It's not as simple as you think. If you use source followers, you still have to limit the positive grid excursion somehow to save the output grids from overheating. If you're not careful, this limiting mechanism will reintroduce the same blocking distortion you were trying to get rid of.
Disclaimer: maybe you can get away without limiting the positive grid voltage in hi-fi if you only ever listen to classical music and don't throw any parties. 🙂 I've seen people try this in guitar amps and it ended in blown MOSFETs and burnt grid stopper resistors as soon as the amp was overdriven.
Crowhurst had quite a lot to say about the behaviour of Williamson-type amps under overdrive, none of it pleasant. Generally the more feedback you use, the worse the circuit reacts to clipping. I think this may be one reason why feedback got its bad rap.
Disclaimer: maybe you can get away without limiting the positive grid voltage in hi-fi if you only ever listen to classical music and don't throw any parties. 🙂 I've seen people try this in guitar amps and it ended in blown MOSFETs and burnt grid stopper resistors as soon as the amp was overdriven.
Crowhurst had quite a lot to say about the behaviour of Williamson-type amps under overdrive, none of it pleasant. Generally the more feedback you use, the worse the circuit reacts to clipping. I think this may be one reason why feedback got its bad rap.
Last edited:
I've been doing this for a few decades and have found that some types of tubes are happier about it than others. I can attest that EL84 is NOT a good choice for being driven with a source follower, thus the different approach taken in the RLD amp. On the other hand, I had great success with 2A3, 6L6GC, and EL37, never melted a grid.
Since clipping in a hifi amp tends to be a transient event, the considerations are different than in a guitar amp which is expected to spend much of its duty cycle in clipping. The main thing in a hifi amp is to prevent the transient effect from causing the amp to "choke" until the capacitances are allowed a few time constants to drain their charge, turning a millisecond event into a several second event. Crowhurst indeed wrote extensively about blocking, and that's what put me down that path initially, back in the late '70s. The availability of MOSFETs in the early '80s caused me to change from cathode followers to the superior source followers, and I've never looked back.
If I want loud music for a party, I have some nice high power solid state amps on hand.
Since clipping in a hifi amp tends to be a transient event, the considerations are different than in a guitar amp which is expected to spend much of its duty cycle in clipping. The main thing in a hifi amp is to prevent the transient effect from causing the amp to "choke" until the capacitances are allowed a few time constants to drain their charge, turning a millisecond event into a several second event. Crowhurst indeed wrote extensively about blocking, and that's what put me down that path initially, back in the late '70s. The availability of MOSFETs in the early '80s caused me to change from cathode followers to the superior source followers, and I've never looked back.
If I want loud music for a party, I have some nice high power solid state amps on hand.
You have to draw the line somewhere. Some people won't even accept silicon diodes in the power supply.
I tend to agree, having two channels reproducing two partly uncorrelated signals is what makes a stereo setup sound "roomy". 🙂
You rang? Actually, SiC Schottky's are really nice devices for the PS, but I get such good results with vacuum rectified choke-input supplies that I don't bother. If I were to use SiC Schottkys I would still use a choke input supply. Cap input supplies blow for a variety of reasons - the sole reason they are used is cost. But I digress.....
Of course I use silicon Schottkys for the heater/filament supplies, also in a choke-input configuration. Heck, if I wanted "light" end "efficient" I would run a class D amp with a switching supply. 😀
The only reason choke-input supplies were ever used in the first place was because cheap, small, high-capacity electrolytics didn't exist. They are about as relevant to hi-fi as vacuum tubes themselves. 😉
Choke input supplies present a far more benign load to the incoming mains supply, so improving power factor and reducing interference with other units on the same supply. They also can reduce the sharpness of charging pulses and hence the risk of induction into nearby circuit loops. They still have a place.
The main part of a choke-input power supply is an air-gapped choke that spews AC magnetic flux in all directions. I don't think the risk of induction is reduced at all.
True, but most of that flux is at 100/120Hz and the airgap is likely to be well away from the rest of the circuitry.
Also, in a choke input supply the AC peak current is equal to or less than the DC current drawn. In a cap input supply the AC peak current is always greater than the DC current.
Also, in a choke input supply the AC peak current is equal to or less than the DC current drawn. In a cap input supply the AC peak current is always greater than the DC current.
Last edited:
CLC supplies are still superior to CRC supplies even today. I use a CLC supply in my Pass Aleph 4 and the PSU noise is remarkably reduced.
CLC is a capacitor input power supply since the first letter is C. A choke-input supply would be LC.
I like the CLC myself, especially for single-ended power amps that have very poor PSRR. The choke mainly carries DC current with only a small ripple, so it doesn't buzz or radiate a hum field. You also don't have the problem of the output voltage skyrocketing below the critical current.
I like the CLC myself, especially for single-ended power amps that have very poor PSRR. The choke mainly carries DC current with only a small ripple, so it doesn't buzz or radiate a hum field. You also don't have the problem of the output voltage skyrocketing below the critical current.
Originally Posted by SY 
If there's sufficient interest, I could do this as my next project and document it, once I'm through with the phono stage I'm currently working on, and a speaker pair that I've been itching about. Probably early spring (Northern Hemisphere). Same basic topology, but using a modern approach to controlling voltages and currents, as well as multiple internal feedback loops and circuitry immune to blocking distortion.
I would certainly be interested in Making such an amp.
I have never used mosfets but have made the Williamson direct coupled to the out put tubes using 6BL7 cathode follower. To me seemed to be much more dynamic.
Phil
If there's sufficient interest, I could do this as my next project and document it, once I'm through with the phono stage I'm currently working on, and a speaker pair that I've been itching about. Probably early spring (Northern Hemisphere). Same basic topology, but using a modern approach to controlling voltages and currents, as well as multiple internal feedback loops and circuitry immune to blocking distortion.
I would certainly be interested in Making such an amp.
I have never used mosfets but have made the Williamson direct coupled to the out put tubes using 6BL7 cathode follower. To me seemed to be much more dynamic.
Phil
That's a good tube to use there. A MOSFET will give somewhat lower distortion, though.
- Status
- Not open for further replies.