Halcro may have reached the level of where the test instruments are NOT the equal of the gear being measured. I do not know this to be true, but it might be a good guess. Also marketing folks have been known to push the envelope once in a while. I am not acusing Halcro of doing this as I would doubt there is any thing lacking in the Halcro measurable or otherwise. It is as usual in a class by itself. It is amazing Mr. Candy continues to even TRY to improve on his already spectacular products. He is so far a head of speaker design science and manufacturing, arguably the hardest of all the audio manufacturing disciplines, You wonder why he does it accept that he CAN.

Isn't the rub of high end is that there is always SOMETHING better than what you have now. If Halcro owners CAN'T sleep knowing they might not have reached the top of the mountain, how can any of the rest of us?

Ditto Jim's comment. I find nothing implausible in the hypothesis that many of today's super-low distortion amplifiers are better than the equipment used to measure them. I believe JA himself has questioned the ability of his measuring machinery to keep up with the latest technical achievements in amplifier and preamplifier design. The key, as you note, is how well the piece performs in real-world listening situations. And the latest Halcro apparently offers some slight sonic improvements over its predecessors. The point about getting one's 40 large worth seems inane, to me. As good old Ben Franklin said, experience runs a tough school, and the stupid will learn only after entrance. I am sure Mr. Candy will arrange for some kind of an exchange, if you bought one and are disappointed in the abstract specs...but don't ask for your money back -- school is now in session, and tuition is unrefundable. I'll get by with my used VTL, which sounded fine in my other system before I sold the Dynaudios. I bought gold and silver with the other 30 grand, a couple years ago, which has about doubled over the interim. I hope nobody bought this thing with a home equity loan. That would be graduate school.

This isn't going to be a replay of the Chord controversy, is it?

Quote:
This isn't going to be a replay of the Chord controversy, is it?

Clifton - You know it's always possible to strike a Chord at these forums.

Groan. Mr. Candy won't strike a bargain with anybody on this thread.

I think there is a lot better chance of correlating measurements of amplifiers with audible performance than is the case with speakers. Still, I'll just listen. I wonder if you could hook the Halcro up to some speakers, print the conventional response characteristics on a few graphs, do the same with my VTL (same speakers), and "see" any differences. I bet you could hear 'em!

Quote:
if you look at JA's 19+20 kHz IM spectra for the Boulder 2050 and the Bryston 14B-SST, they are not that much different from those for the three Halcros reviewed, even though one would not expect those amplifiers to turn in IM performance quite as good as the Halcros.

I am beginning to believe that what is being shown in the IM graphs is the residual of the twin-tone source.

With the Halcro's harmonic distortion, Bruce Candy suggested that this could be due to a hitherto unsuspected non-linearity at very high power in my test load that had been unmasked by the dm88's very low distortion level. I have constructed another load, without switches or relays, to investigate this using the dm38 that I am following up on in the October issue. (The dm88 samples, unfortunately, had to be returned to the distributor prior to HE2006.)

To be honest, the very small shortfall in the dm88's measured performance vis-a-vis its specifications is academic.

John Atkinson
Editor, Stereophile

Delighted to have posts from an Amp designer of the stature of Bob Cordell.

Also a pleasure to read his views.

I concure fully with Cordell's comment :

Quote:

That does not make the amplifier immune to discussion of its technical aspects, however. Some of us like to discuss what's under the hood.

I suspect that the vanishingly low levels of distortion on the Halcros is achieved by Distortion cancellation, from feedforward, rather than only feedback & ofcourse Very Careful Ckt design and layout.

Given the vanishingly low levels of distortion that the Halcros probably have even before Feedforward correction has a go, it seems that the feedforward really needs to be individually tweeked for each unit, to actually achieve the specs indicated.

I have little doubt that the ppb distortion levels CAN be met, but are not met, because in production, that individual level of attention is not given.... and would also be impractical to achieve, and probably also mess up the pcb with repeated component insertions and removals.

Mr Cordell, could you care to speculate on this ? ( Whether Feedforward needs to be individually tweaked on each unit, to achieve optimum results )

The obvious extrapolation of the above is that the new amp that JA receives WILL meet published specs, or bottom out the measurement capabilities of JA's equipment for not only the IM but other tests too.

