John Curl's Blowtorch preamplifier

[PMA]

Joshua,

You sure you want to do this? You know what 'we' will say, whatever your opinion, right? Unless you use a well-controlled, statistically significant DBT....

Jan Didden

Jan, would you arrive to perform the one for me? 😀

Yes, we know what you would say, that's why we save a lot of time recently here 😉

 

[jan.didden]

Jan, would you arrive to perform the one for me? 😀

Yes, we know what you would say, that's why we save a lot of time recently here 😉

You mean doing a well-controlled statistically significant DBT for you?? Are you out of your mind?? Reduce me to a nervous wrack, it would! No thank you. I leave that to the professionals 😀

Jan Didden

 

[Joshua_G]

No, not at all. Two things were going on. On was that your results could only be seen as personal preferences. The other was my effort to try to explain some of well-known properties of perception.

Jan Didden

There are other participants in this thread beside you …
My ability to hear differences was already doubted, due to my age.

 

[Joshua_G]

Jan, would you arrive to perform the one for me? 😀

I'd be happy to come over one of these days. May be when the new Dispree PCBs will be ready.

 

[h_a]

Ovidiu and Edmond, I noticed your PGPs do not use a regulated power supply for the output stage. I guess that's due to high PSRR.

Does anybody know the PSRR (roughly) of a simple source follower?

Sure I can simulate it, but real experiences would be great.

Have fun, Hannes

 

[dimitri]

TPA6120A2. 1.4 ppm distortions, 120dB dynamic range, 120dB S/N ratio, 1300V/uS slew rate, 80mW in 600ohm load.

Have a look how low-frequency distortion depends on the load. Do you like this thermal modulation?

 

[PMA]

I'd be happy to come over one of these days. May be when the new Dispree PCBs will be ready.

You're serious?

 

[Edmond Stuart]

PSRR

Hi Hannes,

That's right, the PGP has a very high PSRR. Simulated > 130dB or so. I don't know the exact figure, as it still sits in my crashed HDD (and I'm to lazy to recover all the data).

How about your MOSFET amp and the FQAxxxx's?

Regards,
Edmond.

 

[syn08]

Have a look how low-frequency distortion depends on the load. Do you like this thermal modulation?

Man, it's really scary 😀. It's shifting from 0.6ppm (10Kohm load) to 40ppm (32ohm load). You do it better with discretes 😀.

But have a closer look at the thermal discussion (ThermalPad, etc...) in the datasheet. If this is really an issue for you (it's certainly not for me), you can cool this baby down to Kingdom Come.

 

[john curl]

Goshua_G, both IC's look interesting. Perhaps they are too complex, internally, to be optimum, but at least they both have a very high open loop bandwidth. That is the secret, (denied by Scott Wurcer) to audio quality. Barrie Gilbert would most likely agree with me.

 

[PMA]

.... in extremely highly averaged measurement (TPA).

 

[syn08]

Ovidiu and Edmond, I noticed your PGPs do not use a regulated power supply for the output stage. I guess that's due to high PSRR.

Does anybody know the PSRR (roughly) of a simple source follower?

Sure I can simulate it, but real experiences would be great.

Have fun, Hannes

The PGP amp has a measured end to end PSRR of about 60dB @ 20KHz, the output stage being mostly responsible for. At low frequency, I was unable to measure but it's most likely over 100dB.

An emitter follower has (in a very simplified model) a PSRR=Uo/(Va-Uo) where Uo is the output voltage and Va is the transistor Early voltage.

 

[scott wurcer]

Barrie Gilbert would most likely agree with me.

Not a chance, ask Charles Hansen.😀

 

[john curl]

I won't quibble, I will stand behind the equations developed by Barrie Gilbert, after Otala from 15 years before. PIM folks, don't ignore it, at your peril.

 

[gerhard]

Goshua_G, both IC's look interesting. Perhaps they are too complex, internally, to be optimum, but at least they both have a very high open loop bandwidth. That is the secret, (denied by Scott Wurcer) to audio quality.

And if the amps were exactly the same, equally fast, but had 40 dB more DC gain:
That would really make them unsuitable. 100 times lower open loop frequency.
How horrible.

Barrie Gilbert would most likely agree with me.

Until he declares you to be his press officer, I'll believe this when I hear it from himself.

regards, Gerhard

 

[cbdb]

but at least they both have a very high open loop bandwidth. That is the secret, to audio quality.

Does this also apply to power amps etc. ?

I will stand behind the equations developed by Barrie Gilbert

Can anyone steer me to these equations?

 

[JPV]

I won't quibble, I will stand behind the equations developed by Barrie Gilbert, after Otala from 15 years before. PIM folks, don't ignore it, at your peril.

Is it your experience that a given opamp or amp will sound better dual pole compensated than single dominant pole ( providing no overload of course) or it depends or it doesn't matter?
And if it depends, then on what?

Thanks

JPV

 

[KBK]

Even more, as there is an everlasting pressure on reduced power consumption.

Much to our detriment. If one thinks that IC's have no place in audio, then what does that say about surface mount electronics?

It's the old dancing bear routine. It is amazing that the bear dances at all-not how well it dances.

Same goes for surface mount components. Proximity and tempco vs area (surface and volumetric) wreak havoc in this highly dynamic application called audio. Sure, surface mount works well for digital devices. But they tend to do great disservice to linear analog audio. This does not say that there are not great surface mount designs and apps but if all things where equal (which they almost never are) a properly made discrete equivalent will outperform the surface mount version. Subtle to some, but not to others.

 

[jan.didden]

Much to our detriment. If one thinks that IC's have no place in audio, then what does that say about surface mount electronics?

It's the old dancing bear routine. It is amazing that the bear dances at all-not how well it dances.

Same goes for surface mount components. Proximity and tempco vs area (surface and volumetric) wreak havoc in this highly dynamic application called audio. Sure, surface mount works well for digital devices. But they tend to do great disservice to linear analog audio. This does not say that there are not great surface mount designs and apps but if all things where equal (which they almost never are) a properly made discrete equivalent will outperform the surface mount version. Subtle to some, but not to others.

There is a whole thread somewhere on using surface mount in chipamps in order to make the fb path as short as possible. Because it sounds so much better. As you say, subtle to some 😉

Jan Didden

 

[KBK]

Everything has it's applications, yes.

For example, I have a CD player that utilities those discrete IC amps from..from..I can't remember. C+C or something like that.

Oddly enough, the sound they produce most resembles the sonic presentation of the OPA627 with a resistive Class 'A' mod.

BTW JAN, I got rid of the whole digital crossover thing.

I gave it a shot.

2 highly modded DCX/2496 and two highly modded DEQX units later, I've given up on giving Digital crossovers a chance to be high fidelity. 🙂

They just don't even come close to doing it right. That's about $10k+ of my money and not counting the time and effort to re-execute either type. I gave them my all..and they just don't cut it.

I might try the 192khz thing, but it likely does still not 'make the grade'. Like the whole digital crossover subject itself, I'll not pre-judge it, but I'm not going into the situation ignorant of my personal conclusions on prior art and works, either.