Hi,
i have a very trivial question
I have noticed one thing. Some very high end NPN bjts do not have the complimentary PNP.
Question: does this mean that NPN are intrinsically better than PNP ?
Second much more weird question: is a NPNs only preamp/amp design possible ?
As a consequence i can say that i am more and more intrigued by quasi complimentary topologies ... very very much.
The nice thing is also that i have only one output bjt model to deal with
Thanks a lot for any explanation.
Kind regards, gino
i have a very trivial question
I have noticed one thing. Some very high end NPN bjts do not have the complimentary PNP.
Question: does this mean that NPN are intrinsically better than PNP ?
Second much more weird question: is a NPNs only preamp/amp design possible ?
As a consequence i can say that i am more and more intrigued by quasi complimentary topologies ... very very much.
The nice thing is also that i have only one output bjt model to deal with
Thanks a lot for any explanation.
Kind regards, gino
Last edited:
It is called Quasi-Complimentary (Almost Complementary). NPN power transistors are usually less expensive than PNP. It stems from an early design from RCA back in the early 70's.
NPN's were historically harder to make in devices such as power transistors, particularly germanium.
An all NPN or all PNP preamp is easily do-able , also quasi output stages can sound really good because of the "unique" distortion spectrum they produce.
And no NPN/PNP complements are mirror images of each other... some just come a bit closer to that ideal than others.
An all NPN or all PNP preamp is easily do-able , also quasi output stages can sound really good because of the "unique" distortion spectrum they produce.
And no NPN/PNP complements are mirror images of each other... some just come a bit closer to that ideal than others.
That is why the output stage needs so much negative feedback, to try and keep the gains as equal as possible and reduce non linear effect of the Quasi output stage. Some issues were ironed out with the employment of a uA741 as a driver/feedback sensor. Now that shows my age!
Thanks a lot for the very valuable information.
But as i said there are NPNs of very very high power, a power that PNPs cannot reach.
So i was thinking that not only are easier to make but also for how they are designed are intrinsically more robust, more reliable ... better.
Thanks again and kind regards, gino
Thanks a lot indeed and very very interesting.
Could you tell me more about this nice distortion sprectrum ... it is very interesting.
Still i think that from a structure point of view NPNs could be more robust ... just a guess of course.
If PNPs are more difficult to make this must mean something.
This is also very interesting.
Another advantage of at least a all NPN output pair is that the parts list is shorter .... the less the different parts the lower the cost and confusion.
Thanks and regards, gino
The majority carriers are different for PNP and NPN devices. This effects the semiconductor design but not the mechanical construction.
Because of the difference in majority carriers, the performance of the two is different with the PNP being more difficult to manufacture to mirror the NPN performance in high power devices.
This is also true for P and N type MOS FETs.
Because of the difference in majority carriers, the performance of the two is different with the PNP being more difficult to manufacture to mirror the NPN performance in high power devices.
This is also true for P and N type MOS FETs.
NPN were the easy ones to manufacture.
PNP required more knowledge of the BJT structure to get them to work at all.
Gradually over the decades the PNPs have now virtually caught up even in the power transistor range.
However in mosFETs it seems that N & P channel can never be equal, unless technology finds another solution.
There is one area where PNP are reputedly better than NPN: less noise.
PNP required more knowledge of the BJT structure to get them to work at all.
Gradually over the decades the PNPs have now virtually caught up even in the power transistor range.
However in mosFETs it seems that N & P channel can never be equal, unless technology finds another solution.
There is one area where PNP are reputedly better than NPN: less noise.
Hi and sorry i expressed myself badly.
I was thinking more to some kind of an " electrical robustness " ... in the sense they are intrinsecally better for high power application, like in a power amp output stage.
Maybe a NPN 500W multiemitter ala Sanken could solve all the problem to audio people for high power amp ?
instead of having to make parallel of many devices with the usual matching issues.
I love Sankens solution immensely.
