Compensation capacitor Cdom

[awpagan]

Jacco
A car saleman with a sense of humor
Would you really buy a car from him??

But the show is amusing.
Wouldn't you like to do the brand car what they do.......
and not pay for it..

Have you heard the Halcro DM58/68. I haven't
what's your impression of it or the topology?

I remember the only Krell i was really impressed with was the KSA50.
After that, yeh another Krell, lost it!
But that's me, my hearing..

allan

 

[David Lewis]

Good point indeed. How many supposed audiophiles are there that have heard a Halcro, or better still own one. A silly question I suppose on a DIY site, but in the big scheme of things quite valid.

If nothing else Bruce Candy has fired a big shot for the pro low THD camp. See below some of the mag review summaries. I can't find a bad one yet but I'll keep googling.


The Absolute Sound, Issue 129, April / May 2001

What’s Hot by Robert E. Greene

"with the ultra-low distortion Halcro amplifier, something of unusual purity emerged."


The Stereo Times
My First Trip to the CES by Henry Wilkenson

"Simply put, these amps are among the best that I have ever heard. They are extremely transparent, yet never hard or overly detailed. They display a relaxed and effortless quality, presenting the music in a way that made you forget about the system. These mono amps are something that you will be hearing a great deal about."

Stereophile, May 2001
Larry Greenhill on Accessories & Surround Sound

"they sounded so pure and beguiling as to almost defy description. (Upstairs) "

Halcro Review Extracts
Lead lines:
"Not Cheap but the very best never is"

"Leading the way to perfection"

"It was not only the best sound we have had in our listening room"

"NO ARTIFICIAL ADDITIVES (Used in TAS Ads)"

"The worlds only Super Fidelity amplifiers"

TAS - absolute sound review (Issue 71)
The Meaning of Zed

"My first impression was of vastly heightened dynamics, combined with extraordinary articulation of musical textures evident on recording after recording."

"The amp appeared to have limitless headroom.
Limpid & natural sounding with the Halcros – no hash, no glare, no glitter, nothing but air, delicacy and crystalline articulation."

"Miles own trumpet was palpably real. Reverberation, whether artificial or native to the recording space, was reproduced with startling precision. Black silences, the like I’ve never heard on an audio system. I don’t believe any consumer vacuum tube component holds an advantage over the Halcros. The more I listened, the more convinced I became that Halcro’s technical excellence represents the preferred approach to amplifier design. If you want the colorations, go to the other people. If you want diamond clarity and utter lack of blemishes or misbehaviour, go to Halcro. "

Stereo Review Harman Japan: 24 October 2001
THE BEST BUY - THE "STEREO" MAGAZINE (published by Ongaku No Tomo Sha)

Organiser: Ongaku No Tomo Sha
Panel of Judges: 14 Reviewers/Critics
Announcement of results: On the December2001 issue of the "Stereo" magazine



POWER AMPLIFIER - Over 1,000,000 Yen

Rank Brand Model Retail Price*

3 Total Received Points

1 Halcro dm68 (GM) ¥5,700,000 105.0

2 Accuphase P-1000 ¥1,100,000 103.0

3 Accuphase M-2000 ¥2,000,000 101.5

4 Mark Levinson No.436L ¥2,200,000 99.0

5 Mark Levinson No.335L ¥1,700,000 94.0

6 Mark Levinson No.33HL ¥3,700,000 93.5

7 Linn Klimax ¥2,400,000 80.5

8 Denon POA-S1 ¥4,000,000 78.5

McIntosh MC1201 ¥2,000,000 64.0

Mark Levinson No.33L ¥5,900,000 64.0

Mark Levinson No.336L ¥2,000,000 51.0

Niro PE-1000 ¥4,000,000 51.0

11 Mark Levinson No.334L ¥1,250,000 49.0

33 Mark Levinson No.434L ¥1,500,000 7.5

HALCRO dm68 Super-Fidelity power amplifier review
by the leading Hong Kong hi-end audio publication Audiotechnique, Issue 237, 2001.

"it has solid and sweet sound performance along with a great level of musical involvement, you will not regard it as a solid-state amplifier; on the other hand, its stunning dynamic contrasts, incredible transient response, and the distinct outline of each stereo image will definitely not make you consider it a tube amp."
Because it offers an ultra-quiet, almost "Pitch Black" super-low noise background, you will find much more never-before-heard details from your favorite CD collections than any of your previous similar experience.
details of harmonics of each instrument became exceptionally abundant, the different levels and three-dimensional sensation of the soundstage were also unprecedented.

I think dm68 is truly one of a kind--its extraordinary sound performance combining all the advantages of both solid-state and tube amps.

