JamJar: an HPA-1-inspired power amp

[JeffYoung]

As long as there is vegemite, I will be there.

I'll be there as long as we don't have to eat the stuff....

 

[2 picoDumbs]

This is the most hurtful thing I have encountered in my life.

I need access to a safe space to discuss my feelings of being marginalized.


Hahahaha

 

[jam]

Pico.......

 

[JeffYoung]

I've made some preliminary topology decisions to get Pico out of his funk.

I'm going to build two. The cascoded push-pull version will be renamed SixPack and go up against Beefcake.

The current-mirror version gets to keep the JamJar name, but now that it no longer needs to go up against Beefcake I can make some changes to address the Early effect in the VAS. I've dropped the rail voltage to 36V, and swapped 2N5401/2N5551 into the mirrors (which have relatively high Early voltages).

Still noodling on the bias circuit (TL431 vs. simple trimmer) and some other details.

As always, comments welcome.

Cheers,
Jeff.

 

[Mark Johnson]

Jeff, I am curious to know, what do you and LTSPICE estimate for the gain of the first stage in the resistor-load amp "SixPack"? Specifically the AC analysis ratio V(Q7base)/V(ampinput)?

 

[jam]

Bias circuits

I hope pico has recovered from the trauma..............

My choice of bias circuit, for consideration , one of a couple of options but for mosfets a simple Vbe (one transistor) suffices. I am not much of a fan of the TL431, it works but sound wise ........I did not use it in the HPA-1 for a reason.

Jam

 

[Mark Johnson]

MJE340 got her a BIG VERY BIG Early voltage.

_

 

[JeffYoung]

Jeff, I am curious to know, what do you and LTSPICE estimate for the gain of the first stage in the resistor-load amp "SixPack"? Specifically the AC analysis ratio V(Q7base)/V(ampinput)?

More-or-less unity. Should I be looking for some gain there?

 

[jam]

Mark ,

Not my device of choice either, just illustrating the topology. The devices chosen for cascodes should work.

In this case all three of Hawksford's examples should work.

Jam

 

[indra1]

... My choice of bias circuit, for consideration ...

No resistor from B to C on Q15?

 

[jam]

indra,

You are correct, there should be a resistor there, my apologies.

Jam

 

[jam]

...now if you really want to get fancy, and have drivers.

Jam

 

[JeffYoung]

Oh yeah, I remember Hawksford's paper from trying to figure out the NAP250 circuit. JV had the added collector resistor, but its value was about 3X what it should have been according to Hawksford's equation. I never did square that circle....

 

[Itsmee]

...I never did square that circle...

The answer may be in the Hawksford paper posted by jam, the paragraph after equation 8.

 

[jam]

Yes the Naim 250 a good candidate for modification.........

 

[JeffYoung]

The answer may be in the Hawksford paper posted by jam, the paragraph after equation 8.

Could be, although the NAP250 is biased very low. (They have been known to go up in smoke, but that's usually due to instability.)

 

[JeffYoung]

How's this look for SixPack?

 

[jam]

Well....

 

[Mark Johnson]

If the first stage's gain is ~1 then all of the amplifier's gain comes from the second stage. Will that be enough?

Also, the second stage's noise contribution is divided by the first stage's gain. Since the first stage's gain is ~1, the amplifier will be noisier.

 

[2 picoDumbs]

I hope pico has recovered from the trauma..............

I am not much of a fan of the TL431, it works but sound wise ........I did not use it in the HPA-1 for a reason.

Jam

Me either, I’ve been pondering about how to replace it with something more dumb, I’ve thought about green leds or zener reference with a voltage dividing trim pot.

I don’t like vbe, but that doesn’t mean there is a rational reason for the dislike, I just prefer dumb.