Death of Gain Clone

Jump to Latest
#62
Upupa Epops said:
2.32 A for 8 W.... really amazing.... 😀 Why not 5 A, superbiased superclass A quadrat ? 😀
Click to expand...
Yes, I thought about that option. But might be a bit too complicated 😉

This little amp would consume like 60 Watt to produce 8 Watt.
Could it try to match Nelson Pass First Watt amplifiers?

He uses like 100 Watt to produce 10 Watt output.
😎
 
#67
A decade ago I did a SE mosfet based amp using a 6SL7 as voltage amp.

It was two stage, with four mosfets per channel. Power supply was 50V, 3A with 22mF-20mH-22mF. Current amp was a P type mosfet with NPN bipolar in CFP hybrid, and current sink was an LED referenced N type. Dissipation was 150W per channel, power was 28W, with 1dB at 15Hz, 75KHz. Output stage distortion was 0.06%, all H2/3. The tube was probably at around 2-3%, since it was a simple plate loaded triode. Zout was 38 milliohms. Energy wastage was horrific. I still feel guilty when I switch it on.

I called it the Glass Harmony, and the interface between the tube and the output stage consumed 18 months of intense R&D. It remains secret. The Glass Harmony sounds like 100 watts up to its limits; grip is very impressive, even with a composite 2000uF coupling cap. Those who have watched my journey over the last decade say it's still the best amp I've done; I don't know, I think the AKSA is better overall because it's more resolving, greater slam and impact, and much more efficient. But I still use it for comparison with all my tube hybrids these days.

When you want it to sound like a tube, you really have to use a tube. Funny, that.

Cheers,

Hugh
 
#69
Luke,

There's IP in there I don't really want to share in a kit.

Secondly, boards are a complete failure as a business model.

Thirdly, 400 volts are involved which is dangerous in a kit, and potentially litigious.

Fourthly, the four heatsinks required weigh no less than 11 Kgs. This is VERY costly for air freight!

Lastly, who would even consider 1 x 28W with 300W dissipation in a world where efficiency is becoming ever more important?

However, if you have reasons why it should be done, let's hear 'em! Heck, I might be dead wrong!

Cheers,

Hugh
 
#70
lumanauw said:
I don't understand the theoritical process of "Current Dumping", why it needs 4 component bridge (I think this should work without 120pF+4.5uH, but don't know about the stability without these 2 components), but as I see it, the finals (911-922) works like a "switch"... turnon-turnoff depending the drop at the 47ohm
Click to expand...

Read the pdf lionked further up in the thread, it explains things. Intuitively, you can also look at it like this: when the output stage switches off (since it works in class B), there would normally be a 'notch' in the output - but keep in mind that the inductor would tend to keep the output current constant - in order to do so, voltage across it will change, moving the now relatively high impedance coupled emitters node of the current dumper 'acrosss' the class B gap.
 
#71
Hi, Ilimzn,

Read the pdf lionked further up in the thread, it explains things. Intuitively, you can also look at it like this: when the output stage switches off (since it works in class B), there would normally be a 'notch' in the output - but keep in mind that the inductor would tend to keep the output current constant - in order to do so, voltage across it will change, moving the now relatively high impedance coupled emitters node of the current dumper 'acrosss' the class B gap.
Click to expand...

You are right, thanks for the explenation 😀
What about the 120pf, what is the main function of this? For stability or for balancing the 4 bridge leg?
 
#74
lumanauw said:
Hi, Ilimzn,
You are right, thanks for the explenation 😀
What about the 120pf, what is the main function of this? For stability or for balancing the 4 bridge leg?
Click to expand...

It's bridge balancing. Again, intuitively, it prevents the NFB loop seeing the discontinuity due to sudden gm change and not correcting propperly - this is precisely how the bridge nulls out the nonlinearity - in order for this to work, you must have the L and C components.
 
#75
Hi Hugh,

I can understand your reasons, but im being selfish of course🙂

I dont leave my amps on all the time, so an hours listening a couple of times a week is not a big deal to me. And I have started using the train to do m bit🙂
I doubt Electricity consumers are the reason for the polar caps melting anyway.

seeya Arthur
 
#77
so you removed 3.300uF//4.7uF blocking cap in feedback

did anything happen?
with offset or else

I guess not, as we have JFET input pair


Even if had been bipolar input pair
I doubt block cap would do much good or be needed.
Not when feedback resistance is as low as here, 500/27 Ohm.
 
You must log in or register to reply here.
Top Bottom