The never ending amp journey

[cpp]

GranteedEV

I don't trust denon's pre-outs to be all that great

mine on my Denon 4311 work as good as the ones on my old Marantz AV7005 and have since removed the Marantz and my Sunfire Cinema Grand Signature amp from my room.

 

[DS-21]

The link with your concerns didn't work for me, but I have significant concerns with this test. Measuring a loud continuous tone tells you something, perhaps, about the average level you might listen to, but nothing of the peaks.

You response indicates to me that you misunderstand the procedure of the test, as well as what it reports. To sum up, here's what you do, and what you get:

1) Set the volume level using music you know to be dynamic, played at your maximum volume. (IMO, with a modern system using a calibrated volume knob, one can skip this step if one knows the maximum volume s/he actually uses.)

2) Download the track. For a modern system, the 220Hz track seems more appropriate to measure mains level.

3) Put in earplugs. (At least, I did. I never listen to test tones bareback, as a safety precaution. I always put in my trusty Etymotic ER-20's. If I'm going to go deaf from this hobby, I want the damage to occur on music I enjoy!)

4) Play the downloaded track.

5) Measure the voltage at the loudspeaker terminals.

6) Multiply the measured voltage by 4 for your peak voltage requirements, as the track is, by design -12dBFS. That 4*measured value is the absolute peak voltage you will ever use, with those loudspeakers in that room, because it will correspond to a 0dBFS digital signal.

It's a very elegant experimental design, using the knowledge that a digital audio source has an absolute level ceiling, 0dBFS to our benefit. My sole comment of the test was in response to an exchange between the test creator and loudspeaker savant Tom Danley of Danley Sound Labs. I could paraphrase, but I'm lazy so I'll just quote:

I think you and Mr. Danley are talking past each other a bit.

Here's what I think your test actually measures: how much power does it take to get your speakers in your room either to the edge of their linear performance to a subjectively-pleasing level of overdrive, on program material you actually listen to.

One thing I've noticed over time is that speakers with linear frequency response and enough efficiency/cone area make it very hard to gauge how loud they're playing without instrumentation. Often, the only way to tell how loud it is without measurements to try to talk to someone. With smaller speakers, one "hears" that it's too loud because of distortion creeping up.

That's, I think, Tom's real objection. But I think also think the data you're collecting are novel and useful,***

Lastly, if you can believe what they tell you in the manual, then perhaps Denon AVRs have equal, or in some cases slightly higher output than Marantz's, now that's just food for thought so don't bother arguing with me.

I would be surprised if the preamp boards were in fact any different. If they are, then D&M Holdings is a very inefficient company.

***I didn't order the AT3000 over the AT2000 because I thought I needed 300w/ch, it's just become my mantra to go-big-or-stay-home. I apparently got lucky.

Do you have that monster on a 20 amp circuit?

 

[PENG]

What's the time resolution of the meter?

Actually I use the Hioki mostly for current. I use the Fluke for voltage. It's response time is 100ms for min/max and 250us for peak measurements.

 

[PENG]

I would be surprised if the preamp boards were in fact any different. If they are, then D&M Holdings is a very inefficient company.

Hard to believe there is finally something we may possibly be able to agree on. Funny it is on something that we are just guessing.

 

[Irvrobinson]

You response indicates to me that you misunderstand the procedure of the test, as well as what it reports. To sum up, here's what you do, and what you get:

1) Set the volume level using music you know to be dynamic, played at your maximum volume. (IMO, with a modern system using a calibrated volume knob, one can skip this step if one knows the maximum volume s/he actually uses.)

2) Download the track. For a modern system, the 220Hz track seems more appropriate to measure mains level.

3) Put in earplugs. (At least, I did. I never listen to test tones bareback, as a safety precaution. I always put in my trusty Etymotic ER-20's. If I'm going to go deaf from this hobby, I want the damage to occur on music I enjoy!)

4) Play the downloaded track.

5) Measure the voltage at the loudspeaker terminals.

6) Multiply the measured voltage by 4 for your peak voltage requirements, as the track is, by design -12dBFS. That 4*measured value is the absolute peak voltage you will ever use, with those loudspeakers in that room, because it will correspond to a 0dBFS digital signal.

