180W folded cascode (just fooling around)

[PanzerLord]

I found a curcuit from one of my text books (Sedra/Smith :Microelectronic Circuits 4th ed.), and I just started to fool around. After some modifications I ended up with a 180W power amp with MOSFET outputs.
It probably contain some flaws and I thought some of you on this forum could come up with some ideas to make it better or if this won't work at all.

The sim says it will work fine and with good S/N and THD numbers, but I won't trust that....

The curcuit shows good stability and great bandwith. The only thing I don't really trust is the emitter follower. I calculated a power dissipation of about 0,25 watts, but the MOSFETS will draw some current at higher frequencies (input cap= 1275pF per MOSFET)...would a 1 watt BJT be sufficient? It should be...but then, that tepends again on the input signal frequency....

(Remember, my electronics skills are a bit rusty, be nice on me. I think no question can be stupid enough...)

And yes, degeneration resistors must be added for practical use...

[PanzerLord]

Here's the ac analysis:

[PanzerLord]

And the fourier-analysis (integrated noise V^2, 20-20kHz = 0,423p)

[darkfenriz]

Providing nicely reguated supply and well matched transistors it can be very nice. Topologically it is creative and pretty 'balanced', in practice you'll probably end up with some degeneration resistors for wilson mirrors (say 33R) and for output mosfets (say 0R22). To simplify Q16/22 don't do much, I'd throw away. And temp. sensing V souce.
cheers

[PanzerLord]

Well, Q16 is part of the (modified) wilson current mirror, but I don't think we'll loose any performance by dropping this one (at least in discrete design).
Dropping Q22 sounds good as well.

Did as told and the circuit is still going strong....

[PanzerLord]

Stupidity shows no limits, Q12,Q13,Q21 have to be replaced. I exceeded the Vce max for these BJT's with ca 50 volts. (The rails where +/-35 volts to begin with)

Maybe the MJE240 will work as a replacement, but I`m a bit sceptic regarding the input impedance...I also found the KSC2628 (Fairchild semiconductors) wich looks even more suitable

I`ll have to do som math and sims when I come home from work...

[Ultima Thule]

Hi Panzerlord,

how did it go with your design, did you ever built it and how did it sound?

Cheers Michael

[PanzerLord]

See my signature...:-). Hopefully soon though, I will probably purchase some output transistors and equipment for making PCB`s in a month or two. I`ve just finished building a power supply for testing my designs (+/- 35 volts, 500VA).

The problem is that I have 3 designs I would like to build, but I haven`t decided wich one to build first...probably this one
Non (global) feedback amp/hawksford I`m a but curious about why so many people favor designs without global feedback...plus I get full balancing...if it works...

In the mean time I`ll stick to my sim...better than nothing, but misleading in a lot of cases for sure..

[Ultima Thule]

Hi again PanzerLord (whats your name btw?),

it's not to often you have been here around for a while, nice to see you back with lot of new ideas (as in your newest thread)!
It looks interesting, I would also like to try something like that, or att least have feedback from VAS.

I recognize myself in your words, having too many ideas I don't know which project I should start!

One idea I have is to use output transistors mounted on a motherboard while having input stage and VAS stage on separate daughterboards one can switch between to evaluate.

Cheers Michael

[CBS240]

Originally posted by PanzerLord
Stupidity shows no limits, Q12,Q13,Q21 have to be replaced. I exceeded the Vce max for these BJT's with ca 50 volts. (The rails where +/-35 volts to begin with)

Maybe the MJE240 will work as a replacement, but I`m a bit sceptic regarding the input impedance...I also found the KSC2628 (Fairchild semiconductors) wich looks even more suitable

I`ll have to do som math and sims when I come home from work...

Have you looked at KSC2784 and KSA1174 from Fairchild for use as small signal instead of 2N5551, 2N5401? High linear gain but definitely small signal. (and 120V Vce low noise)