And not least ...
"amplifiers should operate inclass A for most of the signal time,and derived the corresponding A- to B- class transition level to be about 17 dB below the full output power."
Still relevant ....?
"amplifiers should operate inclass A for most of the signal time,and derived the corresponding A- to B- class transition level to be about 17 dB below the full output power."
Still relevant ....?
Pass likes the "first watt" principle FIRST WATT
interestingly searching Pass' posts here at diyAudio - he has never mentioned Otala or Lohstroth
other lines of looking for the history would be push-pull tube amp output bias, how much "Class A" region into the speaker load after output xfmr
interestingly searching Pass' posts here at diyAudio - he has never mentioned Otala or Lohstroth
other lines of looking for the history would be push-pull tube amp output bias, how much "Class A" region into the speaker load after output xfmr
No, just that high speed, and (perhaps) high open loop bandwidth can be better for some reason. So far, it has been my experience that it is so.
However, low open loop bandwidth is perhaps not the only problem with IC's (in general). It might be the output stage current, and its subsequent open loop non-linearity.
However, low open loop bandwidth is perhaps not the only problem with IC's (in general). It might be the output stage current, and its subsequent open loop non-linearity.
In the paper linked here, 6th page, Figure 6-A, is that 1 kHz response unusual?
The square wave shape.
You can set the bias in an AB output stage for lowest distortion at full power which will get you at the 'optimum bias'. With that setting, at low levels, Gm-doubling and xover distortion will produce high non-linearities in narrow time-domain places which, although having a very low THD value, are audibly Not Nice.
Since we listen to an average level of a watt or so, one choice you can make is to set the bias higher, for best linearity at that level, but this then gives worse linearity at full power.
This may not be a problem because at very high SPL you're hardly hearing the distortion anyway.
Pick your devil.
Jan
The Lohstroh and Otala paper is an interesting read with a lot of helpful insight into decisions about the internal structure. Its similar to the older HP and Tek literature where they actually explain what things were done and why.
I would take issue with reading "rules" into the discussion. Most of it is a rational effort to solve clear problems, like getting 1 MHz response from an amplifier and how to compensate it so it doesn't oscillate. The gain target (a very high by today's standard of 38 dB) and the extended bandwidth would dictate more than a few of the decisions made. Two stages of internal gain plus the VAS would be needed to have enough open loop gain at 1 MHz. The compensation tricks and internal degeneration would all help to keep some phase margin and the obvious DC coupled path and differential circuitry reduces what the feedback needs to deal with. Note the very high drive at the VAS, needed to drive the driver and output transistors at high frequencies.
I think the effort to categorize the internal HF overload as a specific "Transient Intermodulation" distortion made it too easy to focus on a narrow understanding of the broader issue of handling ultrasonic signals in an amplifier. While its neat to have a marketing spin on something like that the polarizing aspects in effect hide many of the other straightforward aspects that make this a valid quality amplifier that meets its design goals. Unfortunately its hard to generate excitement with straightforward rational engineering, even in the dull world of the AES.
I would take issue with reading "rules" into the discussion. Most of it is a rational effort to solve clear problems, like getting 1 MHz response from an amplifier and how to compensate it so it doesn't oscillate. The gain target (a very high by today's standard of 38 dB) and the extended bandwidth would dictate more than a few of the decisions made. Two stages of internal gain plus the VAS would be needed to have enough open loop gain at 1 MHz. The compensation tricks and internal degeneration would all help to keep some phase margin and the obvious DC coupled path and differential circuitry reduces what the feedback needs to deal with. Note the very high drive at the VAS, needed to drive the driver and output transistors at high frequencies.
I think the effort to categorize the internal HF overload as a specific "Transient Intermodulation" distortion made it too easy to focus on a narrow understanding of the broader issue of handling ultrasonic signals in an amplifier. While its neat to have a marketing spin on something like that the polarizing aspects in effect hide many of the other straightforward aspects that make this a valid quality amplifier that meets its design goals. Unfortunately its hard to generate excitement with straightforward rational engineering, even in the dull world of the AES.
Joachim, the original O/L amp had 20 dB of GNFB, read the text and you'll see, and you'll find their original measurements. With much better measurement results. The point is, the Norwegians went wrong somewhere.
If memory serves, the first series of Electrocompaniet were actually made somewhere in Scotland, and if you drink beer in the same pub as Ivor Tiefenbrun, well hell baby, you just have to have great sound.
I heard the Electrocompaniet only once, but I had my sample of the original at the time, and it did NOT sound as good as the O/L amp. Not much worse, but not quite like it. That was my impression.
Because statistically the CONTINUOUS power we normally use for everyday listening is often well below 1W, but of course, that will depend on the speaker efficiency.
Therefore, if you make your class AB amp run in pure class A up to that point, or a bit above it, most of your listening will happen in pure class A.
The O/L amp ran in pure class A up to 2.9/5.8W into 4/8 Ohms, bias current was 300 mA per each of the two output pairs.
I agree with this.
Statistically speaking, the majority of amps I like do have a reasnobal wide OL bandwidth. I cannot swear on the Holy Bible that this is THE key, but I also cannot overlook the fact that in many cases, to me, this does appear to be an important point.
Let me put it this way - I think that a wide OL bandwidth may not guarantee a good sound, but it seems to provide better chances for it. Ultimately, if I can, why not?
Only because ONE aspect of the amp was taken as the absolute, which in itself is ridiculous, as no single point will ever be all that important to guarantee a good sound.
That is just as narrow minded as the high NFB factor group hypes up high NFB as most important. That's another story we'll never hear the end of.
Bravo, Demian!
That's eaxctly hitting the nail on the head. The instant anything as relatively complex as an audio power amplifier is reduced to just one aspect, the mistake becomes built in.
No amp will sound good just because it has a wide OL bandwith.
No amp will sound good just because its has 60+ dB of GNFB.
If it were as simple as that, we would have had the Prefect Amp a long time ago.
Found this on TAS.
Setting Up A Phono Cartridge | The Absolute Sound
They use Adjust + at some point. :roll eyes:
I like wow & flutter feature the most on it.
- Status
- Not open for further replies.