I appreciate the reply! All too often, I READ a word, but never hear it pronounced. I invariably get it wrong. Words like Piezo, Baucharot, or even the names of people. I read it, but never hear it pronounced, sometimes, with embarrassing results. Fine for writing and posting articles, but not so good for oral presentation! Phonics DOESN'T work for me!
A good video, thanks. You do a good job of covering the practical application and discussion of the merits alongside the theory. It's all very helpful to a complete novice.
I'm really enjoying this series of videos, John - covering the theory from quite a practical perspective. Also nice to see the Snickers' contribution to the eternal output topology debate at 17:52...
Great video, as always! Thanks John! As a musician, learning about audio at such a 'raw' level is so interesting, and building my own hardware is so satisfying. Thank you for sharing your knowledge and allowing me to learn about all these amazing things!
You are awesome John... Greetings from India. I love all your videos... You are very resourceful.... I love your kitty too... I have had kitties all through my childhood... I miss them now!
Great stuff ! ... I think you could measure output distortion with only an Opamp to compare input to output. If the output did not track exactly with the input, the opamp would show the difference.
Back in the day, they would often add a diode in series with the base of the CFP, so the voltage from the rail would be the same as the complementary "Darlington" pair. Once complimentary pairs of power transistors came available, I went with the CFP design, because it didn't add the negative temperature coefficiency of the drive transistor to the power transistor.That means only one compensation diode, instead of two, to give overall temperature stability. The diodes needed to be thermally coupled as close as possible, to the power transistors.(the emitter follower resistors are there to pick up the difference)
A rather old integrated Kenwood KA-300 in my collection has the Quasi Complementary output stage. The schematic calls it "SEPP" (single ended push-pull), which may sound rather confusing, as SE and PP are mentioned together. One of the advantages of this circuit is its simplicity, it needs fewer components to deliver a decent sound than the Darlington setup. BTW, not only does the CFP not need a sensor on the heat sink for thermal stability, on the contrary, a thermal sensor on the heat sink would be a bad idea.
The term SE is kind of confusing as in the tube amp world, it refers to a single output device rather than a push pull circuit. In solid state, SE seems to refer to an output stage that has one output connection and the other terminal is circuit ground as opposed to a bridged configuration.
You still need some compensation for the negative temperature coefficiency of the DRIVE transistor, not the output transistor. That would allow you to put the compensation diode on the PC board,(in close proximity to the driver transistors) unless, the driver transistors are also mounted to the same heatsink as the power transistors. In that case, the diodes do need to be mounted on the heatsink. If the drive transistors do need to be mounted to a heatsink, that's where to put the compensation diodes.
There is another advantage with the CFP: you can get more voltage swing at the output and therefore more power. This yields especially for lower operating voltages. With Darlington configuration I can get 10 W at +/- 12 V rails and an impedance of 4 Ohm. With Sziklai configuration I get 13 W. At higher operating voltages this advantage vanishes more and more.
It would be interesting to do some speed comparisons between standard Darlington and the CFP configuration. The Onkyo A-9150 is designed to achieve 500V/us slew rates, and it uses CFP in the driver stage.
May be getting out of the "audio" subject, but I've been playing with LED flashers, and if you put a 50K or so resistor from + 12 Volts to a 2.2 uF capacitor to ground, then put the Emitter of a S9018 NPN Transistor to the R/C junction and then an LED from the collector to circuit ground, The LED will flash. (Base of transistor is N/C) Changing R/C changes frequency. Almost any NPN Transistor will do this, But a PNP will not. I use this to pulse the power to the LEDs and save power. The NPN is "backwards" but flashes when it's breakdown voltage is met. (I think) instead of the LED you can drive another transistor for even more power. Why a PNP won't work has me confused. Different transistors must have different voltages in order to "flash".
Sir I am a subscriber of your channel. Well in your class d amp reviews the pam8403 inspired me most. But wen I bought the amp I realized that it had a lot of bass. And its a low power amp so giving output filters will decrease the sound output so sir is there any way to decrease the bass from the input? And also I found that 104 pf can decrease some bass but not that much.
