I call is Semi Class A.
Bias the output with 275mA. Gives Class A with more than 0.5 Ampere into 8 Ohm = 1 Watt.
This makes the amp go in Class A for the 1 Watt level and the rest in Class AB.
Is this a good idea?
The First Watt is the important. Is it not?
Bias the output with 275mA. Gives Class A with more than 0.5 Ampere into 8 Ohm = 1 Watt.
This makes the amp go in Class A for the 1 Watt level and the rest in Class AB.
Is this a good idea?
The First Watt is the important. Is it not?
It's called AB... That's how AB works - one device turns off for less than 50% of the cycle.
My amp designs generally output ~2WRMS as class A before entering class B operation (bias the output for 40mA 320V (tube amp)).
Indeed the "First Watt" is important. If you biased your amp to 20mA, it wouldn't sound the same.
My amp designs generally output ~2WRMS as class A before entering class B operation (bias the output for 40mA 320V (tube amp)).
Indeed the "First Watt" is important. If you biased your amp to 20mA, it wouldn't sound the same.
Ya. It's almost required with tubes to have a large Class A window (when using triodes - they are harder to turn off than on so you get crossover distortion if they run too cold).
A Class B (tube) amp is inherently higher in distortion than class A. The came up with AB as a compromise.
This will never get resolved, LOL. It may be possible to run a MOSFET output stage so the side that's in the "off" half cycle doesn't fully switch off at all, so it's kind-of class-AB trying to be push-pull class-A with a low idle current.
With class B, it irks me that the distortion curves have a downward slope. Without hearing such an amp to make sure, I would presume it would have some kind of audio-correlated "buzz" at a constant volume, so it relies on loud signals to mask it.
With class B, it irks me that the distortion curves have a downward slope. Without hearing such an amp to make sure, I would presume it would have some kind of audio-correlated "buzz" at a constant volume, so it relies on loud signals to mask it.
But a higher % distortion at a higher power can be less audible. If 90% of the waveform is reproduced completely free of the crossover region. The higher distortion during the peaks is masked.
I play around with very high power amps, many running in class H where the power supply rail transitions are clearly visible on the scope. Because it happens when the music is LOUD and for short periods of time you simply don’t hear it. A similar glitch happening at the zero crossing sounds like a mistuned FM radio (Half music, half static). It happens more often, on every cycle, and down at the average level where most of the signal is.
There is no ”optimum class B” for tubes like there is for emitter follower BJTs. The optimum is a class AB with significant bias current and highly dependent on tube type. Class AB in tubes takes advantage of gm doubling rather then being hindered by it. You want to bias the tube stage such that the sum of the two gm’s, times it’s effective load R (which also changes as you transition out of class A) remains relatively constant from the crossover up to as high a power as practical. There is no analog of this in a solid state amp - the turn on’s are just too abrupt.
You are asking the question answered by my thread on class AB biasing. https://www.diyaudio.com/community/threads/class-ab-biasing-article.393566
Synopsis: bipolar transistors in class AB have an optimum bias that minimizes distortion.
Ed
Synopsis: bipolar transistors in class AB have an optimum bias that minimizes distortion.
Ed