Darlington or Sziklai (compound) pair output stages?

GregH2 · 30th August 2011, 10:55 AM

Hi all,

Which do you prefer and why? On paper the sziklai pair is superior - lower distortion, fewer high order harmonics and better temperature stability. However, in real life the driver transistor of the sziklai pair can oscillate on switch off requiring compensation capacitors and perhaps degrading sonics.

Theory aside, what do you prefer in a practical circuit and why?

Look forward to your thoughts!

kenpeter · 12:50 PM

Its easy enough to make a non-switching output stage that always maintains
a hundred or so milliamps in bypass of the load. CFP would never need to see
the "driver switch off" condition you so worried about.

I prefer Aleph (BJT MOSFET) or Allison (BJT BJT) super pair for my outputs.
1st BJT turned backwards (emitter coupled) as a depletion mode vreference.

GregH2 · 30th August 2011, 01:02 PM

Thanks Ken,

Correct me if I'm wrong, but aren't you referring to class a output stages? I am talking about their use in class ab outputs. Guess i wasn't clear. Please explain your concept further if I have misinterpreted you as I woud like to understand.

Cheers,

Greg.

kenpeter · 30th August 2011, 02:10 PM

I could easily make these curves cross a lot lower. if you need to
see them squashed flatter against the 200mA reserve current...

Is it AB or Aa? Hard to tell from this graph of transistor currents.
But if you were to look at my Schottky diodes, you then see AB!
Only diodes without recovery probs turn off, transistors never...

AB like this would not bother your siziklai pairs...

Bigun · 30th August 2011, 02:57 PM

You should read more from Doug Self on CFP output for some of the subtleties. For example, although thermal runaway is less of an issue because the drivers have firm control over the outputs, the optimal Class AB bias current (temperature dependent) needs to be controlled much more accurately than regular EF output, for minimum distortion.

And I think Doug found that low signal level distortion for CFP was worse than for EF even though large signal linearity favours the CFP. I believe the reason is that the change in gain around cross-over is more abrupt than EF.

So my take on it is use CFP for Class A only, such as the drivers to an EF stage like the Roender amplifier - or you need some kind of adaptive bias as Kenpeter shows.

AndrewT · 30th August 2011, 03:03 PM

Hi BG,
I think your interpretation of D.Self is correct.
It seems to fit with the general consensus we hear around here.

ingenieus · 30th August 2011, 03:08 PM

CFP outputs are trickier as far the stability is concerned.They are more prone to oscillation when driving capacitive loads. With careful design of compensation circuitry this can be overcome.

tiefbassuebertr · 05:06 PM

to the same topic also this thread exists:
Why collector/drain output stages are so rare???
by my post #59 are additional URLs for more informations.
Go also to
Abacus rieder 60-120 amplifier

GregH2 · 31st August 2011, 12:06 AM

Ah thanks all,

That thread makes good reading.

GregH2 · 31st August 2011, 05:46 AM

Does anyone have any thoughts on which topology would be preferable for highly biased class AB, say 0.5A of bias?

30th August 2011, 10:55 AM	#1
GregH2 diyAudio Member Join Date: Oct 2009	Darlington or Sziklai (compound) pair output stages? Hi all, Which do you prefer and why? On paper the sziklai pair is superior - lower distortion, fewer high order harmonics and better temperature stability. However, in real life the driver transistor of the sziklai pair can oscillate on switch off requiring compensation capacitors and perhaps degrading sonics. Theory aside, what do you prefer in a practical circuit and why? Look forward to your thoughts!

30th August 2011, 12:46 PM	#2
kenpeter diyAudio Member Join Date: Nov 2007 Location: Dallas	Its easy enough to make a non-switching output stage that always maintains a hundred or so milliamps in bypass of the load. CFP would never need to see the "driver switch off" condition you so worried about. I prefer Aleph (BJT MOSFET) or Allison (BJT BJT) super pair for my outputs. 1st BJT turned backwards (emitter coupled) as a depletion mode vreference. Last edited by kenpeter; 30th August 2011 at 12:50 PM.

30th August 2011, 05:03 PM	#8
tiefbassuebertr diyAudio Member Join Date: Jul 2005 Location: D-55629 Schwarzerden	to the same topic also this thread exists: Why collector/drain output stages are so rare??? by my post #59 are additional URLs for more informations. Go also to Abacus rieder 60-120 amplifier Last edited by tiefbassuebertr; 30th August 2011 at 05:06 PM.

Thread Tools	Search this Thread
Show Printable Version Email this Page	Search this Thread: Advanced Search

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
Germanium and silicium transistor in darlington ro sziklai	Captain archer	Solid State	28	5th April 2010 02:55 PM
Sziklai pair as Vbe Multiplier?	sam9	Solid State	12	15th September 2009 02:11 PM
Sziklai pair - 1fet+1bjt.	beppe61	Solid State	27	27th February 2006 08:26 AM
sziklai-darlington diamond buffer (SDDB) for power output stages	mtlin12	Solid State	5	1st April 2005 06:55 AM


	Please consider donating to help us continue to serve you. Ads on/off / Custom Title / More PMs / More album space / Advanced printing & mass image saving

30th August 2011, 01:02 PM	#3
GregH2 diyAudio Member Join Date: Oct 2009	Thanks Ken, Correct me if I'm wrong, but aren't you referring to class a output stages? I am talking about their use in class ab outputs. Guess i wasn't clear. Please explain your concept further if I have misinterpreted you as I woud like to understand. Cheers, Greg.

30th August 2011, 02:57 PM	#5
Bigun diyAudio Member Join Date: Jan 2009 Blog Entries: 2	You should read more from Doug Self on CFP output for some of the subtleties. For example, although thermal runaway is less of an issue because the drivers have firm control over the outputs, the optimal Class AB bias current (temperature dependent) needs to be controlled much more accurately than regular EF output, for minimum distortion. And I think Doug found that low signal level distortion for CFP was worse than for EF even though large signal linearity favours the CFP. I believe the reason is that the change in gain around cross-over is more abrupt than EF. So my take on it is use CFP for Class A only, such as the drivers to an EF stage like the Roender amplifier - or you need some kind of adaptive bias as Kenpeter shows. __________________ "The test of the machine is the satisfaction it gives you. There isn't any other test. If the machine produces tranquility it's right. If it disturbs you it's wrong until either the machine or your mind is changed." Robert M Pirsig.

30th August 2011, 03:03 PM	#6
AndrewT diyAudio Member Join Date: Jul 2004 Location: Scottish Borders	Hi BG, I think your interpretation of D.Self is correct. It seems to fit with the general consensus we hear around here.

30th August 2011, 03:08 PM	#7
ingenieus diyAudio Member Join Date: Apr 2011	CFP outputs are trickier as far the stability is concerned.They are more prone to oscillation when driving capacitive loads. With careful design of compensation circuitry this can be overcome.

31st August 2011, 12:06 AM	#9
GregH2 diyAudio Member Join Date: Oct 2009	Ah thanks all, That thread makes good reading.

31st August 2011, 05:46 AM	#10
GregH2 diyAudio Member Join Date: Oct 2009	Does anyone have any thoughts on which topology would be preferable for highly biased class AB, say 0.5A of bias?

New To Site?	Need Help?
Register to Participate Search Privacy Statement Contact Us	Frequently Asked Questions Did you forget your password? Mark Forums Read