I exaggerated nothing. It really has a superb sonic performance so different from the others. But, the tremendous sound effects such as the life-like 3D imaging, pin-point focus and precise object location remind you continuously that they all benefit from the existence of dm68 and it’s the best sound that we have had in our listening room for the last three years. Searching my heart, I found no flaw in dm68’s superlative performance by any standard
it produces only 1,000 parts per billion (0.0001%) harmonic distortion at 20kHz at full power!

Since dm68 has very low distortion in all stages, it’ unnecessary to make correction by applying feedback owing to the minimal accumulative errors. No compromise in component quality - it employs industrial grade components
The power supply of dm 68 features switch mode design
The switching frequency of dm68 is above 100kHz, in another word, far beyond the audible range. Ironically, all the test equipments have greater distortion than the amplifier itself. It’s truly a "real" sound. I have to say that I’ve never heard an audio component which can sound like this. It has sweet voice of tube amp, almost limitless frequency response, which is full of nuance and delicacy, hi-resolution reproduction similar to the magnifying ability of a microscope, and no matter what type of music or instrument it plays, the extending of ultra highs and attenuation of harmonics of the percussion instruments are both dealt in completely natural ways.

 

[jacco vermeulen]

the show is amusing

yeah, after all the fuss overhere about the Halcro amps i did.
The beauty of this cheese place is that there's an audio palazzo at a 2-4 hour drive in any direction.
(don't know a more fitting word for these mega luxury shops)
Visiting at least one audio store on every trip i make abroad has become a tradition after doing that routine for +2 decades. The brand was/is not imported overhere afaik, i had to visit one in Paris for the 58s.

I'm not trashing B.H.Candy, i welcome any techn. innovation and see the existence of these amps as serving a purpose. For what it's worth, the Halcro 58 is a piece of art. But if i had my eye and wallet on a Veyron, it better do 250mph or the riot salesman will be in knee deep doodoo.

(anyone visiting Paris : Adagio, 10 minute walk from the Gare du Nord trainstation)

 

[gaetan8888]

Hello

I think that the Halcro amp do take account of the memory distortion problems and correct it, but is that type of distortion are a so big problems ?

Here's a web page about it;

http://peufeu.free.fr/audio/memory/memory-10-annoyance.html

At least it seem that the input diff pair transistor can be the problem.

Gaetan

 

[David Lewis]

Hello

I think that the Halcro amp do take account of the memory distortion problems and correct it, but is that type of distortion are a so big problems ?

Here's a web page about it;

http://peufeu.free.fr/audio/memory/memory-10-annoyance.html

At least it seem that the input diff pair transistor can be the problem.

Gaetan

The consensus seems to be that Halcro acheives low distortion by applying big amounts of global as well as nested fb. However, I haven't personally seen a schematic.

- I'm not sold on the idea of cascoded current sources just yet. Particularly in the VAS where you would expect the Cdom to totally swamp any non linear VAS loading from a current source.
A good reason for cascoding the VAS is to load the LTP with an EF stage with AV=1, which negates the impact of the very non linear Miller capacitance.

 

[jacco vermeulen]

As this turned to Halcro stuff, a video of the Logic series: here

In French, but it is nice to watch.
During the interview it's stated that the software of these models can be updated from the Halcro factory through the web.
(MS goes Oz)

 

[awpagan]

The consensus seems to be that Halcro acheives low distortion by applying big amounts of global as well as nested fb. However, I haven't personally seen a schematic.

Wouldn't large amounts of nfb slow the amp??



gaetan8888
"memory distortion problems"

Thermal drift ???
using the component within it's "sweet" spot. ?

allan

 

[David Lewis]

Wouldn't large amounts of nfb slow the amp??

According to Bruce Candy the output stage has a 1MHz bandwidth with feedback applied around it. I Gather that it uses Mosfets but I'm not sure in what configuration. Large amounts of feedback doesn't make the ampifier slower, but it's fair to say that it usually requires limiting the bandwidth at the input to avoid slew induced distortions or TIM. The major problem is stability and apparently that's an area that has been paid attention to. The amplifier layout more resembles a radio transmitter than a conventional audio amplifier (Candy quote) with individually shielded stages and global feedback applied via a terminated transmission line.

 

[AKSA]

Why?

Hugh

 

[Samuel Jayaraj]

Matching the LTP devices for VBE and HFE, and then IDENTICAL current flow gives even better results. You need to simulate identical loading of the VAS , on the unloaded LTP side.
A close approximation in many cases, could be as simple as a
1% 22Megohm resistor from the tied base and collector of the current mirror, to the -VE rail. I am not using this method,but my.....
SandyK

I just tried this today in an amp where the current mirror is from the +ve rail with 330E resistors. I just put two 10Meg resistors in series from the tied Base/Collector of the current mirror to the +ve rail.