It's a very elegant experimental design, using the knowledge that a digital audio source has an absolute level ceiling, 0dBFS to our benefit. My sole comment of the test was in response to an exchange between the test creator and loudspeaker savant Tom Danley of Danley Sound Labs. I could paraphrase, but I'm lazy so I'll just quote:

Do you have that monster on a 20 amp circuit?

To answer your last question first, yes, I have a dedicated 20amp circuit for the amp, wired with 12/3. I'm guessing that for my use a 15amp circuit would work fine.

You are correct, I was multi-tasking when I read the test instructions, and didn't notice the x4 factor to allow for peaks. My apologies.

I'm still not bought in to the use of a single test tone, but I understand that without a test tone a simple volt meter won't have the resolution to tell you anything. I suppose this test tells you *something*, but my sense is that it probably underestimates peak power requirements for real music. I'm not sure it's really any more accurate than the estimation method I used.

I was browsing for scopes that night. They're cheaper than I thought.

 

[PENG]

I didn't think they'd be relevant at such a short distance.

I bet they are to some extent, but you can take measurements to find out just to be sure. Your prepro's test tone may not be the best for this but should still work. I did a quick check on mine and it was relevant (about 2 dB) at such a short distance in my environment.

 

[DS-21]

To answer your last question first, yes, I have a dedicated 20amp circuit for the amp, wired with 12/3. I'm guessing that for my use a 15amp circuit would work fine.

In that case, it's more likely that more power could've actually led to something in your system, given that you use very inefficient speakers (10dB/W/m less than my reference speakers), probably have a huge room (given that you can fit such a behemoth amp in it, and you actually have the current capability from the wall to make use of the amp's capabilities (20A circuit.)

I'm still not bought in to the use of a single test tone *** I suppose this test tells you *something*, but my sense is that it probably underestimates peak power requirements for real music.

Why? If anything, it could overstate actual peak power requirements, if one doesn't listen to music with peaks at the measured frequency. The upper bass is where voltage peaks are most likely to be, however. So "Pano" really did an excellent job with the test design. Comparing it to in-room measurements with an SPL meter (calibrated how? How recently?) is frankly laughable.

PS: How's this for irony: I bought a separate amp last night! It's not even a particularly modern one! It doesn't even have a switch-mode power supply, let alone Class D output stage. My new amp is a bog standard insanely heavy class AB unit with seven "blade" style monoblocs fed by a common PS with dual output transformers, Sherwood's A-965 I spent about 20min looking around for them after finding that T$S review by Dr. David Rich I remembered reading before, to provide expert support for my points about safety certification and the maximum power an AB amp can draw from a 15A 120V circuit...and it turns out there was an NIB one available locally for sale at a price I couldn't resist.

(Note also that the RCA inputs and binding posts are properly color-coded per CEA-865B, which is a nice touch.)

It lacks level/attenuation/gain controls for each channel, sadly, but for my purposes all that really matters is that the gain structure is the same across six channels. It also doesn't have balanced inputs (not a concern for me) or the Speakon outputs I'd prefer. But it'll do. I tested each channel with a very efficient tweeter (Eminence APT compression driver on a small square horn; not a high-fidelity device by any means, but for $30 and 100+ dB/W/m sensitive it is a great tool to weed out noisy amps) and couldn't hear any noise beyond ~3" from the horn mouth. Low noise is way more important to me than high power. And I'm actually glad the case metal (faceplate excluded) is on the thin side, because I wouldn't want to have it weigh another 10+ lbs!

Why would I bother with a separate amp, given what I've written on this thread? I'm going to remove the passive crossovers on my big Tannoys and drive them actively. (Processor as-yet undecided. I might go with a pro matrix-routing processor, or the miniDSP 8x8. The seventh output will likely end up driving my high-mounted subwoofer.) Since I'll likely be mounting my mains up high, removing their very heavy (all air-core inductors, poly caps, and beefy resistors) crossovers is a virtue, as is the finer control over their sound that can be achieved with selectable crossover slopes/frequencies and parametric EQ. So I needed a six-channel external amp anyway. This giant boat anchor is, of course, total effing overkill for mains that even with the passive xo were 96dB/W/m.

