I had 0,47 @ one Toshiba 2SK176 and it was great; but I wanted more slam; so selected 2 quads of each - each with their own source resistor of 0,47. It did not image rigcht. The zero Rs is great in imaging. But 'softness in sound' is something else.
What I like very much is Nelsons new invention of a partial decoupling of the top source resistor in his single-Vb version, so both halves start to have a different amplification through less degeneration/more local feedback; and that influences the H2 amount and phase.
Here, with P3 in the F5 v2, the same is done. The amplification changes a bit through different local feedback.
In my original Le Monstre there is also a potmeter, it sets DC but also the local feedback of the input pair; I had 200/20 ohms! [Original selected pairs by the shop; the L'Audiophile design is with fixed Rd.]. So same effect in reducing the amplification in the two halves, top and bottom.
What I like very much is Nelsons new invention of a partial decoupling of the top source resistor in his single-Vb version, so both halves start to have a different amplification through less degeneration/more local feedback; and that influences the H2 amount and phase.
Here, with P3 in the F5 v2, the same is done. The amplification changes a bit through different local feedback.
In my original Le Monstre there is also a potmeter, it sets DC but also the local feedback of the input pair; I had 200/20 ohms! [Original selected pairs by the shop; the L'Audiophile design is with fixed Rd.]. So same effect in reducing the amplification in the two halves, top and bottom.
)
I'll think about it. I have the models too. Somewhere. But models give no phase. Hey wait. Should add a brick filter! generator 1,000 Hz; filter 1,200Hz at >60 dB/octave. I ponder a bit and ask in the appropriate thread.
Agree that distortion in the input probably leads to H4 and higher. The trick is in setting both halves of the virtual long tailed pair at a slightly different position (instead of N-N device, now N-P device) - there can be a specific position.
Agree that distortion in the input probably leads to H4 and higher. The trick is in setting both halves of the virtual long tailed pair at a slightly different position (instead of N-N device, now N-P device) - there can be a specific position.
I once did this in a symmetrical tube preamp, where the output pair of tubes E90CC had common Rk but separate Vg each side with a potmeter, so I could slide along the loadline and I saw the distortion move from positive to negative. That was in the eighties, my generator from that time went bust; I am now at my forth scope and already have destroyed two inputs again. So it gets harder to do new work. I that old tube build it was most beautiful sound I got; but after years I found output of the power amplifier was mostly the same, so I reasoned the preamp signature dominated everything. Stupid me, I have it on a shelve now.
Amount and phase of H2 depend on a few factors:
- gm of each jfet which is directly related to Idss and Vp
- jfet drain resistor
- gm of each mosfet
- Mosfet source resistor
- amount of global feedback
You could use 2 x GST60A24-P1J in series per amp but I would not recommend it.
A well designed linear supply will sound better.
I have played with this - you need 2 of them and have to set them up just right - it works but it's also can be unstable (you have to make sure the power supplies can be setup together to make a dual pole power supply). When I played around with this Idea I also used an SMPS from Connex Electroic and the amp draw caused it to have a low level noise (they wouldn't help because they stated their SMPS were not to be used for class a amplifiers at the time). If you want to experiment you could ask for a custom V version of this SMPS. Just be aware that you may have a small amount of SMPS noise in the system unless you know how to deal with that.
It would be great if someone made an SMPS designed for Class A amplifier experimentation - when I played with this it was almost 10 years ago so maybe something newer and better is out there? In the end your best bet is to build a traditional PS as the cost savings is small for the SMPS vs. having a PS that simply works.
I've been using SMPS in Class A amplifiers for a long time now. The two issues you face are differential turn-on (which are a bit of a pain but not so much in the F5) and average to peak draw from each supply. In a push-pull amp the ratio of standing current to peak current can be very high, and you tend to get some dynamic compression as the SMPSs try to keep up with the increase in demand due to the lowish switching frequency. Problem is made worse by the fact that in stereo music, usually both channels have the same demands at the same time.
One trick used in Class D amplifiers to solve this issue is to invert the phase of one channel and also invert the loudspeaker terminal polarity, thus maintaining absolute phase, at the same time balancing the PSU load by loading the opposite supply. This prevents bus pumping, which is a bit of a nuisance. Push-pull amps do not draw a constant bias from two disparate supplies (though they do do so from a single bipolar supply). That said, at low output (or with massively oversized supplies) you probably will be OK. I run a small 'Juma-style' F5 with a couple 24V SMPS and it runs great, about 8 years on.
SE amps OTOH have no issues with SMPS. One design we have runs on SMPS, scales from 12V to 45V rails for 15-80W output, and has a peak to average ratio of just 4:1, which is something an SMPS can do without an issue, even without the phase inversion technique. The ACA is another design, that runs officially on SMPS.
Clean up of SMPS switching noise is mandatory when using them. Our preferred route is a cap multiplier, but even a simple LC filter works well. Even some added capacitance on the output can be of great help, but the inductance is to make sure the SMPS does not suffer internal resonances when connected directly to large capacitors.
One trick used in Class D amplifiers to solve this issue is to invert the phase of one channel and also invert the loudspeaker terminal polarity, thus maintaining absolute phase, at the same time balancing the PSU load by loading the opposite supply. This prevents bus pumping, which is a bit of a nuisance. Push-pull amps do not draw a constant bias from two disparate supplies (though they do do so from a single bipolar supply). That said, at low output (or with massively oversized supplies) you probably will be OK. I run a small 'Juma-style' F5 with a couple 24V SMPS and it runs great, about 8 years on.
SE amps OTOH have no issues with SMPS. One design we have runs on SMPS, scales from 12V to 45V rails for 15-80W output, and has a peak to average ratio of just 4:1, which is something an SMPS can do without an issue, even without the phase inversion technique. The ACA is another design, that runs officially on SMPS.
Clean up of SMPS switching noise is mandatory when using them. Our preferred route is a cap multiplier, but even a simple LC filter works well. Even some added capacitance on the output can be of great help, but the inductance is to make sure the SMPS does not suffer internal resonances when connected directly to large capacitors.
ACA Mini aka Amp Camp Amp Mini is basically a mini F5 run on a single supply with input and output caps.
https://www.diyaudio.com/community/threads/diy-aca-mini.379037/
Could you tell me which one you went with and if yoy needed to add any aditional filtering?
This is the one designed for class a use:
https://micro-audio.com/store/product/smps500-cla/
I have one in my F5, plus one PCB of extra filter caps. It’s worked flawlessly.
If you contact Sami through his website, he’ll make sure you get what you need.
https://micro-audio.com/store/product/smps500-cla/
I have one in my F5, plus one PCB of extra filter caps. It’s worked flawlessly.
If you contact Sami through his website, he’ll make sure you get what you need.
