H
HAYK
Nicely put. The actual standing current in the output doesn't matter as the output will be dc. It's just a bias level.
AB is different - the A aspect does drive the output and at some level switches to B.
Class C. A pulse of some size driving a tuned circuit. Dubious for audio but used for RF to produce a sine,
Then what's a lightly biased MOSFET amplifier configured as common drain push-pull?
Say, lateral 'FETs running 100-200mA idle don't appear to fully switch off at all.
The above was simmed with Exicons running on +/-15V rails, and around 3.x V between the gates. The 2 squished curves show voltage across 100m ohm sense resistors (drain side) for the N and P MOSFETs, (20mV = 200mA) and the output was loaded with a not-too-shabby 8 ohm + LRC speaker model.
Distortion? Circa 0.00005%.
At higher output levels, the valleys get progressively more squished close to the 0mA floor.
There are at least 2 active threads with Exicons, and they both have "hard class A" running at extremely high bias, which appears to overlook that 'sticky' property where, as the current drops near zero, both the temp-co and the increasing Vds work together to maintain a bit of current leakage.
There is also Ian Hegglun's square law class A with a quiescent current of 1/4 the peak output current.
My amplifier has a quiescent current of about 100 mA, but the current through one side doesn't go below about 50 mA in the peaks of the current through the other side. I call it non-switching class AB.
My amplifier has a quiescent current of about 100 mA, but the current through one side doesn't go below about 50 mA in the peaks of the current through the other side. I call it non-switching class AB.
I was using the forward-bias term in Ebers-Moll. After I posted, I realized that someone would point out the reverse-bias term.
Ed
@abstract in your case where the opposing transistor doesn’t ever switch off, then that looks to me like class A or AB, but this seems to be a consequence of the low supply voltage as the modest bias current doesn’t really put it anywhere close to class A as usually understood. According to the conduction angles it would be class A, the kind of distortion figures you get also makes it look rather like Class A!
How do you think about this one? An outlier or hard and fast (something)?
(My depth of knowledge and experience is being challenged here, but that is good as far as I’m concerned as I find this topic very interesting and I hope I’m getting more right than wrong 🙂)
How do you think about this one? An outlier or hard and fast (something)?
(My depth of knowledge and experience is being challenged here, but that is good as far as I’m concerned as I find this topic very interesting and I hope I’m getting more right than wrong 🙂)
On large signals, the impedance of the output stage of a blameless class B
is much more linear than the impedance of a class AB.
is much more linear than the impedance of a class AB.
Last edited:
So, just to loop back on class B amplifier chips, this one from NXP is an example https://www.nxp.com/docs/en/data-sheet/TDA1517.pdf
Unfortunately not a great chip, 6 W per channel stereo with 0.1% THD at 1W and 10% at full power.
Unfortunately not a great chip, 6 W per channel stereo with 0.1% THD at 1W and 10% at full power.
I have been looking at two possible Class B builds the last month or so. I cannot comment on their topology or how true to Class B they are.
https://www.diyaudio.com/community/threads/the-blomley-class-b-amplifier.322744/
https://www.diyaudio.com/community/...ee-pole-compensated-class-b-retro-amp.385080/
I am planning on building the H-ENG 60V and will get boards from member @002 early next year.
Thank you all for the orderly and very informative discussion, greatly appreciated.
As far as I can observe the single practical example from @Mooly can be interpreted as a Class C amplifier.
The information and examples posted so far leads me to conclude with my initial suspicion, that Class B is mostly a theoretical type of solution that does not have any particular real-life use.
If anyone have a practical use-case of any Class B amplifier I would still like to know about it, hoping this could possibly debunk my suspicions to some limited extent. Especially teachers love to preach about the Class B amplifiers but what is the point if it is a sub-par solution that does not really exist.
As far as I can observe the single practical example from @Mooly can be interpreted as a Class C amplifier.
The information and examples posted so far leads me to conclude with my initial suspicion, that Class B is mostly a theoretical type of solution that does not have any particular real-life use.
If anyone have a practical use-case of any Class B amplifier I would still like to know about it, hoping this could possibly debunk my suspicions to some limited extent. Especially teachers love to preach about the Class B amplifiers but what is the point if it is a sub-par solution that does not really exist.
i have a feeling that some have problems realising that a bias current in an output stage has nothing to do with driving a loudspeaker. When the drivers are actually running at that level the amp is outputting DC. Class AB is entirely different. Class C. Definitions of what that is can be found on the web.
KaffiMann - Class B and class C are widely used in RF.
SSB radio transmitters are class B (push-pull with zero-bias). The modulation is applied before the power amplifier. Thus, the power amplifier has to be linear (it is even referred to as a "linear amplifier"). A tuned circuit eliminates harmonics.
AM radio transmitters are class C. The modulation is applied to the RF amplifier's power supply. Class C amplifiers are single-ended. Once again, a tuned circuit eliminates harmonics.
Audio is limited to class A and AB because audio is wide-band.
Ed
SSB radio transmitters are class B (push-pull with zero-bias). The modulation is applied before the power amplifier. Thus, the power amplifier has to be linear (it is even referred to as a "linear amplifier"). A tuned circuit eliminates harmonics.
AM radio transmitters are class C. The modulation is applied to the RF amplifier's power supply. Class C amplifiers are single-ended. Once again, a tuned circuit eliminates harmonics.
Audio is limited to class A and AB because audio is wide-band.
Ed
Last edited:
Depends on which of the two definitions you use, see this whole thread.
On second thoughts the 2 sides do appear to fully switch, but it was clearly just a nice 'AB' setup that needed more voltage swing to reach that point.
So I was reading up on the cube-law amplifiers. For one thing the floating PSU technique seems like a really neat way to run the output transistors with gain / "common source". But I'll have to play around with the distortion cancelling stage to see if it's something I'd want to do.
An extract from a book
Class B or Class AB?In a class B amplifier, the top and bottom halves each conduct for one-half cycle and the handoff from one side to the other is abrupt. In a class AB amplifier, the stage behaves largely like a class B amplifier, but some overlap of conduction is permitted, so that the transition of contribution to the output signal is a bit more smooth and broad. In the most widely used output stage, a small optimum bias is applied, called the quiescent bias Iq. In the region of ±2Iq the stage is essentially in class A, with both transistors contributing transconductance. This chapter is focused exclusively on class AB output stages. Other authors have described this mode of operation as class B, and this is really just a matter of semantics
https://pearl-hifi.com/06_Lit_Archi...dell_Bob/Designing_Audio_Pwr_Amps_Cordell.pdf
So with this bias arrangement we are talking about mA in classA. Hardly significant in terms of driving a loudspeaker. In Class AB the A drive capability could be whatever some one wants and rather pointless for it to be in a similar mA area to a typical bias setting.
So if some one on the basis of semantics wants to confuse things fine. Seems my father with his a Armstrong amp bought in the 70's was using an AB amp.
Class B or Class AB?In a class B amplifier, the top and bottom halves each conduct for one-half cycle and the handoff from one side to the other is abrupt. In a class AB amplifier, the stage behaves largely like a class B amplifier, but some overlap of conduction is permitted, so that the transition of contribution to the output signal is a bit more smooth and broad. In the most widely used output stage, a small optimum bias is applied, called the quiescent bias Iq. In the region of ±2Iq the stage is essentially in class A, with both transistors contributing transconductance. This chapter is focused exclusively on class AB output stages. Other authors have described this mode of operation as class B, and this is really just a matter of semantics
https://pearl-hifi.com/06_Lit_Archi...dell_Bob/Designing_Audio_Pwr_Amps_Cordell.pdf
So with this bias arrangement we are talking about mA in classA. Hardly significant in terms of driving a loudspeaker. In Class AB the A drive capability could be whatever some one wants and rather pointless for it to be in a similar mA area to a typical bias setting.
So if some one on the basis of semantics wants to confuse things fine. Seems my father with his a Armstrong amp bought in the 70's was using an AB amp.
Class B means that that there is conduction during half of a cycle, at least.
As a silicium bipolar transistor only begins to be conductive when the bias
is a slightly less 0.4 V, your definition of class is not correct
A transistor alone can work in Class A and in class B.