 

[AcuDefTechGuy]

Why would I bother with a separate amp, given what I've written on this thread? I'm going to remove the passive crossovers on my big Tannoys and drive them actively.

I predict that you won't hear a significant improvement.

Perhaps some improvement, but not significant.

 

[Irvrobinson]

Why? If anything, it could overstate actual peak power requirements, if one doesn't listen to music with peaks at the measured frequency. The upper bass is where voltage peaks are most likely to be, however. So "Pano" really did an excellent job with the test design. Comparing it to in-room measurements with an SPL meter (calibrated how? How recently?) is frankly laughable.

I'm using a Dayton Audio OmniMic set-up:

Dayton Audio OmniMic Precision Measurement System 390-790

The spectrum analyzer screen, with the microphone positioned where my right ear would be in my listening seat, has fascinated me for some time now. One of the real-time monitoring functions is average and peak loudness.

It was also a great tool for placing and equalizing my sub. Velodyne's software is quite unimpressive.

 

[DS-21]

I predict that you won't hear a significant improvement.

Perhaps some improvement, but not significant.

I expect you're entirely correct in that. I may even find that I can't do an active crossover that improves at all on the passive one, or even that I can't properly effect a clone of the passive with the tools I have available. (I tried to design a passive that was better, and failed. As did a well-known crossover guru I commissioned to try. No slight to that person. Tannoy's engineers knew what they were doing, and when you're selling a speaker with an inexpensive finish for like $3-4k a piece the engineers have some room to play!)

The main reason I'm doing it is to make the mains lighter. The crossovers are very heavy. The biggest inductor in them alone is about 2lbs! (It is not even on the board, but remotely located and screwed down separately to the cabinet.) If I want to put them on shelves cut to fit, or use wall-mounting brackets, lighter is useful.

My Nathan Funk-built cabinets are already designed to be as light as possible for the size, with the single biggest panel (the side/back monocoque) being a thin CLD panel made (from outside-in) of plastic laminate, glue, sonotube, Liquid Nails, and damping panels placed strategically inside based on panel vibration measurements once I received them from Nathan. Bracing is also just the minimum necessary, with a curved shelf brace that supports the woofer magnet of the Dual Concentric driver (coupled via poster-tak, just as on the originals) and a dowel between the top and bottom end caps.

If I end up getting a front projector in my new place, the amp could also power three of Keele's new CBT kits from Parts Express for front mains...

 

[AcuDefTechGuy]

If I end up getting a front projector in my new place, the amp could also power three of Keele's new CBT kits from Parts Express for front mains...

That will be sweet.

Since the CBTs are 61" tall, I assume you will use those sound transparent screens?

I don't know why they don't offer the crossover + speakon cables along with the kit for additional $$.

Would you order the Audio Artistry's XO?

 

[DS-21]

That will be sweet.

To be clear, I probably won't do that. The Tannoy's sound awfully good. I've never felt the need to upgrade them.

But it's not entirely impossible, either. Especially if we end up moving up to NYC in a few years, and I decide to do a dedicated theater above the garage of our Lake George house. IF I were to buy new speakers, they would either be CBT's or Danley Synergy Horns. (No GedLee Abbeys, or JBL LSR6332's, because I value symmetrical vertical polar response. Sometimes I like to stand up and pace when enjoying music at home. That's why I'm such a big fan of Tannoy, KEF, Gradient, and TAD.)

So why buy the amp now? Simply because the chances of lucking into something of that quality for that price is unlikely to happen twice without expending way too much effort searching the second time around.

Since the CBTs are 61" tall, I assume you will use those sound transparent screens?

I'd do a Geddes-style screen: thin round posts holding a white sheet backed by black muslin in tension. The "sound transparent" screens I've heard have all been awful-sounding.

Would you order the Audio Artistry's XO?

They're going to have a bespoke unit? I thought they were using a standard Berry DCX.

If they do offer a nice-looking one, my choice would depend on how much space I have to hide the electronics. I am not willing to look at the blinky lights of a Berry DCX.

Still, I'd probably use the 8x8 miniDSP. I like their interface, and it's OSX-compatible. I'll happily pay more to support companies that make Mac-compatible products.

