age, Mike is going by the old-skool definition of balanced. As in, positive phase, negative phase, and center tap (ground.) And by that definition, there are no balanced cartridges—at least none I know of. Maybe for a house-priced phono rig? I don’t know. I’d buy a house, not an overengineered fancified device to spin super-outdated, flaws-baked-in niche media.

“Wait a sec! I thought all cartridges were balanced!” someone is crying.

Yeah. Go back. That big block of type up ahead of your cry? Re-read that.

“Well, most people say cartridges are inherently balanced because they don’t have a ground reference,” I told Mike.

Mike looked at me skeptically, as if he was trying to figure out if I was kidding.

“No, seriously,” I said.

Mike sighed. “No center tap, not balanced. Yeah maybe you could kinda sorta fake it, but you ground one end of the cartridge for a reason.”

(Actually what he said was a bit less complimentary, so I cleaned it up a bit. You get the idea.)

“But I figured we’d do a prototype of a balanced preamp, you know, maybe just in case—”

“Maybe if you want 6dB more noise,” Mike grumbled, citing another problem with balanced designs—as in, they can be noisier. Common-mode noise may drop out (like power supply garbage) but thermal noise—Johnson noise—will be higher for an equivalent implementation. This is why most MC step-up stages are very, very simple—as in, unipolar and single-ended.

“If the network impedance is the same,” I told Mike. “But in this case, the network impedance is even lower than Mani 2, and that’s a lot lower than Mani, so it should be quieter.”

Mike looked at me doubtfully. “Let me know how it goes.”

Here’s how it went: the prototype that became Mani 2 measured very well, significantly better than Mani. Listening tests had us choosing Mani 2 every time. Very good. The balanced “Manius?” It measured even better than Mani 2! Really nice.

However, those measurements were from the analyzer. In practice, hooked up to an actual cartridge, the balanced prototype frequently had noise issues. Hum, buzz, and sometimes rhythmic clicks and other strange sounds issued from the speakers.

“Because it ain’t really balanced!” Mike said.

Maybe. Maybe it was just a bit unstable. It was based on differential op-amps that we’d used before to great result, but who knows how stable those were when hooked up to a cartridge?

In the end, we decided not to pursue the balanced/Manius prototype, at least in part because there are really no actual balanced catridges, but more because we’d want to improve the feature set if it was going to become a real product—as in, have front-panel adjustment of load and gain, with even more options for load. That—and the oversight to make sure front-panel, real-time adjustment is totally safe—meant it was a totally different product. And, if we decided to do a higher-end phono preamp, it would really have to duke it out with a discrete design and maybe even tubes.

Aside: not tubes so much anymore. Tube phono pres are rough. And with tubes getting harder to supply, that’s probably never in the cards.​

​

Aside to the aside: again, the balanced/not-balanced/higher-end phono pre is a notgonnahappen.com deal. We are not working on it. There are no plans for one.​

But the Mani 2 prototype—no noise problems. The lower impedance RIAA network, the simplified gain stage with DC coupling, all that worked fine.

Aside: Mani 2 eliminates all 3 coupling capacitors of the original Mani design, as well as one gain stage per channel—it’s now much more akin to the Jotunheim/Ragnarok board than the original Mani. It also significantly increases the power supply rails—from +/-5V to +/-16V, for over 3X the voltage swing capability and much less chance of overload. And, finally, it increases the number of loading options, from two resistive and one capacitive in the original Mani, to four resistive and four capacitive in the Mani 2. Oh yeah and the one- or two-pole LF filter…​

So was the prototype of the first Mani 2 perfect?

No. As I said, it didn’t have the low cut filter, mainly because I didn’t want to add another gain stage, and I hadn’t looked into the feasibility of a passive filter at the time.

It also had significantly different noise performance between the two channels—as in, one channel had much higher power supply noise. This meant I needed to take a look at the grounding—which meant another prototype.


Path, Part 3: The Details

The second Mani 2 prototype had much better grounding…

…but still didn’t perform to expectation.