The amplifier is of course linear across the frequency range, but as you found it's sometimes not desired :) Due to 1/f energy distribution, cutting off a little bit off the low end where your speakers aren't audible anyway, can give you a good bit of leeway on gain. Perform input filtering rather than output filtering. Most economical way do to this is to look at the existing input AC coupling capacitor and reuse that and connecting a resistor across the input. Look up "RC high-pass filter", you'll find how to calculate the filter. The AC coupling capacitor forms effectively such a low-pass filter together with IC's input impedance as resistance, but since IC's input impedance is very high, you want to effectively lower it.
"Tanks" again!! Will try this out in some of my devices. Will, sooner or later, make a video of these things and other stuff. Am an empirical experimenter who does not calculate, not math.....just try it and fry it, as of late don't burn up many components. Must keep the smoke inside of them.
Same to me.im an inventor.and successfully developing my random automatik network switches.which one switch will cancel the previously switches which took me 15 years to really successful.my concept is " do and try first,calculation.. later".
Try the BD140 - 16, 2SA1358, or a ZTX753 (lower power). Older/oldest pnp driver transistors are the BC 161, BC461, BCX29 which are available in the metal TO5, TO39 cases which can make attachment to a common heatsink difficult (the collector is connected to the case with these transistors).
sir have a good day.im.just asking what is the best value of coupling capacitor or input capacitor for.preamp? what is the best coupling capacitor electrolytic capacitor? or propylene capacitor?tnx.sir
@@JohnAudioTech my input value capacitor of my preamp with tone control is 2.2uf electrolytic capacitor and the output is 1uf electrolytic caps.is that ok.sir the values of the capacitor? and can.i change polypropylene caps?
You did not mention that CFP gives you an advantage in signal swing: it subjects your signal to just one diode voltage drop instead of two in the darlington pair. Also, you did not mention what the stability conditions are for sziklai vs darlington pair. Overall, nice video.
I can't find an image anywhere but on page 75 of Small Signal Audio Design by Douglas Self there is good CFP unity-gain buffer. I am making an active crossover with that buffer.
In the second circuit the emitter of Q2 is supplying 0.51 Amps so you cannot get more than that from its collector!!! what is 5100? It looks to me that the formula is incorrect
Note that the units aren't matched in the diagram; the base of Q1 is listed as ".1" but that is .1 mA, so .0001 amps. That's why it is 5100. He does clearly say "a hundred microamps" when he writes the .1.
The correct pronunciation was added by a Hungarian before you made your comment. Since we know you didn't even know, maybe in general with pronunciation you'd do better to use the phrasing "I pronounce it as" or "I've heard it pronounced as" instead of "it is."
Pronunciation of Sziklai: seek-laa-ee
Greets from Hungary!
Thanks! I was hoping someone from Hungary would chime in.
Well, I am glad I am close. I have always said 'seek-lie'
I appreciate the reply! All too often, I READ a word, but never hear it pronounced. I invariably get it wrong. Words like Piezo, Baucharot, or even the names of people. I read it, but never hear it pronounced, sometimes, with embarrassing results. Fine for writing and posting articles, but not so good for oral presentation! Phonics DOESN'T work for me!
A good video, thanks. You do a good job of covering the practical application and discussion of the merits alongside the theory. It's all very helpful to a complete novice.
I'm really enjoying this series of videos, John - covering the theory from quite a practical perspective. Also nice to see the Snickers' contribution to the eternal output topology debate at 17:52...
Great video, as always! Thanks John!
As a musician, learning about audio at such a 'raw' level is so interesting, and building my own hardware is so satisfying.
Thank you for sharing your knowledge and allowing me to learn about all these amazing things!
Glad you enjoyed it! Musicians are amazing to me. They create all the wonderful music I listen to through these electronic components.
You are awesome John... Greetings from India. I love all your videos... You are very resourceful.... I love your kitty too... I have had kitties all through my childhood... I miss them now!
Great video, as always! Thanks John!
Just a great video; you are so clear, thank you.
