John Curl's Blowtorch preamplifier part II

[simon7000]

Originally Posted by Esperado

Adding two paralleled diodes upside down between the center transformer wire and ground ?

Correct! Now who did it first?

[simon7000]

Originally Posted by simon7000

Pavel,

Would you care to start the list of all the advantages?

ES

Originally Posted by sofaspud

It was unclear to me also, but is now making sense. I'd be interested if there is more to the story.

Originally Posted by SY

You'll want to consider primary and secondary common and differential mode filtering, interwinding shielding, and snubbing (especially the secondaries). A broad-based approach to making quiet supplies is more effective than just focusing on one or two aspects and trying to rhodium plate them.

SY,

Do you care to elaborate on your additions to the list?

I am not big fan of inter-winding shielding as it does increase line coupling to the shield which creates the problem of where do you return those currents. I prefer dual or split bobbin designs and isolating the core from the chassis. (See a Citation 12 for an example.)

ES

[Esperado]

Originally Posted by simon7000

Correct! Now who did it first?

J.C. ? He did everything first.

Originally Posted by simon7000

I am not big fan of inter-winding shielding as it does increase line coupling to the shield which creates the problem of where do you return those currents. I prefer dual or split bobbin designs and isolating the core from the chassis. (See a Citation 12 for an example.)

I wonder why R core are so rare and expensive, that is the best solution and easy to manufacture.

[a.wayne]

Borbely .....

[SY]

Originally Posted by simon7000

I am not big fan of inter-winding shielding as it does increase line coupling to the shield which creates the problem of where do you return those currents.

Chassis. It's not a problem, it's a feature. See, for example, how I did this in my phono and line preamp designs which are extremely quiet and don't have issues with line noise pickup at any levels above ridiculous.

There's an excellent and thorough treatment of these issues (as well as the others I mentioned which don't seem to concern John) in "Valve Amplifiers" 4th ed. You'll also see an excellent survey of proper ways to do snubbing in an upcoming Linear Audio (I think the next issue).

Agreed, split bobbin is a great way to go. That's even more important in tube amps, which have some unique ways for noise to creep in.

[simon7000]

Originally Posted by a.wayne

Borbely .....

Borberly showed amazing ability to spot an improvement, redo it with his style and present it. But the first place I ever saw it was.... well I won't feed John's ego!

[simon7000]

Originally Posted by SY

Chassis. It's not a problem, it's a feature. See, for example, how I did this in my phono and line preamp designs which are extremely quiet and don't have issues with line noise pickup at any levels above ridiculous.

There's an excellent and thorough treatment of these issues (as well as the others I mentioned which don't seem to concern John) in "Valve Amplifiers" 4th ed. You'll also see an excellent survey of proper ways to do snubbing in an upcoming Linear Audio (I think the next issue).

Agreed, split bobbin is a great way to go. That's even more important in tube amps, which have some unique ways for noise to creep in.

But SY you know I go for ridiculous levels!

Bipolar junction transistors are much more susceptible to the problem than red hots. The consumer electronics industry is going to double insulated and plastic chassis. So the current design trend is to use different shielding techniques.

Now Rick Miller mentioned using higher voltage diodes as they have less capacitance than lower voltage types. What was not mention is the difference in switch off voltage between the two, attached is a shot showing the difference between a 1N4004 rated at 400 PIV and 1 amp versus a 2.5 amp 1000 volt generic replacement diode.

Very little difference. It is getting harder to find soft recovery diodes and from my measurements commodity diodes are getting much quieter.

So what else do you want to expand on?

[bcarso]

Originally Posted by simon7000

But SY you know I go for ridiculous levels!

Bipolar junction transistors are much more susceptible to the problem than red hots. The consumer electronics industry is going to double insulated and plastic chassis. So the current design trend is to use different shielding techniques.

Now Rick Miller mentioned using higher voltage diodes as they have less capacitance than lower voltage types. What was not mention is the difference in switch off voltage between the two, attached is a shot showing the difference between a 1N4004 rated at 400 PIV and 1 amp versus a 2.5 amp 1000 volt generic replacement diode.

Very little difference. It is getting harder to find soft recovery diodes and from my measurements commodity diodes are getting much quieter.

So what else do you want to expand on?

Although the forward drop is considerably higher, SiC is quite impressive. I wish they made some small ones. Ditto GaN.

Pease once pointed out that many (he may have said "all") silicon schottkies had a "parasitic" P-N junction diode along for the ride, which spoils the fine-structure in reverse recovery.

[PMA]

Originally Posted by SY

I don't think he's ready for his close up, Mr. DeMille.

[SY]

Originally Posted by simon7000

So what else do you want to expand on?

Nothing, really. These are solved problems, and people who want to read about them have plenty of resources; I'm sure your review will be interesting and educational.

Diode switching can be handled with proper snubbing and layout. Having soft recovery makes that easier, and these days, soft recovery diodes are dead cheap. Circuit design and layout for noise rejection is even more important, but doesn't make good stories, apparently.