John Curl's Blowtorch preamplifier

[PMA]

High resistor values

The high resistor value is an interesting point to me.

In my experience, high resistor value tend to bring problems. When I compare 10k pot to 100k pot, the 10k pot sounds better. When I compare 10k/100R to 100k/1k divider, the 10k/100R sounds better. When I compare low impedance output to medium impedance, the low sounds better.

Why? I guess the answer is easy. Low resistor values are less influenced by parasitic capacitances, stray capacitances, and are inherently less noisy. I do not see any reason to keep high resistor values, unless it is a cooperation with low output current, thus high distortion tube gear.

 

[john curl]

I agree, BUT this resistor is the load for the open loop fet follower. They won't use V parts, as they are more expensive for them. Only Bl (at best) It is a real tradeoff.
I was hit below the belt in 'The Audio Critic' about 10 years ago for using this buffer, so I had to take steps to reduce the loading.
Just for fun, I should try to find what we used to use, before the fet buffer.

 

[PMA]

I understand the buffer issue, and I see it is wise to use it.

My objection is to input pot values (not JC-1 now), I never use higher volume pot than 10k. I guess that 100k pots are for the reason not to load previous circuits.

 

[Wavebourn]

Re: High resistor values

thus high distortion tube gear.

What???

 

[KBK]

Re: High resistor values

The high resistor value is an interesting point to me.

In my experience, high resistor value tend to bring problems. When I compare 10k pot to 100k pot, the 10k pot sounds better. When I compare 10k/100R to 100k/1k divider, the 10k/100R sounds better. When I compare low impedance output to medium impedance, the low sounds better.

Why? I guess the answer is easy. Low resistor values are less influenced by parasitic capacitances, stray capacitances, and are inherently less noisy. I do not see any reason to keep high resistor values, unless it is a cooperation with low output current, thus high distortion tube gear.

High current tends to reject external and internal interference strictly due to drive mass. The backlash tends to be loss of the finest and more rarefied levels of sonic detail.

 

[Wavebourn]

No. Low distortion solid state gear likes to be driven by 50 Ohm signal generators in order to shine, otherwise variations of it's complex input impedance damage the perfect picture....

 

[KBK]

If you are talking video, it's a carrier system. The application is different. Audio signals have to bridge some difficult gaps. Or more correctly, greater complexity in the modeling of the interference considerations which affect the desired signal propogation-compared to that of any energized carrier system.

 

[Edmond Stuart]

Syn08, the 3k83 sees low output impedance of the input jfet buffer.

If you throw SW1, it sees even zero impedance. So, what's your point?

 

[PMA]

No. Low distortion solid state gear likes to be driven by 50 Ohm signal generators in order to shine, otherwise variations of it's complex input impedance damage the perfect picture....

Correct. As low output impedance of driver circuit as possible.

 

[PMA]

If you throw SW1, it sees even zero impedance. So, what's your point?

My point is that is a nonsense to calculate noise of the 100k resistor.

 

[jacco vermeulen]

yep, why not skip the balanced stuff alltogether.

 

[john curl]

I use it, but then, I have a CTC blowtorch preamp to drive it.

 

[Edmond Stuart]

nonsense?

My point is that is a nonsense to calculate noise of the 100k resistor.

100k resistor? Weren't we talking about that 3k38 resistor?

Anyhow, Ovidiu's remarks are far from 'nonsense'. The noise contribution referred to the input is mainly determined by R40 (3k38) or referred to the output by R40 and R44 (100k). Whether the other side of R40 (pin 2) "sees low output impedance of the input jfet buffer" or seeing an even lower impedance by flipping SW1, doesn't matter at all.

 

[john curl]

Much noise about nothing.

 

[syn08]

Much noise about nothing.

Perhaps, but I was/still am curious:

a) How did you spec'd the JC-1 noise numbers.

b) If (quote) "THIS IS A POWER AMP. Not a phono stage, or even a line stage." which I heartfully agree, then why using those ultra low noise expensive and rare 2SK389/2SJ109 duals? Wouldn't anything cheaper/readily available as the 2N5458/2N5461 do? If not, why? You could even find dual (not matched) JFETs as the MMBF4416/MMBFJ177 pair from OnSemi, to avoid thermal drifts (although having a servo is probably good anyway).

 

[john curl]

It used to be cheap and easy to get Toshiba parts. Why bother with less? We have used them 'everywhere' for the last 15 years.

 

[Wavebourn]

Correct. As low output impedance of driver circuit as possible.

Now, we need higher input resistance to minimize the role of their highly non-linear and frequency dependent output resistance.

 

[Wavebourn]

It used to be cheap and easy to get Toshiba parts. Why bother with less? We have used them 'everywhere' for the last 15 years.

I love them too, especially for condenser microphones.

 

[PMA]

Now, we need higher input resistance to minimize the role of their highly non-linear and frequency dependent output resistance.

Not an issue.

 

[john curl]

For the record, the input stage shown was an early edition and NOT complete. In fact, the fundamental limiting noise source is not even listed on the schematic. As time goes on, and IF we still make JC-1's in the far future, we will have to change the input devices to something else, hopefully nothing too compromised.
Of course, when I was first working on this design configuration more than 35 years ago, I used Motorola 5458/5461 type devices, as they were the only ones available in 1970-2.