My feelings exactly.
Amazing explaination. Thumbnail is quite innovative.
Great stuff ! ... I think you could measure output distortion with only an Opamp to compare input to output. If the output did not track exactly with the input, the opamp would show the difference.
Interesting that people differ over this....cheers.
Back in the day, they would often add a diode in series with the base of the CFP, so the voltage from the rail would be the same as the complementary "Darlington" pair. Once complimentary pairs of power transistors came available, I went with the CFP design, because it didn't add the negative temperature coefficiency of the drive transistor to the power transistor.That means only one compensation diode, instead of two, to give overall temperature stability. The diodes needed to be thermally coupled as close as possible, to the power transistors.(the emitter follower resistors are there to pick up the difference)
A rather old integrated Kenwood KA-300 in my collection has the Quasi Complementary output stage. The schematic calls it "SEPP" (single ended push-pull), which may sound rather confusing, as SE and PP are mentioned together. One of the advantages of this circuit is its simplicity, it needs fewer components to deliver a decent sound than the Darlington setup. BTW, not only does the CFP not need a sensor on the heat sink for thermal stability, on the contrary, a thermal sensor on the heat sink would be a bad idea.
The term SE is kind of confusing as in the tube amp world, it refers to a single output device rather than a push pull circuit. In solid state, SE seems to refer to an output stage that has one output connection and the other terminal is circuit ground as opposed to a bridged configuration.
You still need some compensation for the negative temperature coefficiency of the DRIVE transistor, not the output transistor. That would allow you to put the compensation diode on the PC board,(in close proximity to the driver transistors) unless, the driver transistors are also mounted to the same heatsink as the power transistors. In that case, the diodes do need to be mounted on the heatsink. If the drive transistors do need to be mounted to a heatsink, that's where to put the compensation diodes.
There is another advantage with the CFP: you can get more voltage swing at the output and therefore more power. This yields especially for lower operating voltages. With Darlington configuration I can get 10 W at +/- 12 V rails and an impedance of 4 Ohm. With Sziklai configuration I get 13 W. At higher operating voltages this advantage vanishes more and more.
HELLO THERE . THERES A TIP142 NPN TRANSISTOR IM TRYING TO USE IN MY ONE TRANSISTOR SOUND AMP . SO TELL ME HOW THE CIRCUIT WILL LOOK LIKE ??? THANK YOU
Great vid ty.
I am reading a schematic that uses CFP to drive multiple output transistors. Seems like a good idea.
It would be interesting to do some speed comparisons between standard Darlington and the CFP configuration. The Onkyo A-9150 is designed to achieve 500V/us slew rates, and it uses CFP in the driver stage.
What would happen if 3 transistor in this skilai configuration?
Dear John would you please suggest some audio amplifier designs/diagrams to illustrate your discussions points?
May be getting out of the "audio" subject, but I've been playing with LED flashers, and if you put a 50K or so resistor from + 12 Volts to a 2.2 uF capacitor to ground, then put the Emitter of a S9018 NPN Transistor to the R/C junction and then an LED from the collector to circuit ground, The LED will flash. (Base of transistor is N/C) Changing R/C changes frequency. Almost any NPN Transistor will do this, But a PNP will not. I use this to pulse the power to the LEDs and save power. The NPN is "backwards" but flashes when it's breakdown voltage is met. (I think) instead of the LED you can drive another transistor for even more power. Why a PNP won't work has me confused. Different transistors must have different voltages in order to "flash".
Sir I am a subscriber of your channel. Well in your class d amp reviews the pam8403 inspired me most. But wen I bought the amp I realized that it had a lot of bass. And its a low power amp so giving output filters will decrease the sound output so sir is there any way to decrease the bass from the input? And also I found that 104 pf can decrease some bass but not that much.
