Sziklai buffer circuit: anyone seen this before?

rjm

Member
Paid Member
2004-05-02 2:41 pm
Kyoto
phonoclone.com
OK, I posted this on the blog, but just thought I'd throw this out here for your inspection.

I searched, couldn't come up with this particular implementation either here or on Google images.

I like it. I mean, throw in a few more resistors over the basic diamond buffer and you can freely control the output stage bias current. Simple and convenient.

Now, there are residual questions regarding offsets, stability, thermal runaway, and so forth, but the FFT at least sims out good.
 

Attachments

  • buffer 03f.png
    buffer 03f.png
    58.7 KB · Views: 600
Last edited:

dadod

Member
2006-04-18 3:11 pm
Zagreb
OK, I posted this on the blog, but just thought I'd throw this out here for your inspection.

I searched, couldn't come up with this particular implementation either here or on Google images.

I like it. I mean, throw in a few more resistors over the basic diamond buffer and you can freely control the output stage bias current. Simple and convenient.

Now, there are residual questions regarding offsets, stability, thermal runaway, and so forth, but the FFT at least sims out good.

Look the output buffer here, not the same but similar. http://www.diyaudio.com/forums/solid-state/235695-no-nfb-line-amp-gainwire-mk2.html
 
OK, I posted this on the blog, but just thought I'd throw this out here for your inspection.

I searched, couldn't come up with this particular implementation either here or on Google images.

I like it. I mean, throw in a few more resistors over the basic diamond buffer and you can freely control the output stage bias current. Simple and convenient.

Now, there are residual questions regarding offsets, stability, thermal runaway, and so forth, but the FFT at least sims out good.

Seen it before, though as a long wave, QRP rig. Works just fine as an RF amp, though in that case, it works into tuned circuits and/or BPFs to eliminate unwanted harmonics. Might require some bias adjustment to get it more towards Class A operation for decent sonic performance.
 

rjm

Member
Paid Member
2004-05-02 2:41 pm
Kyoto
phonoclone.com
R5 and R6 might best be located between collectors and load.
You might also then spread the offsets a little, let feedback to
emitters Q1 and Q3 control the output stage currents.

(those emitters would be Q3 Q4 in my drawing. Some reason
I named my resistors like yours, but forgot to name the Q's
same like your drawing. I was asleep I guess???)

If you were really after unity gain, diodes could get you there...
There would be 2 diode drops between collectors, therefore R8
plus R7 would control the sum of output stage currents in linear
class A.

I'm not so sure this circuit is self limiting if one rail is powered
and the other for whatever reason refuses to come up (or down).
 

Attachments

  • knowyourcurrents.gif
    knowyourcurrents.gif
    20.3 KB · Views: 414
Last edited:

rjm

Member
Paid Member
2004-05-02 2:41 pm
Kyoto
phonoclone.com
It doesn't matter that its not quite unity gain. It ends up close enough.

There may even be a usage case for higher gain, but the performance seems to degrade quickly for these configurations.

Collector vs. emitter resistors? That's something to consider. Intuitively placing them on the emitter seem like more effective "brakes" ...
 
The emitter of the output transistor is in series with the collector
of the input transistor, making less effective than first intuition.
If not for R3 and R4 reducing the impedance of your Q1 and Q3
collectors, the emitter resistors R5 and R6 would have no effect
whatsoever...

Which better blocks current? Q1 collector in parallel with 1K. Or
the 4.7 ohm emitter resistor? If they should disagree about the
base-emitter current for Q4, which impedance dominates?

For this pair, the output collector (in parallel with input emitter)
is the node more sensitive for this kind of degenerative feedback.
Locate that resistor where the 1st emitter can "see" the output
stage current as a degenerative feedback, and then no problem.

Intuitive or not, whatever scheme best reduces the base-emitter
current in the event of excessive collector-emitter current must
be the better brake.
 
Last edited: