1: What constitutes an amplifier as being "pure class A"?
2: What are some brands of solid state PURE class A home audio amplifiers?
I use to think it was pretty straight forward and thought that class A meant the transistor is constantly on and biased as such to drive the transistor between cutoff and saturation. So for me pure class A meant that this would be so in all stages.
All the different claims of class A operation is confusing. Some of the older technics designs of class A+, class AA, new class A come to mind, Denon's optical class A, Krell's sustained plateau bias and on and on. you have single ended, push pull.... help simplify things for me if possible. Am I really confused or is it as simple as you have pure class A and all other variations/flavours of class A operation?
2: What are some brands of solid state PURE class A home audio amplifiers?
I use to think it was pretty straight forward and thought that class A meant the transistor is constantly on and biased as such to drive the transistor between cutoff and saturation. So for me pure class A meant that this would be so in all stages.
All the different claims of class A operation is confusing. Some of the older technics designs of class A+, class AA, new class A come to mind, Denon's optical class A, Krell's sustained plateau bias and on and on. you have single ended, push pull.... help simplify things for me if possible. Am I really confused or is it as simple as you have pure class A and all other variations/flavours of class A operation?
ClassA can be either single ended or push-pull.
looking at single ended first, valve or solid state.
A single active device is biassed to pull the full output current into the worst loading the speaker can demand. for 10W into 8ohms this will be Vpk=12.6Vpk output, Ipk=2.1A (for Re=6r0).
The supply rail/s will be set up for about 2*12.6V+~5V~=30V.
The amplifier can supply upto ~2.1Apk to the load and if designed well sound really nice. The transistor must never turn off and must never saturate. If it did the distortion in the output signal would sound terrible. Keep that thought, we'll need it for the push-pull version. I chose low supply rails and low ClassA power to limit the dissipation in the active device. 30V * 2.1A=63W, very hot.
A push-pull ClassA amplifier is biassed to half the peak current into the worst (normal) load the speaker can demand. The push-pull can deliver much more than ClassA peak current if allowed to do so (but not in ClassA, it automatically converts to ClassAB when overloaded, but without the enormous increase in distortion of the single ended version).
Lets set up the PSU and active devices as for the single-ended. 30Vdc and 2.1A of bias.
The upper active devices pass 2.1A and the lower active devices sink this same current. As shown in those earlier graphs the upper device changes current with signal level and the lower device changes current in the opposite direction. The difference in current between the upper and lower devices is delivered to the load. The ClassA limit for this is approximately two times the bias. so if biassed to 2.1A the maximum ClassA current to the load is about 4.2Apk. if this is pushed through a 3r0 load then it needs Vpk=12.6V. exactly the same as the single ended. The maximum 8ohm ClassA power is also the same as the single ended.
Now we come to what happens if the worst load is below 3r0.
let's assume that we connect a 2r0 resistor for a load.
the maximum output voltage is 12.6Vpk. the resistor will draw 6.3Apk. This is well above the 2times bias value. HOW? WHY?
the upper device supplies all the current, provided the PSU can deliver and provided the designer has not incorporated current limiting. The lower device tracks the upper device as before until it's minimum current drops to a very low value This is the lower transistor effectively cut-off. The amp is no longer in ClassA. The big difference now between single-ended and push-pull is that the cut off transistor does not cause gross distortion, it simply takes no part in controlling the output current. For negative peak current the opposite happens. The lower device sinks extra current and the upper device sources less current as the signal changes. Again the limit of ClassA arrives when the upper devices cut-off.
Now look at the device dissipation of the PP ClassA 30V*2.1A=63W but shared equally between upper and lower devices AND it's capable of delvering twice the ClassA current. That is a very large increase in output/input efficiency and each device dissipates just half the power. No wonder Push-Pull is so popular.
The definition I adopt for Push-Pull ClassA is that both devices must actively control the current in the load with neither becoming near constant current and/or cut-off.
This differs from the general definition where it is stated that neither device should cut-off. I believe Leach's explanation of drivers and ClassA falls into that trap. Go read his Low Tim design paper to see the alternative view.
looking at single ended first, valve or solid state.
A single active device is biassed to pull the full output current into the worst loading the speaker can demand. for 10W into 8ohms this will be Vpk=12.6Vpk output, Ipk=2.1A (for Re=6r0).
The supply rail/s will be set up for about 2*12.6V+~5V~=30V.
The amplifier can supply upto ~2.1Apk to the load and if designed well sound really nice. The transistor must never turn off and must never saturate. If it did the distortion in the output signal would sound terrible. Keep that thought, we'll need it for the push-pull version. I chose low supply rails and low ClassA power to limit the dissipation in the active device. 30V * 2.1A=63W, very hot.
A push-pull ClassA amplifier is biassed to half the peak current into the worst (normal) load the speaker can demand. The push-pull can deliver much more than ClassA peak current if allowed to do so (but not in ClassA, it automatically converts to ClassAB when overloaded, but without the enormous increase in distortion of the single ended version).
Lets set up the PSU and active devices as for the single-ended. 30Vdc and 2.1A of bias.
The upper active devices pass 2.1A and the lower active devices sink this same current. As shown in those earlier graphs the upper device changes current with signal level and the lower device changes current in the opposite direction. The difference in current between the upper and lower devices is delivered to the load. The ClassA limit for this is approximately two times the bias. so if biassed to 2.1A the maximum ClassA current to the load is about 4.2Apk. if this is pushed through a 3r0 load then it needs Vpk=12.6V. exactly the same as the single ended. The maximum 8ohm ClassA power is also the same as the single ended.
Now we come to what happens if the worst load is below 3r0.
let's assume that we connect a 2r0 resistor for a load.
the maximum output voltage is 12.6Vpk. the resistor will draw 6.3Apk. This is well above the 2times bias value. HOW? WHY?
the upper device supplies all the current, provided the PSU can deliver and provided the designer has not incorporated current limiting. The lower device tracks the upper device as before until it's minimum current drops to a very low value This is the lower transistor effectively cut-off. The amp is no longer in ClassA. The big difference now between single-ended and push-pull is that the cut off transistor does not cause gross distortion, it simply takes no part in controlling the output current. For negative peak current the opposite happens. The lower device sinks extra current and the upper device sources less current as the signal changes. Again the limit of ClassA arrives when the upper devices cut-off.
Now look at the device dissipation of the PP ClassA 30V*2.1A=63W but shared equally between upper and lower devices AND it's capable of delvering twice the ClassA current. That is a very large increase in output/input efficiency and each device dissipates just half the power. No wonder Push-Pull is so popular.
The definition I adopt for Push-Pull ClassA is that both devices must actively control the current in the load with neither becoming near constant current and/or cut-off.
This differs from the general definition where it is stated that neither device should cut-off. I believe Leach's explanation of drivers and ClassA falls into that trap. Go read his Low Tim design paper to see the alternative view.
Thanks for your response and time Andrew, its very much appreciated. Cheers
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.