The amplifier is of course linear across the frequency range, but as you found it's sometimes not desired :) Due to 1/f energy distribution, cutting off a little bit off the low end where your speakers aren't audible anyway, can give you a good bit of leeway on gain. Perform input filtering rather than output filtering. Most economical way do to this is to look at the existing input AC coupling capacitor and reuse that and connecting a resistor across the input. Look up "RC high-pass filter", you'll find how to calculate the filter. The AC coupling capacitor forms effectively such a low-pass filter together with IC's input impedance as resistance, but since IC's input impedance is very high, you want to effectively lower it.
"Tanks" again!! Will try this out in some of my devices. Will, sooner or later, make a video of these things and other stuff.
Am an empirical experimenter who does not calculate, not math.....just try it and fry it, as of late don't burn up many components.
Must keep the smoke inside of them.
Same to me.im an inventor.and successfully developing my random automatik network switches.which one switch will cancel the previously switches which took me 15 years to really successful.my concept is " do and try first,calculation.. later".
Can you recommend a good in production PNP, to build a "PNP" Sziklai with a 2n3055 for a quasi complementary output stage? Thanks.
Try the BD140 - 16, 2SA1358, or a ZTX753 (lower power). Older/oldest pnp driver transistors are the BC 161, BC461, BCX29 which are available in the metal TO5, TO39 cases which can make attachment to a common heatsink difficult (the collector is connected to the case with these transistors).
@@robertneill3057
Thanks. I've also seen SJE210 used, in a stereo receiver, but it's not available anymore.
If you've not already you should do a video on MOSFET output stages.
sir have a good day.im.just asking what is the best value of coupling capacitor or input capacitor for.preamp? what is the best coupling capacitor electrolytic capacitor? or propylene capacitor?tnx.sir
Polypropylene
@@JohnAudioTech wow..thank you sir..what about the best value of capacitor? for input coupling caps polypropylene?
@@cortessarge5399 I can't say without knowing the input impedance of the circuit.
@@JohnAudioTech for my preamp tone control sir?
@@JohnAudioTech my input value capacitor of my preamp with tone control is 2.2uf electrolytic capacitor and the output is 1uf electrolytic caps.is that ok.sir the values of the capacitor? and can.i change polypropylene caps?
good to see a new kitty!
thanks for the great instruction
You did not mention that CFP gives you an advantage in signal swing: it subjects your signal to just one diode voltage drop instead of two in the darlington pair. Also, you did not mention what the stability conditions are for sziklai vs darlington pair. Overall, nice video.
Awesome video. I have a request can u make a video on high power amplifier say 250W RMS
Send John a 250W/ch amplifier, and he will make a video on it.
Excellent information 👍
I would like to see a link to a schematic with the parts identified for the work you did from 9:30-11:00.
I can't find an image anywhere but on page 75 of Small Signal Audio Design by Douglas Self there is good CFP unity-gain buffer. I am making an active crossover with that buffer.
Excellent video!
nice explanation! maybe triple emitter follower would be good for the next video?
Nice video men! Love that cat n.n
great! Thank you!
Sziklai: availabilty as a single chip?
Try the SAP16.
thanks for the expanation =Romania
In the second circuit the emitter of Q2 is supplying 0.51 Amps so you cannot get more than that from its collector!!! what is 5100? It looks to me that the formula is incorrect
Note that the units aren't matched in the diagram; the base of Q1 is listed as ".1" but that is .1 mA, so .0001 amps. That's why it is 5100. He does clearly say "a hundred microamps" when he writes the .1.
17:52 priority interrupt! :-D
Excellent!!!!
awesome
NPN: you're under my power! >:)
PNP: whyyyyyyy :'(
You said point 1 mill amp that would be .0001 what you drew is .1 which is 100 MA.
Thanks!
I pronounce it as sak-lai pair
Good God, feed the poor cat. 😊
It's pronounced "sick lie"
Sick lie hahhhhaaaa
The correct pronunciation was added by a Hungarian before you made your comment.
Since we know you didn't even know, maybe in general with pronunciation you'd do better to use the phrasing "I pronounce it as" or "I've heard it pronounced as" instead of "it is."
It's Hungarian, pronounced sick-la-eeee where 'la' is as in la-la-la-la-I-can't-hear-you. OK, maybe not 4 e's, maybe only 2 or 3