Quote:
I suspect that the vanishingly low levels of distortion on the Halcros is achieved by Distortion cancellation, from feedforward, rather than only feedback & ofcourse Very Careful Ckt design and layout.

I had thoughts along these lines when I read about the first Halcro amps reviewed by Stereophile. It does make one wonder.

www.avahifi.com Dave Hafler was the FIRST to use MOSFETS in an audio amp back in teh 70's I think...Vanalstine continues teh Hafler tradition of top quality, high performance priced for mortals. An OmegaStar EX AVA in teh Hafler chasis I use a few of em, the P500 OmegaStar EX ckts are doing about 1200-1400W MONO into my speakers...speced at .01% or less distortion 800v/uS MOSFETS baby Exicon from England in the ckts. Faster more linear according to AVA. 12 in the DH500 12 in the P500 with a bigger mains transformer 16 in the XL-600 going down into 2 OHMS with stability and cranking them watts WATTS there is no substitute.Halcro costs how much? Chord costs how much? How much do they cost when meeting specs? Are they the hand selected ones that meet specs? McIntosh guarantees it meets specs don't they? When you pay alot you expect alot. When you pay less and get more....now that's called great design. John, maybe another measuring session, do some measurments of the AVA amps lets see how it meets specs?

From Hawksford and error-correction to MOSFETS to Hafler...to another Van Alstine commercial. DUP, have you no shame?

Hi, Bob --

Actually, your posts are very interesting. An actual designer talking design theory (unlike a mindless statistician raving the superiority of abstract theory over the reality of listening) is quite fascinating. There were attempts, as I recall, by various reviewers a couple of decades ago (Julian Hirsch?) to measure amplifier performance through speaker taps, but nothing came of them. In today's oscilloscope-centric universe, it would be educational and fun to see those attempts revisited...especially since amplifier design has progressed sonically as well as theoretically. Amplifier/speaker interface is, after all, a trial and error, subjective inquiry as things now stand. And if the objective side of things has true real-world validity, one ought to be able to measure amplifier performance through speakers playing music. I suspect the results would be contaminated, initially, by the sheer variety and number of variables presented to the technician, but that has never stopped science before. I admit I have little patience with those "look" for music in a laboratory dominated by visuals, be they LED's, graphs, or blips, and draw sonic value judgments out of a non-aural matrix. But the design perspective is a different animal, as abstract theory must be documented with measuring instruments. Again, I enjoy your comments a great deal. The 40 grand? Discretionary spending is up to the individual. I can afford this for an amp, but shall not pay it, because there are other demands on my discretionary cash. Cheers, Clifton

Quote:
Halcro may have reached the level of where the test instruments are NOT the equal of the gear being measured.

I always wonder when I hear something like that.

Mr. Tavegia, if what you postulate is true, then how would Halcro know the specs?

I'm a firm believer in "hearing is believing" but I am also opposed to bullshitting with specs.

A product should perform "as specified." Otherwise, "specify" it differently!

I don't have my issue handy, did Halcro address this in their Manufacturer's Response?

HAFLER had on their XL-600 model I believe an internal trim pot for user adj distortion tweaking, this was back in the early 80's...HAFLER was probably as usual ahead of his time. He had an entire procedure et up in the manual for using one channel against the other as a reference. When ever I read, that the device under test is beyond the limits of teh test equipment, it means they have the WRONG test equipm,ent. You gonna try to convince me Fluke, Agilent, Tektronix don't have test equipment better than any audio product by any mfg? Surely you are deluded...how then as mentioned does teh audio maker know of what he makes? What is the slew rate I didn't cath it on the write up on the Halcro in this month's issue..should be something astronomical for this price....better be above 800V/uS damping factor should also be better than anything ever made As these amps get into the stratosphere in prices, it seems just because they can.....How do they get to these list prices anyway? That new McIntosh multi pice absurdity....looks great, but also seems like a bunch of nonsense to have to make so many different modules....I think it's just to impress someone, no logic to the design...Halcro looks like a monster, for what reason? I wanna see a water cooled amp with 10's of thousands of watts, running off 240V twist locks!!! For $40K..CROWN had them needed 240V supplys they where so potent!!! Ya can't get more than what teh line delivers either

Yo, Buddha --

Bruce Candy may have test procedures/equipment more advanced than other lab facilities. I don't know. I guess you have to make some "superlative" claim to justify the price. I would (as you suggest) be more comfortable if Mr. Candy had said, "...and you will hear less distortion in this revolutionary design than otherwise thought possible..." or some such. Why resort to the numbers as "proof," when the proof always has been in the listening?

