John Curl's Blowtorch preamplifier part II

[Joachim Gerhard]

To do that most designers parallel a lot of output transistors. I have that seen first in early ML gear and in Treshold amps. Here in Europe that was an unusual sight. Using a lot of output deviced needs a very stable driver stage though, so some designers here go back to a single pair of output transistors and claim audible benefits. I am no expert in power amps and i only post my observations here.

[PMA]

Originally Posted by Joachim Gerhard

so some designers here go back to a single pair of output transistors and claim audible benefits.

Marketink hype, nothing else. Single pair of usual output power devices works well only for low power amplifier.

[Joachim Gerhard]

Gammut uses only one pair of very high power Fets and do a high power amp that way.
Some years ago i marketed a class d amp hat also used only one pair of transistors per side of an H-bridge that could do 800W in 4 Ohm. I do not want to start a discussion if one pair sounds better then several. I certainly heard amps that sound excellent with many output devices for example my silicon reference Audionet Max2.

[EUVL]

> Here in Europe that was an unusual sight.

Not for the output stage, but here is an exception (see Abb.7)

GRUENSCH MSE - Monaural Reference Amplifier, Mono Endstufe, Verstärker, Schaltnetzteil, MOS-FET


Patrick

[john curl]

The real, practical reason for using so many output devices in parallel is to improve the high voltage safe area, so that higher voltages can be used in the power supply. A side benefit is increased peak current available and spreading of the heat generated by a large idle current along with a high voltage, over a large surface area.

[Joachim Gerhard]

Hy Patrick ! What i ment are the old times, before 1980. Now you see a lot of transistors in the output stage of many amps but i think that all started in the USA. They really had the edge that that time.
What i liked so much in the Treshold 400A and Stasis ( I could not afford ML back then, but at some time i owned a Levinson active crossover for my self made HQD system ),
was that the treble was sweet and there was a lot of depth in the soundstage. I think my Preamp was an NS10. i liked the big and interesting shaped knobs.

[Joachim Gerhard]

Later i had an SL10. That really blew my mind.

[1audio]

I have a problem with the measurement process and the results as I understand Joachim's description. The core of it being that a low source impedance will cause the impedance curve to be imposed on the frequency response as derived from a transient from an MLS signal. I have seen the opposite in measurements myself and using the same technique I have never seen the impedance curve affect the response that way. Usually its the opposite if anything significant happens.

However re: Damping factor. As a marketing tool a silly number is good. That number is usually only present at very low frequencies. Its possible in such an amp that the DF would be lower at higher frequencies than an amp dependent only on the output Z of the output followers.

Putting the two above together it is conceivable that the impedance variations in the HF section of a speaker from various crossover and driver issues could interact significantly with the changing output Z of different power amps. John Curl has strongly advocated against a series output inductor on his amps, probably for just this reason. And the constant loop gain in the audio passband will help create a constant output Z from the amp. Not as mystical or exciting as a new form of distortion perhaps but probably a real and measurable effect. If I had some samples if amps and speakers with these qualities handy I could try measuring these differences the next time I measure some speakers.

[Joachim Gerhard]

John or his associates could send an amp of his to Test Factory and look how it behaves.

[1audio]

Trying to read through the article (my German is zilch) I think they say there is a difference between sine wave and MLS measurements. That is true and easily explained- the sine wave measurement uses a wide band detector so harmonics will add with the fundamental, the MLS is intrinsically narrow band, it needs to be to derive the individual frequency response. Further the MLS system is band limited by the A to D and D to A conversions.

After that I really could not make sense of the article.

If I can understand their measurement process I could duplicate it and see if I can get the same results.