 

[AcuDefTechGuy]

They're going to have a bespoke unit? I thought they were using a standard Berry DCX.

If they do offer a nice-looking one, my choice would depend on how much space I have to hide the electronics. I am not willing to look at the blinky lights of a Berry DCX.

Still, I'd probably use the 8x8 miniDSP. I like their interface, and it's OSX-compatible. I'll happily pay more to support companies that make Mac-compatible products.

Don Keele replied to my email saying the completed CBT speakers will come with "the crossover", without amps & cables. All for $8500.

 

[GranteedEV]

I'd do a Geddes-style screen: thin round posts holding a white sheet backed by black muslin in tension.

I would use the seymourAV AT screen material

They're going to have a bespoke unit? I thought they were using a standard Berry DCX.

At RMAF they used a DEQX, which might account for a lot of the markup over the kit. Dunno, though.

 

[DS-21]

FWIW, if someone's dead-set on a standalone amp for whatever reason, and doesn't mind a humungous Class AB boat anchor, there are two superb super-cheap options on A-gon right now: a Sherwood A-965 and its clone Boston Acoustics A7200. While I didn't click through to see condition, they're both "maybe I shoulda waited a couple days" cheap.
(Though new from an authorized dealer has its perks, too.)

While I couldn't find any online measurements for the amp, here are Home Theater's measurements for a Sherwood AVR with similar amps, which Sherwood rated as doing 120W/8Ω/2ch, with no "all channels driven" rating (The A-965 is rated 100x7/8Ω, 160x7/4Ω, all channels driven, at ≤0.007% THD) .

Five channels driven continuously into 8-ohm loads:
0.1% distortion at 128.3 watts
1% distortion at 155.7 watts

All channels driven continuously into 8-ohm loads:
0.1% distortion at 109.6 watts
1% distortion at 126.0 watts

***[with] two channels driving 8-ohm loads, [the R-965] reaches 0.1 percent distortion at 134.0 watts and 1 percent distortion at 167.1 watts. Into 4 ohms, the amplifier reaches 0.1 percent distortion at 229.7 watts and 1 percent distortion at 261.8 watts.

And look at the decimal place on the distortion graph, while noting that it's not just the amp, but the amp+front end.

According to Dr. Rich's T$S review, the differences between the R-965 AVR and A-965 amp are that the AVR's amp has single transformer rather than duals, 2/3 the primary filter capactitance, half the power transistors, and "less room for heat sinks since all the front-end electronics fills half the box." The AVR is also listed at 52lbs, compared to 80lbs for the separate amp. So if nothing else, the separate amp should perform at least as well as the AVR.

IMO, you won't get a better class AB 7-channel amp period, unless you have a 20A+ available to power the thing. (I would prefer a a good Class D amp of similar or slightly greater power, because it would be more energy-efficient and smaller. All current Class D implementations do measure worse than this amp in terms of THD, though not audibly so.) And you certainly won't get anything else of close quality for near the price. Amps with similarly superior performance carry names like "Bryston," "McIntosh," and "Anthem Statement."

Here's how Dr. Rich - who, not being a mouth breather, did not "listen" to the amp and write idiotic purple prose about its "liquid midrange" or whatever, but rather applied his expertise to analyze the circuit design and parts/build quality - finished his T$S review:

The A-965 passes FTC preconditioning with 160 watts into 4 ohms, all channels driven, and still gets the UL safety stamp of approval.

Combine that with a power amp designed for low distortion and high bandwidth before the feedback loop is closed, and the price of $1,500 is a real bargain. *** If you want about the most horsepower that you can get out of one AC outlet in a 7-channel AV configuration, this 80-pound monster should fit the bill and remain state-of-the-art for many years to come.--DAR

Besides its size and mass, the only criticism I can lodge on the A-965 in good faith is that the binding posts, while higher quality than one would find on an AVR and helpfully color-coded per CEA-865B, are a bit short for bananas, or at least the sawtoothed ones I use. The banana sticks out a little. I might switch to spades for pet safety reasons.