Do you remember some of those 2nd generation transistor amps in the '80's claiming they had whipped "notch" distortion? The oscilloscope traces looked mighty impressive, especially compared with older models that typically topped out around 1% IM and 1.5% THD. But they still sounded like buzz-saws.

Hafler amps always sounded veiled to me, compared with the competition -- they lacked clarity and transparency, even though they sounded harmonically "right." DUP, are you sure you took your earmuffs off from last winter?? They put out a lot of veiled watts, though, that's for sure. Cheers, Clifton.

That reminds me. Did you see that old W.C. Fields movie, where he is in a cabin in the frozen wastes of the Yukon? There is a howling blizzard outside, and he and his babe are trapped in some leaky cabin. He takes a zither off the wall, where it had been hanging on the end of a tenpenny nail, and begins to strum. He looks soulfully at the babe (Mae West??), saying softly, "Do you mind if I play with my mittens on?"

The Hafler -- mosfet mist? There is a reason why everyone who frequently heard live music preferred tubes, in those days, in spite of the relatively bad measured performance. Nowadays, the gap between tube and transistors has certainly been closed; even though there are still some slight differences you can sometimes hear, it is difficult to say one is better than the other, provided both types are top-notch designs. But the measured differences still remain -- whenever JA sums up a tube amp that he likes, he seems to always write, "these are excellent numbers, for a tube amp," or, "this product seems well-designed, for a tube amp." Still, most reviewers couldn't tell a VTL from a Krell in terms of the quality of the listening experience, although even I can say, "they sound a little different."

Quote:

Do you remember some of those 2nd generation transistor amps in the '80's claiming they had whipped "notch" distortion? The oscilloscope traces looked mighty impressive, especially compared with older models that typically topped out around 1% IM and 1.5% THD. But they still sounded like buzz-saws.

Exactly!

So, why specify that way?

Why take relatively low priority "numbers" and "over-specify?"

The only reason to care about Halcro's over-specification is to wonder why they'd feel the need.
____________________________

I do figure that maybe they are that low and it's a question of measurement technique...

Veiled? That's why Frank Vanalstine has upgraded from the original Hitachi MOSFETS in the Hafler DH500,XL600, P500..to EXICON from England faster, more linear according to Frank....If you ain't heard the latest OmegaStar EX ckts, you are missing a lot, ain't no veils, this ain't Mecca, or Baghdad, we don't cover it up to look like a ghost, we let it free to fly and sound REAL, LIVE with SLAM, IMPACT and REALISM. How can Halcro justify $40,000 price tag on those things, AND they appear to be like CHORD, don't meet advertised specs? FTC loves those kinds of bargains....If you bought a car advertised at 286 HP it only have 145...you would be pissed, and moving mighty slow ...same with WATTS, you be listening mightly congested, flat, lifeless, where's the SLAM, Dynamics, realism...where's MY WATTS. MOSFET smooth FAST WATTS....Didn't teh Phase Linears use Bi-Polar gritty transistors? The DH500 came out and smoked anything out there SS wise in it's time...

Looks who back in town....cool, another inovative pioneer for lotsa good watts.. http://www.ampzilla2000.com/

I agree.

Quasi Complementary was Phoo-phooed for years, once matched pnp power transistors became available.

Now we have the darTZeel power amp.....

Exactly. The proof is always in the listening (after all, the reason for the amplifier's existence), in real listening rooms, through real speakers, and with real-world recordings of music played by real musicians. Most of the transistor amps of the '70's and early '80's sounded quite unrealistic, whether MOSFET designs or bipolar designs. I remember. I heard most of them. That's what drove me to tubes, even after consideration of the typical tube amp's own bag of sonic compromises. I suppose you could argue that tube design was a mature technology, building on 50 or more years of trial-and-error and innovative progression, and that transistor design was still in its infancy, with the errors dominating the various solid state design approaches. Currently, solid-state (and hybrid) design has certainly evolved to the point where many of these amplifiers are preferable to tubes. I had a Steve McCormack amplifier for about 5 years (the DNA-1), with Steve's own upgrades added in as they became available. It was a MOSFET design, as I recall, but had none of the flaws I remembered hearing in earlier MOSFET amps. It was superbly transparent and smooth. I suspect design will always triumph over materials.