They should develop the ultimate power NPN 😱
Thanks a lot and kind regards, gino
Last edited:
Hi and thanks a lot. So there is not a intrinsic superiority of NPNs for high power applications. I thought differently.
This is very interesting indeed.
So in an input stage a PNP pair could be better than the usual NPN pair.
This is very interesting indeed. Very.
Like PNP for low power apps and NPN for high power apps.
Kind regards, gino
Last edited:
Quasi stages have an unavoidable "kink" in the transfer characteristic as they pass through the crossover region, however that asymmetry causes a spray of harmonics which many seem to find pleasing to the ear with the quality often described as "musical" and non fatiguing. The absolute numbers are still vanishingly small... but seems to give it an audible character non the less.
Hi and thank you very much indeed. Very interesting as i said before.
I do not know if it easier to get a larger SOA for NPN for instance
One thing is true. That the most powerful NPNs do not have complimentary PNPs.
This should mean something. For me would be enough to try to use only NPNs were power is concerned, with a suitable circuit of course.
Thanks again, gino
What's wrong with complementary pairs at 250W each?
That gives a quite powerful output.
Double up to a 2pr output stage and that covers most domestic amplifiers upto very powerful.
Only PA duty requires more.
That gives a quite powerful output.
Double up to a 2pr output stage and that covers most domestic amplifiers upto very powerful.
Only PA duty requires more.
Hi and they have the same SOA ?
I mean is the PNP less reliable than the NPN ?
Yes i understand this is the norm and i think must work because PA amps must be very reliable. But just a single pair will avoid any the matching issue
I have noticed lately that even lower power integrated the review are more positive when there is a single pair at the output.
I know ... it is more a obsession that something with a scientific basis.
What you think ... i single pair can sound better ?
I do not like these long rows of bjts ...
Thanks and regards, gino
The SOA are the same. I have not seen any manufacturer data nor any reports that one is more reliable than the other.
Have a look at MJ15003/4 and MJ21193/4
Long rows is for high dissipation. Only required if your design needs it.
One pair of 250W devices is good for a 100W amplifier.
two pair of 250W devices is good for a 200W amplifier.
For domestic use I doubt you need to go to 3pair.
Thank you sincerely again. No more silly questions
Also because i was looking at power bjts on ebay.com
Very often they are sold already in npn/pnp pairs, matched pairs to be precise.
This means to me that complementary pairs are the just the norm for output stages.
No need to make strange choices ... things are already very complicated.
I would like to take profit of your kindness and knowledge 😱
I am attaching a schematic i cannot get out of my mind for a simple question
The question: do you think it will work just fine with
R32=R33=R34= 2K2 ????
I would like to reduce parts number as much as possible an buy nice ones.
I have already bought npn and pnp
Then i would buy a PS kit based on tl783 chip.
Maybe i will be able to build my first line stage ever
Thanks a lot indeed
Kindest regards, gino
Attachments
Last edited:
I would not put too much weight on 'Matching" of devices as sold on ebay unless matching data is provided. Even then I'd take it with a grain of salt.
Hi yes Sir ! You are right 😉
Someone pointed me to this file here 🙄
http://www.acousticpsychos.com/Files/Radford-HD250-ZD22-int-sm.pdf
On page 15 the schematic of the line stage can be found
Not needing the tone controls section and the ouptut section, i am focusing on just one gain stage, the input one.
Vsupply is uncommonly very very high at 60V, single supply.
For this i will use the tl783.
But i would like to simplify it even more reducing parts types.
Using all 2K2 resistors it would be easier to build.
It is my very first attempt to build something.
Do you think it could work with the following values
R 32, 33, 34 = 2K2 ?
What could be a good Ibias value the npn bjt "TS7" ?
I am using overrated parts anyway ... avoiding in this way the need of any heatsink on bjts.
Thanks a lot and kind regards, gino
Last edited:
- Status
- Not open for further replies.