The sound stage is expansive and the energy from the amp is superb. This was already a well tweaked amp and it only got better. I am not sure about this, but the noise floor of the amp seems to have lowered and the depth perception is a little too much that it almost sounds unrealistic.

I also found that this type of loading does change the tone of the amp (or timbre of instruments) ever so slightly. The do sound a bit harder now as compared to very, very natural without the mod.

Is there a rule of thumb value for the loading resistor? Does this depend on the value of the current mirror resistors?

Another question is, which is the best place to put an offset trim if not somewhere in the LTP degeneration resistors?

Thanks.

 

[AndrewT]

The output offset set trim could be fitted to vary the VAS current, or on the collector load of the LTP, or a DC current injection into either the inverting or non-inverting input. There can be a variety of others if you're ingenious, eg, SKA, but I can't divulge that one, it is simple and clever.

 

[Samuel Jayaraj]

I am not familiar with varying the VAS current method, while the other two are familiar. However, it is much easier to inject a small amount of DC in a fully symmetrical balanced design (where there are two mirrored halves of the whole amplifier). Servo is not accepted by many for its intrusive nature.

Any comments on passive loading of the unloaded side of the LTP?

 

[sandyK]

compensation capacitor Cdom

Sam
I only suggested that resistor , because it was a simple thing for someone else to try. It was based on a guesstimate of the HFE of the transistor loading the other side, and it's bias current. It is not the method that a couple of us use.For best results you need the temp coefficient of a diode in there too.Now that the soundstage is further improved, have a look at your front end input filter. Can it be given a higher rolloff frequency ? Bandwidth increase should bring back the sweetness and give further improved localisation of instruments and voices. You may even be able to slightly reduce the value of what this topic is all about. Don't use a ceramic capacitor.Use a suitably rated polystyrene.
I am not prepared to further elaborate to anybody who just wants to stick things in a Simulator and say the result is a load of crap, and just another "Blameless and Sterile" D.Self clone !
I have been there, done that. I have been trying to get other people to try with an open mind since 1987.
Thanks for taking me seriously Sam. Remember that the LTP transistors should be closely matched for HFE AND VBE, preferably thermally coupled too.
SandyK

 

[AKSA]

Sandy,

I take you very seriously...... one would be a fool not to. However, I've found that pinpoint imaging and sound stage is simply a function of greater resolution. Fix the detail, get the image. It is not some strange notion of phase coherence, which I've always found tricky to decipher, but simply the fine detail of spatial cues which are usually lost in the amplification.

Your ideas are fascinating, please go on!

Cheers,

Hugh

 

[sandyK]

Hugh and Samuel
Further to my original reply. Now that the soundstage is bigger and bandwidth increased (The English designer Graham Slee says you should preserve as many harmonics as possible before rolloff, to assist in this regard. I agree with Graham. His suggestions further improved results) Much of the ambience information is at a very low level , often mixed in with crossover distortion and noise, from source player too. You really need to feed the front end with a very low noise,VERY low impedance supply to >100KHZ . If you are using CD/DCVD players,STBs etc. which use switchmode PSUs, it is best to reduce their noise level using faster bridge rectifier diodes such as BYV26C and a small 630V capacitor across the main electrolytic. Chassis dampening also helps. These measures will further assist in rescuing the low level ambience.The difference is quite noticeable via SPDIF into a high quality DAC, or even HDMI. Even picture quality !
To further illustrate this low level ambience issue, a reviewer of the Kim Carnes -Bette Davis Eyes CD track, commented about the subtle low level echoes etc. , and remarked "Turn off the air conditioner !"

 

[pooge]

Hugh, I have to agree with you that improving the resolution will improve imaging, assuming the speakers are not totally screwing up the phase. And by improving the resolution, I'm assuming you're including reducing the noise, as Sandy says. I've noticed improved imaging and reduced noise by simply changing the resistors in my crossovers, and the I/V resistor in my DAC to Mills wire wounds. So no change in speakers, yet better imaging. The better imaging also comes about when steps to reduce congestion are taken. When things get congested, sounds start imaging closer together, both in width and depth. While the resistor changes mentioned above reduced congestion, others things like capacitors will do the same. Not sure why a resistor can do this, unless there is something to the hysteresis effect with steel leads.

Certainly, congestion and imaging problems can come from diffraction and speaker resonances, also. But reducing those increase resolution, as you say.

On some Dorian CDs, I can hear the air conditioning, as well as trucks going by the recording venue on the street outside. I live in a very quiet area, and I couldn't figure out why I was hearing so many trucks going by all of the sudden. They even sounded like they were coming from my street.