In contrast, also I checked out a Stereophile review of that $1500 (used!) Theta stingray-vagina thing mentioned in the first post of the thread. Wow, is it ever a steaming pile of mid-fi crap! First, their shows all of the attention to quality detail one expects from the dead-oops-I-mean-high end, with a quarter-decible channel imbalance. Also, the design itself seems pretty substandard, with an input impedance low enough to cause 1dB swings into their simulated load. Lastly, Theta's claimed power ratings seem to be is about as accurate as an 1980s ghetto blaster's: rather than being a 100W/8Ω, 200W/4Ω amp, into a standard 1% THD it's actually a ~40W/8Ω, 75W/4Ω amp. And the clipping "knee" (point at which distortion starts to rise dramatically) is much lower than that, about 7W (!) into 8 and 4Ω, and between 2 and 3 W into 2Ω! See here:

But on the bright side, such a poorly engineered device built with such slapdash quality control will sound different from a well-designed and well-built amp...I guess

Still, I can't see why someone would actively consider shelling out $1500 for someone's decade-old midfi-at-best castoff. $1500 will buy a lot more clean power from a brand new amp.

 

[Irvrobinson]

the design itself seems pretty substandard, with an input impedance low enough to cause 1dB swings into their simulated load.

You mean output impedance, and that it is high versus low. I'm a fan of very low amplifier output impedance myself, but there's a lot of controversy over how low it needs to be. My view is the closer to zero the better.

I agree with most of your intent though, which is why I wish there were more people doing comprehensive bench tests on amplifiers. Channel matching issues that vary by frequency aren't very good for precise imaging. SNR can matter. Low distortion isn't a given; it can be achieved routinely, but as the Theta proves it is possible to screw even that up in a solid state amp.

I used to like the old Audio Critic reviews with the circuit analyses. A fancy case doesn't mean there's a well-designed circuit with high quality parts in it.

I popped the cover on my new ATI AT3005 and I'm pretty impressed. Nowhere near as over-engineered as my old Levinson amps, but I think Kessler made some very thoughtful choices. I know the ATI is still quite expensive by the standards of this forum, but I actually think it's a pretty fair value in terms of parts cost to selling price ratio. (Selling prices are substantially less than MSRP, if you shop.)

 

[GranteedEV]

Low distortion isn't a given; it can be achieved routinely, but as the Theta proves it is possible to screw even that up in a solid state amp.

The problem is, audiophiles are resistant to the idea of feedback circuitry in their amps.

 

[Irvrobinson]

The problem is, audiophiles are resistant to the idea of feedback circuitry in their amps.

Not this one...

 

[GranteedEV]

Not this one...

That's why your Salon2s are so sibilant

 

[DS-21]

You mean output impedance, and that it is high versus low.

I think you're right. The line from the review is "the source impedance was on the high side for a solid-state design, at 0.65 ohm."

That's like the "triode sound" tap on a Sunfire amp.

I used to like the old Audio Critic reviews with the circuit analyses. A fancy case doesn't mean there's a well-designed circuit with high quality parts in it.

Ditto. As an aside, the gentleman who did those great circuit analyses for TAC is the same Dr. David Rich whose T$S review of the Sherwood amp I quoted above. That amp wasn't on my radar screen until this thread, because I thought I remembered reading something about NRTL labs and how most amp makers/marketers are too cheap to get the proper safety certifications. So I went back and found that review. Then I bought the amp he reviewed.

I popped the cover on my new ATI AT3005 and I'm pretty impressed. Nowhere near as over-engineered as my old Levinson amps, but I think Kessler made some very thoughtful choices.

ATI makes a quality amp. FWIW, here's TAC's review of the 12-channel model.

I know the ATI is still quite expensive by the standards of this forum, but I actually think it's a pretty fair value in terms of parts cost to selling price ratio. (Selling prices are substantially less than MSRP, if you shop.)

IMO, at a bit over $3k MSRP, and I'd imagine a lot less for a good negotiator, I agree that it's a good value, if one considers the additional cost of a dedicated 20A line as well, for the power it offers. To look at it another way, it's much cheaper than a Bryston, McIntosh, or Anthem Statement amp of equivalent power. (I mention Anthem Statement in the same breath only because of these rather stunning test results. All that power in an AB amp is pointless for people who (unlike you) don't have 20A circuits to feed it, though.