You shouldn't cringe at the fact that labels inevitably accrue when the product doesn't perform. The bad reputation may be wrongly attached to the designer's materials, instead of the designer, but that's merely language at work. Today's designers have plenty to emulate on the positive side of the evolutionary curve, and many negative stereotypes are, indeed, labels that are time-specific, to the past, and no longer applicable in the present. I was speaking of a '70's amplifier (the Hafler) that was veiled, compared to the best tubed designs of its era. And, yes, "mosfet mist" was the tag that summed up this situation. Somewhat analagous to the choice between AR-3's and JBL's, on the speaker side of things, the choice between "mosfet mist" and "bipolar spit" was Hobson's choice. But, fortunately, things have improved.

If all it matters is listening why are there specs? Cus' specs matter, how else do you have any idea what the thing even starts to do? any product that is so overpriced and so under performing, would indicate a defect? How's that CHORD doing at less watts than ya paid for? and ya be paying an awful lot for nutin'.

At the point of the Halcro I think the specs only matter to the designer who is trying to go better than the DM 38. I am amazed at those who can hear the difference between the resolution and presentation of the 38 to the 88.. God love'em. At my age of 59 this month...that is not happening for me.

The question is what are you going to use to measure the improvment in the DM88 to the soon to come DM99!!!!!! Or what ever is coming next from Halcro. Until some speaker can do what the Halcro does I am not worried at all.

I like the cigarette lighter analogy: we all want everything to work as advertised, even if we will never use the lacking feature (or, in this case, hear the missing spec). And, since you are a designer, and thus fussy about everything you put your name on, you are going to be a bit hyper-critical about something designed by one of your peers. I like to think I can still hear the distortion that matters, and I don't really care about the number of zeroes behind the decimal, as long as I don't hear any grit. But I don't have the professional perspective that you have...and professional pride in the details is as much responsible for progress, in audio products, as gross and subtle improvements in the perceived sound. So, yes, Mr. Candy should clean this up, even though I don't really care about this particular product. Cheers, Clifton

Quote:

I am beginning to believe that what is being shown in the IM graphs is the residual of the twin-tone source.

I would believe this. I have always encountered problems when tried to generate ultra-low distortion differential tone test signal. The best results were obtained by summation of 2 single frequency sine signals on resistor network.

Is there any special reason that you usually use 50Hz sinus for THD spectral measurements? I would guess that 1kHz or even higher fundamental frequency (up to 10kHz) would show more of the spectral behaviour of the amplifier, especially the one as good as Halcro.

Quote:
Is there any special reason that you usually use 50Hz sinus for THD spectral measurements? I would guess that 1kHz or even higher fundamental frequency (up to 10kHz) would show more of the spectral behaviour of the amplifier, especially the one as good as Halcro.

I do perform FFT analysis of higher-frequency tones, but don't publish them unless they are different from the LF tests. I use 50Hz as it is close to the 60Hz AC supply frequency and at high powers will reveal power supply deficiencies, an idea first proposed by Martin Colloms in the early 1980s.

John Atkinson
Editor, Stereophile

Quote:

I use 50Hz as it is close to the 60Hz AC supply frequency and at high powers will reveal power supply deficiencies, an idea first proposed by Martin Colloms in the early 1980s.

John Atkinson
Editor, Stereophile

It may probably mask the power supply residual spectral lines due to finite PSRR (multiples of 50Hz or 60Hz), which can be clearly seen in case that we use 1kHz tone.

Quote:

Quote:

I use 50Hz as it is close to the 60Hz AC supply frequency and at high powers will reveal power supply deficiencies, an idea first proposed by Martin Colloms in the early 1980s.

It may probably mask the power supply residual spectral lines due to finite PSRR (multiples of 50Hz or 60Hz), which can be clearly seen in case that we use 1kHz tone.

Maybe, but you can can clearly see spectral lines developing at 70Hz and 190Hz with some amplifiers (120Hz