On the other hand, I do think that it is not simply a function of resolution. Phase can have a great affect on imaging. Perhaps you are referring more to the focus aspect of imaging, but I have had instances when messing around with connections, or what not, that a sound caused my head to snap around to the side because and image presented itself there, and I couldn't attribute it to a change in reflections in the room.

 

[AKSA]

Hi Pooge,

Nice to see you here. You wrote:

Hugh, I have to agree with you that improving the resolution will improve imaging, assuming the speakers are not totally screwing up the phase. And by improving the resolution, I'm assuming you're including reducing the noise, as Sandy says.

Absolutely. There are analogies with cooking, in fact. Reducing the salt content will start to remove body and texture from the taste, something we don't normally associate with a salty taste, but at lower levels, this is one effect.

Similarly, increasing top end will tend to detract from the bass, and conversely, increasing bass will somehow remove from the top end. We hear the musical event as a whole, apparently, not as discrete sectors.

Resolution is intimately related to noise levels. A moment's thought shows that any noise will mask low level detail; if this detail is spatial information/cues, then the imaging will suffer. Keep the noise down and awareness of the sound space will increase. However, it's not so hard keeping noise down on power amps, the big challenge is vinyl preamps, and generally, for MC, one needs to use JFETS to achieve this.

When you examine the pole constraints on stability, you find that group delay on a SS amp amounts to one half the periodicity of the waveform at the pole frequency. Most are around 500KHz, so this implies 180 degrees of phase shift at this frequency with roughly linear phase shift up to the point of instability.

Thus, we have a propagation/group delay around 1uS, and at 20Khz, this is 7.2 degrees.

Since we are hearing much diminished harmonics, if at all, at this frequency, it follows that at say 5KHz, a more realistic upper limit to the musical artefacts, we have less than 2 degrees of phase shift as signal passes through the amp. This phase shift will be more pronounced the higher the frequency, and this is the problem....

By comparison, the very sluggish electromechanical speaker drivers will introduce far more group delay, arguably two or three orders of magnitude more, in fact.

I confine my comments to amps, not to speakers, but you make a very good point, essentially that much more reward would be forthcoming if we minimised cone/VC mass and maximised force factor. Presumably only this, discounting the impedance mismatch of the driver/air interface and troublesome back reflections from the enclosure, would reduce phase shift at the driver.

I've noticed improved imaging and reduced noise by simply changing the resistors in my crossovers, and the I/V resistor in my DAC to Mills wire wounds. So no change in speakers, yet better imaging.

Phase shift on speakers is hugely important; it can deviate easily 10 or 15 degrees across the crossover point if the filter is not correctly designed. We've found that changing out resistors can impact on this phenomenon, and there is nothing magical about this as a phase plot on a good software package will demonstrate very clearly. And this does create palpable imaging anomalies. On a recent two way TL design, we managed to get it down sub one degree each side of the crossover at 2500Hz.

The better imaging also comes about when steps to reduce congestion are taken. When things get congested, sounds start imaging closer together, both in width and depth. While the resistor changes mentioned above reduced congestion, others things like capacitors will do the same. Not sure why a resistor can do this, unless there is something to the hysteresis effect with steel leads

You should be able to compensate for this effect by altering the inductors slightly - a quarter or half turn should do it.

And the round, metallised caps, the Solens, have considerable self inductance - showing the entire crossover is a compromised system and tweaking is essential once the raw dimensions drop out of the forumulae.

Cheers,

Hugh

 

[sandyK]

Hugh
If you care to email me , I will give some background on why I,
and a couple of friends are doing these very debatable changes.
It is interesting though, that someone across the other side of the world (Samuel), tried the very rough and ready suggestion that I made, and heard an increase in soundstage ?

SandyK

 

[gaetan8888]

Hi Pooge,

Nice to see you here. You wrote:


...
Resolution is intimately related to noise levels. A moment's thought shows that any noise will mask low level detail; if this detail is spatial information/cues, then the imaging will suffer. Keep the noise down and awareness of the sound space will increase. However, it's not so hard keeping noise down on power amps, the big challenge is vinyl preamps, and generally, for MC, one needs to use JFETS to achieve this.
....

Cheers,

Hugh

Hello Hugh

About noise level and sound imaging, there is noisy tube amps that was sounding with a superb sound imaging and definitions.

Maby wen there is much lower crossover and high frequency distortions, the noise level are a lower problems to the sound imaging and definitions ?

Gaetan

 

[AKSA]

Gaetan,

Yes, true, many tube amps have fantastic depth of image.

However, they are generally poor in image width and focus when compared to the best of the SS brethren.

And tube amps age very fast..... together with the high maintenance/replacement costs, this is serious disadvantage.

Cheers,

Hugh