Now, there’s a difference in power supply noise between both channels of the original Mani. That’s not super surprising, especially when you’re looking at almost 1000x gain. But Mani 2 should be better…and I knew I could do better.

So I did a third layout. I was about to send it out when my thoughts about a low-cut filter came to a head.

Here’s the thing: many phono preamps have a “rumble filter.” Most of these are fairly steep, op-amp based filters that work just fine to filter out subsonic noise, but also throw another op-amp based gain stage in the path…one that has significant phase shift in the low frequencies to boot. This trade-off led to a bifurcation in the design of phono preamps, a “high road” and a “low road” approach:

• High road: you don’t need a subsonic filter, you heathen! Get a better turntable, or better records, and make sure you use a record weight, and for the love of god make sure there are no resonances between your cartridge and arm fundamentals, and everything will be fine! (AKA “we leave out the subsonic filter because it sounds kinda ass-y, so better optimize your system, bro.”)
• Low road: yeah we have a subsonic filter, here’s the chart, look at how much it cuts down on low-frequency wobble, so if your woofers jiggle like jell-o, switch it in and it’ll stop all the boogie. (AKA: but we know it sounds kinda like ass, but better to throw it in so you won’t complain about it.

I realized that what I wanted was a “Mid road.” As in, something that helped, but would still be sonically transparent—preferably without another gain stage.

And I realized that I could do just that with a passive filter. I mean, the op-amps had plenty of current to drive a fairly low-impedance passive stage, and I could even cascade two of them for more LF rejection (though at the cost of some overdamped droop.)

Aside: cascading two passive filters will have more droop in the passband, because it will end up being extremely overdamped. You can see this in the Mani 2 APx report, showing both the one-pole (not cascaded) and two-pole (cascaded) filter characteristics. I consider this to be a fine trade-off (and it harkens back to literally one of my first circuits, a passive filter for bi-amping car stereo systems, when bi- and tri-amping were just getting started). You may not consider this a fine trade-off. I get it. Tomato, to-mah-toe.​

So I added a switchable one-pole or two-pole passive filter on to the third proto, in addition to completely changing the layout and re-grounding it.

Aaaaahhhh.

Now the channels were within a couple of dB in terms of power supply noise.

And the filter worked very nicely, too. One pole, or 6dB/octave, had virtually no effect at 20Hz, but cut down subsonic woofer excursion nicely. Two poles, or 12dB/octave, yeah, you lost a couple dB at 20Hz, but subsonics were way, way down. All 100% passive.

So we were done, right?

Well, not quite.


Flushing the Flyshit

Shortly before it was time to send Mani 2 to production, Alex looked grim. This was in the spring of 2021, when the global supply chain ****ery/parts shortages were starting to get stupid.

“Your flyshit regulator,” he said. “That’s a problem.”

The flyshit he was referring to was a neat TI voltage regulator that combined both positive and negative rails into a single part. Neat, but unfortunately not quiet—it needed to have additional filtering to get rid of the HF noise coming out of it. Not dissimilar to the ancient LM317/337 parts we’ve been using since the earth cooled.

“Problem as in problem, or problem as in supply?” I asked.

“Both,” Alex told me. “Our PCB assemblers don’t like them, because with your large thermal vias, they need special fixturing or touch-ups—”

“But they need the thermal vias—” I began, interrupting Alex.

Alex held up a hand. “And they might be hard to get on a continuous basis, starting now.”

Crap.

Well, that was that. I needed to redesign for the older regulators—and that meant getting new prototypes to confirm the performance of the product, and possible delays in getting it to market.

“How far out are we?” I asked Alex.

“We’re fine, as long as we don’t let it sit,” Alex said.

So I didn’t. I had a new layout that day, and new boards in a week. A hand-stuffed prototype worked fine—it actually performed better than the blindingly expensive flyshit TI part—so I ordered a set of 10 boards for first article production.

