Yes, the OT dominates the LF response in most reasonable amplifier designs.
indeed, but recently i made a ppp 6bq5 amplifier using the original copy of the A471 from Triode electronics and boy i can hear the kick drum distinctly, not as loud as i wanted but it is there....
another learning for me, is that you need not used big big coupling caps either, i thank SY for that....
Last edited:
That's because HP and LP are mirror images of each other. With the same R and C, they both give exactly the same -3dB point!
Have a look at a simple gain stage. I have deliberately drawn it so the PSU dropping resistors (R1 / R2 / R3) are 'vertical' to make it obvious that, from the point of view of the triode, they are really part of the anode load resistor.
In the first case there is no PSU bypass cap, so the triode sees a load of 56k+10k = 66k and acheives a flat gain.
In the second case a 1.5uF bypass cap is used. At very low frequencies the cap looks like an open circuit, so the gain is as before. But at high frequencies the cap looks like a short circuit (it 'bypasses'), so the triode sees a load of 56k, so the gain is less. In between is a transition -creating a sort of bass boost. Usually we want this boost to be outside the audible range, or at least small within the audible range.
Making the cap much larger, as in the third case, pushes the boost down lower, so the gain stays flat from 20Hz upwards.
The amount of boost may be small, which is why you may not be able to hear any difference. In this example I threw together it is not even 0.1dB! But hopefully you understand the mechanics. There will be other situations where the frequency response deviation could be larger, esp when coupling transformers get involved. And remember, the power supply is not really a battery as in this ideal example, but a whole series of other filters 'upstream'! You can also play tricks with the cathode bypass cap, to cancel out any boost, but this is just a forum post. It's why people write whole books on these subjects.
Attachments
Last edited:
Exactly.
With low Rp tubes like a 300B it's really easy to get 4Hz @3-dB with small signal, which means likely 3Hz at least with larger (1W for sure normally) signal with FeSi cores. The permeability is not perfectly constant in gapped FeSi cores and it will increase up to the max rated power if the transformer is well designed. Inductance decreasing before max rated power is reached is not a good design....
Having said this the requirement of 560 uF in my opinion is excessive because one thing is the frequency response at small and medium signal and another thing is at max rated power where the charge-discharge in the PSU will become most critical. In such case 30Hz is normally the "reference" frequency for max output in a well designed transformer without exaggerating with size with EI cores. No lower than 20-25Hz with C cores. For me, personally, it's 30Hz regardless of the core type.....
This means power bandwidth (-3dB) starting from 20Hz.
The "music with massage" is no good for standard loudspeakers anyway (99.9% of them easy, I think). So this is not really a weakness of the typical SE amp.
Best case of 300B amplifier. Let's assume we have a top quality output transformer with HiB C-core and this is 5K and is optimized to work with 80mA DC current to give best inductance and max headroom at 20Hz.
What does it mean optimized for 80 mA max headroom at 20 Hz? It means that the DC induction is taking no less than 50% of the headroom, normally slightly more and the rest is for the signal until the transformer saturates. Let's assume that the max rated power @20Hz is 12.5W. this means we are applying 250V rms signal.
What is the power we can apply at 10Hz and 4 Hz before the transformer saturates? Easy the voltage we can apply is inversely proportional to frequency so we get just over 3W @10Hz and 0.5W at 4Hz....
What's the point of sizing the cap for large persistent transients at 4Hz?
Of course one should not exaggerate in the other direction but as always find a good balance among all requirements.
What does it mean optimized for 80 mA max headroom at 20 Hz? It means that the DC induction is taking no less than 50% of the headroom, normally slightly more and the rest is for the signal until the transformer saturates. Let's assume that the max rated power @20Hz is 12.5W. this means we are applying 250V rms signal.
What is the power we can apply at 10Hz and 4 Hz before the transformer saturates? Easy the voltage we can apply is inversely proportional to frequency so we get just over 3W @10Hz and 0.5W at 4Hz....
What's the point of sizing the cap for large persistent transients at 4Hz?
Of course one should not exaggerate in the other direction but as always find a good balance among all requirements.
Or with imaginary buckets and long or skinny hoses. SPICE does NOT understand the problem. That is your job. It may not be an easy job; few practical engineering problems are.
45: This is a good point. Loesch uses 4 Hz as an ideal, but does mention that a higher frequency might make more sense in the real world. The ideal criteria set out by his article, 4 Hz and 40 MS in his 300B example, may be unrealistic and completely unnecessary most of the time, and I think that we understand that this is the case. What I want to know is whether or not the method is valid. If so, a person using his method can insert whatever criteria they want.
Well, let's see what an utterly practical man might do. This is a Fender amplifier. Leo was an accountant who fixed radios and built a few instrument amplifiers. Work-out the time constants for the several stages of power supply.
