JLH 10 Watt class A amplifier

[Nico Ras]

2SK1058 includes gate zener in the package.

[shaan]

I just simulated the following DOZ schematic with the CCS config as shown in the attachment. It passed all the tests and gave lower THD level, higher output swing and higher slew rate than the original DOZ with the same output transistors. So I was wondering how it may sound.

The CCS runs okay in the sim and to me it looks like a better option as it will have less Iq increase with temperature increase. Still, can someone please tell me if there's some problem waiting for me if I use this CCS?

Thanks in advance.

[Nico Ras]

You also retained the bootstrap...

[shaan]

Yup I like it. So we actually have two current sources here.

Simulation shows the effect of the bootstrap source very clearly. Voltage swing is +19 -19. Slew rate +180V/uS -400V/uS.

[Jay]

Aesthetically I prefer a higher (R16+R3). Of course I have seen smaller and bigger value. But I don't do amp design so I don't know if this has an audible effect. But if I do amp design I will make sure that I know.

[shaan]

Jay, It doesn't have an audible effect. But it has an effect on the minimum value of C5, which increases as (R16+R3) decreases, in order to effectively prevent any power supply hum or noise from getting into the input. I used this combination cause I have a lot of 20K pots in stock, in the past I used 100K and that was okay too.

[Jay]

Originally Posted by shaan

Jay, It doesn't have an audible effect. But it has an effect on the minimum value of C5, which increases as (R16+R3) decreases, in order to effectively prevent any power supply hum or noise from getting into the input. I used this combination cause I have a lot of 20K pots in stock, in the past I used 100K and that was okay too.

If this is true then higher value of (R16+R3) is better because you can use smaller value of C5 which means you can choose higher quality cap, or means less dollar. But... I doubt it.

From my limited experience, there is a "sweet spot" for this RC filter. I have temporarily assumed that this is true: you have to aim for a specific RC values (not just minimum R*C) which may means a specific frequency, but it is a DC anyway

[shaan]

With this 20K pot I use 47uF electrolytic. There is absolutely no hum or any other kind of badness in the speaker. With 100K pot I used 33uF and to me that was okay too, with same results as now. However, I smooth my PS with a simple CRC pi filter which I have found to be as silent as a capacitance multiplier, with 2Watts less dissipation than the cap-mult.

[Jay]

Originally Posted by shaan

However, I smooth my PS with a simple CRC pi filter which I have found to be as silent as a capacitance multiplier, with 2Watts less dissipation than the cap-mult.

Have you ever tried to remove the R in the CRC and compare how they "sound"? I have done this many times in other people design (of course) and I found that I prefer the filter without the specified R.

In another occasion I experimented with CLC and found out that there is this "sweet spot", meaning that we cannot just throw an arbitrary inductance and capacitance there. For another limited experience, I experienced the same thing with RC filter. So I guess that explained why I prefer to remove the R from the CRC in those circuits. But may be it was just because the R was too big that it affected the impedance seriously. May be it should be less than 0.1 Ohm?

[shaan]

Originally Posted by Jay

Have you ever tried to remove the R in the CRC and compare how they "sound"? I have done this many times in other people design (of course) and I found that I prefer the filter without the specified R.

Yes I have done it and it resulted in clearly audible hum. I use a 15000uF+1ohm+15000uF CRC filter Only for my Class-A experiments with >1A bias. For my class-AB amplifiers I have never used it as it's just not necessary with an amp with good PSRR and 200mA Iq.

In another occasion I experimented with CLC and found out that there is this "sweet spot", meaning that we cannot just throw an arbitrary inductance and capacitance there. For another limited experience, I experienced the same thing with RC filter. So I guess that explained why I prefer to remove the R from the CRC in those circuits. But may be it was just because the R was too big that it affected the impedance seriously. May be it should be less than 0.1 Ohm?

According to my limited experience and knowledge the second C in the CRC/CLC filter should present a low impedance to the amplifier and also minimize the negative effect the R or L would cause in case of transient production. With my abovementioned simple filter I have not yet noticed any illness in the speed(drumstrokes, steel guitars, cymbals etc.) of my class-A amplifiers, be it a simple MOSFET source follower or ZEN or JLH. I think doubling the second C will be an even better step, than decreasing the value of R. Sweet Spot? I have yet to find one as I haven't spent as much time on the PS as I have on the amplifiers.