And that brings us to this chapter. Because, after verifying the 10 boards, we’re off to production. And I figured best to write this while we wait.

Especially after a funny anecdote. I was running one of the 10 first article boards on the APx we have in Corpus Christi, and we had a new guy walk by when I was doing it. He saw the results, and did a double take.

“That’s…a phono preamp?”

“Yep,” I said.

“Did you know you have probably the best-performing phono preamp out there?”

I shook my head. It was a lot better than the original Mani, but I hadn’t thought of it in those terms. So I started looking around, and discovered…holy moly, he was right! And that was comparing our results to stuff measured with the inputs shorted. As in, not connected to anything except a shorting plug. That’s not a realistic use. Every phono preamp is connected to a cartridge via cables. They aren’t shorted.

Now, we do measure versus 2V RMS out, because (a) this is standard line-level, and (b) Mani 2 has tons and tons of headroom—with +/-16V rails, it has over three times the swing of Mani. So maybe I should do a cheat sheet for evaluating phono preamps, which usually have something like this for specs:

Signal to Noise Ratio: >80dB, A-weighted, inputs shorted, referenced to 1V RMS, MM gain.

Let’s break that down:

• >80dB. That’s the raw measurement. Meaningless in itself.
• A-weighted. This is a weighting that suppresses low frequency noise, which typically results in better measurements for products that have high power supply noise, or decreasing noise at high frequencies (like phono preamps). Most phono preamps are measured A-weighted, so we follow this standard. If we didn’t use A weighting, our numbers are about 6dB worse—still at the top of their class.
• Inputs shorted. This means the phono preamp isn’t hooked up to anything. This is 100% a cheat and 100% bull. Our numbers are almost 4dB better if inputs are shorted. We don’t measure this way, because this is not an actual use of the phono preamp.
• Referenced to 1V RMS. Most phono preamps reference to 1V RMS. We do 2V because we have the headroom, and because it is the standard line level output. If we referenced to 1V RMS, our numbers would be 6dB worse.
• MM gain. On most phono preamps, this is “low” gain. We have 4 gains. Their “MM” gain is usually a little lower gain than our “MM” gain, by 2-4dB.

“Enough of the measurements, what does it sound like!” someone cries.

And yeah, I hear you. No pun intended. But the best way to answer that question is probably to answer it for yourself. It’s not like Mani 2 costs a mint.

We really hope you enjoy it!

 

[quasor]

Tyrs

 

[ArmchairPhilosopher]

Ah, come on, you’ve heard the saying: “Ya cain’t polish a turd!”

So, Mani's internal code name is Dorodango, then? ;p

 

[Ripper2860]

Mani 2 seems very nice!! I wonder when they'll be done polishing the Urd?

 

[corndog71]

New and improved Schiit?!

Count me in!

 

[dstrimbu]

Tyrs

I was so excited... still excited... still waiting tho...

Mani 2 looks beautiful; congrats on another potential winner!

 

[US Blues]

Mani 2 seems very nice!! I wonder when they'll be done polishing the Urd?

I'll get the pitchforks, you grab the torches.

 

[audioelements]

just ordered a Mani 2. will report back

 

[LarryMagoo]

Vinyl is such a waste of time...

 

[bcowen]

Vinyl is such a waste of time for me.

FTFY.

 

[Stevko]

An asgard and a modius in one box would be nice.
399$?

 

[433521]

Vinyl is such a waste of time...

The best way to waste time listening to music is to interact with it. Personally I find it reminds me of my yut. B-52's sound better when the reminder to PLAY LOUD is staring up at you!

 

[ArmchairPhilosopher]

An asgard and a modius in one box would be nice.
399$?

An Asgard and Modius in one box doesn't really make sense. Modius' whole point is that it's fully balanced, but Asgard is single-ended.

What would make sense would be an Asgard with a built-in single-ended DAC.

…which already exists: Just configure an Asgard with an ES card and you're set. And it's 50 bucks below your proposed 399$, too!