It is not an easy task to beat the 2700 series 100x.
http://ap.com/products/2700.htm
http://ap.com/pdf/2700Brochure4_04.pdf
http://ap.com/products/2700.htm
http://ap.com/pdf/2700Brochure4_04.pdf
Workhorse said:
Whichever intrigues me most for the moment. Why have favourites?
You see, even the gurus can change their mind, which is another good reason not take anything anybody says as an absolute truth. That is not the same thing as disregarding what they say. The interesting thing is usually not what, but why. It can be interesting to know that he used lateral MOS before and has now swithched to BJTs, but it would be hundred times more interesting to know why he has done so.
Of course, picking John, Charles and Bob and those specific issues were of course only examples to illustrate my point. I could have choosen any other professionals or any other issues.
AndrewT said:
Jones, England
Jones is Welsh.
Try Smith or anything but Jones (Sue)
I had no idea.
I have always thought Jones is as archetypically english as Smith, and I would think most non-native speaker share that delusion. Jones doesn't sound welsh to me.
So, does that mean that the term "keeping up with Joneses" has a hidden meaning, making fun of the welsh?
What is the problem here? I personally have made amp prototypes with Vfet, vertical mosfet, bipolar driven by bipolar, lateral mosfet, and bipolar driven by vertical mosfet. The latter is what I use with Parasound, but if I could make an all vertical mosfet amp that I felt was rugged enough, I would do so.
Bob Cordell avocates vertical mosfet driven by bipolar, but I have not seen anything from him except a prototype using +/- 35V supplies. I don't know if he has a proven, customer proof design for higher supply voltages. At this time, I use +/- 90V, so 35V is not going to cut it with me, even with balanced bridged operation.
No serious designer is completely limited to one type of design. I am sure that Nelson Pass has tried just about everything as well.
Charles Hansen should be commended for trying new design concepts, rather than sticking to modifications of his previous designs.
Bob Cordell avocates vertical mosfet driven by bipolar, but I have not seen anything from him except a prototype using +/- 35V supplies. I don't know if he has a proven, customer proof design for higher supply voltages. At this time, I use +/- 90V, so 35V is not going to cut it with me, even with balanced bridged operation.
No serious designer is completely limited to one type of design. I am sure that Nelson Pass has tried just about everything as well.
Charles Hansen should be commended for trying new design concepts, rather than sticking to modifications of his previous designs.
Bob Cordell said:
Hi Bob,
You are asking a lot, but never mind. Besides, it’s a couple of years ago I’ve developed this thing, so, at this moment, I can’t remember all ins and out or exact numbers, but I’ll do my best. Basically, my system is based on a high resolution sound card (24bit, 192kHz), a PC and specialized software. Such arrangement has a THD threshold of a modest -105db or so. Therefore, I purified the sine source (from the DAC) by means of a bank of passive band filter (1kHz, 5kHz and 20kHz), which also provides a galvanic isolation between PC and DUT. Now, we have a rather clean sine source. But that’s by far not enough. The ADC itself suffers also from distortion. To cope with this problem, I first have a run without the DUT to determine the residual of the ADC (and DAC) and saved this waveform (as a reference). Actually, this run comprises millions of cycles to suppress noise. Next, I determine the frequency and phase response of the DUT, as we will need this later on. Now we do the final run with the DUT included (this takes also a million cycles or so). Of course, the residual from this run is contaminated by the distortion from the sound card, but as we do know the distortion from the latter, we ‘simply’ subtract this one, that is, after a (digital) correction has been applied, based on the frequency and phase response of the DUT.
In a nutshell, this is it.
PS1: The sound cards I’m using: Waveterminal 192X, heavily modified by means of better op-amps and resistors and even a slightly modified topology (a hell of job with these tiny SMD components!). Yes, two of them, as for IMD measurements, the IMD interaction between the channels of single chip DAC of this card spoils to much.
Maybe in the future, I’ll purchase a sound card from Lynx, the best there is, and expensive, but still far less expensive than an AP.
PS2: Measurement bandwidth: <= 98kHz (limited by the sound card or at will lower under software control)
PS3: This setup measures THD, not THD+N!
Cheers,
John, I don't know if you were referring to what I wrote, but if so, I think you misunderstood me. I was absolutely not suggesting that any of you is stubborn enough to always stick to one technique or anything like that. In fact, I was not really intending to say anything about the professionals like youself or Charles, but rather about those who listen or don't listen to you, and how they do or don't do that.
PMA said:
Hi Pavel,
Sure, it isn't easy, and the drawback of my arrangement is a far longer measurement time. Besides, the reference run and DUT run should be done under exactly the same conditions, that is, same (input and output) voltages and temperature (maybe even equal humidity
) wrt the sound card.Cheers,
john curl said:
Hi John,
Probably the best example out there of my chosen design style and philosophy is the Halcro, which Stereophile declared to be their best amplifier. It has plenty of power, as I'm sure you know. It uses VFETs driven by bipolars, and its Error Correction circuit is a fairly direct descendant of the one I published 25 years ago, right down to the way it is compensated (see his patent; the change he made to make it patentable was to drive his EC with a floating power supply bootstrapped from the output signal, as opposed to my use of a limited amount of feedforward supply for the EC and drivers, similar to what Hawksford did).
Cheers,
Bob
john curl said:
Aaahhhh, John, you are asking a very good question!
I don't know, but I know what you are getting at. I didn't say the Halcros were good at everything
. They do poop out a bit into 2 ohms, don't they! At least on the continuous power testing that JA did.Here is my speculation, but it is only that. First, I suspect that on short-term peaks it can put out a lot of current. VFETs are good at that. Even my puny 50 watt MOSFET amp could put out a 40-amp peak into a 0.5 ohm load. So I don't think its the VFETs.
It could be deliberate current limiting, but that is not what I think it is.
I beleive it is Bruce Candy's beloved switching power supply, which he claims to be an advantage in power amps (kind of like the Chord, which didn't meet any of its output power specs). I think his power supply poops out.
I'd love to get my hands on one of those suckers and test it.
Cheers,
Bob
Have you looked at the same peak output current from a JC-1 as measured by the same magazine? How about 5 times more, or so?
This is called 'hard current limiting' and it makes mosfets safe from overload. Now, is that really an advantage? Could it be a problem? What about a difficult electrostatic load?
Think of the cost savings that Halcro got by using mosfet output stages!
This is called 'hard current limiting' and it makes mosfets safe from overload. Now, is that really an advantage? Could it be a problem? What about a difficult electrostatic load?
Think of the cost savings that Halcro got by using mosfet output stages!
Hmm, looks like the current of the DM88 is hard limited to just a little over 16A, doesn't it? Pretty anemic, especially for $40k a pair.
I don't see this as being necessary for a MOSFET output stage though, especially into a resistive load with lots of parallel devices. What's going on? Is he just trying to keep his switching power supply from blowing up?
I don't see this as being necessary for a MOSFET output stage though, especially into a resistive load with lots of parallel devices. What's going on? Is he just trying to keep his switching power supply from blowing up?
I'm new to this thread and have been catching up for the past couple of weeks (great thread by the way).
While I'm a big fan of measurements I also have an observation on the weight that should be applied to listening tests. A while ago I experimented a little with the effects of placing resistors in series with the drivers (I recall someone raised this earlier). What I noticed was a tendency to improve IM performance as a result. The results varied from unit to unit and over frequency and ranged from only marginal improvement (outside of the driver resonance region) to a maximum improvement of about 9dB. Where I saw improvements in IM they are always accompanied by an audible improvement in clarity.
I think the more tests we can do the better but at the end of the day the tests are only truly valid if performed on a closed system. A listening test is a closed system test while tests performed using a distortion analyser on an amplifer output are not.
Has anyone tried measuring amplifiers using microphones in front of the speakers they are driving?
While I'm a big fan of measurements I also have an observation on the weight that should be applied to listening tests. A while ago I experimented a little with the effects of placing resistors in series with the drivers (I recall someone raised this earlier). What I noticed was a tendency to improve IM performance as a result. The results varied from unit to unit and over frequency and ranged from only marginal improvement (outside of the driver resonance region) to a maximum improvement of about 9dB. Where I saw improvements in IM they are always accompanied by an audible improvement in clarity.
I think the more tests we can do the better but at the end of the day the tests are only truly valid if performed on a closed system. A listening test is a closed system test while tests performed using a distortion analyser on an amplifer output are not.
Has anyone tried measuring amplifiers using microphones in front of the speakers they are driving?
PMA said:
Correction: it is 15A (too low, though).
See page 11 of
http://www.onahighernote.com/documents/products/dm58dm68manual.pdf
They say: "As we do not wish to set your house on fire! we have limited the peak output current
to 15A."
