Why is it so hard to answer this question. (Which transistor to select).

[chinoy]

Im a bit of a two stroke engine nut and run a forum where I answer peoples complex questions in simple english.

Ive had this one question nagging me ever since I took up audio. Ive tried it on various forums maybe even here. But never got a simple answer.
I know the answer is not simple. But if you understand a subject part of that is being able to explain complex questions with simple english.

Here is my question.
I have a bag full for transistors. Or the person who has designed the amplifier has given me a list of transistors that can be used.
How do I select the best one.

My application is music.
my objective is to have the best power output efficiency and most of all lowest noise and distortion.
How do I select from a bunch of options.
Say circuit calls for TO92 NPN
I have 10 options with me. Which datasheet parameter should I be looking at. Or what can I measure other than THD and Noise to see which one works better. ?.
Or like some say it really does not matter.

[peter-h]

It does matter, within a certain broad band, but you need to know quite a bit about electronics, so a quick forum answer is not possible. If you post a circuit diagram, someone may be able to help.

[Wallace Gasiewicz]

Make sure that the transistor has the appropriate voltage rating and current rating for the job you want it to do. There are big variations in the current handling and voltage of TO92 package transistors., some applications even require a heat sink cap.

You did not say if you are replacing a transistor on a board or developing a new circuit. If you are replacing a transistor the pinout should be the same, it will be easier to put it in.If you are replacing a transistor you could put in a transistor socket and try different transistors easily.Chose a transistor made for audio, Usually low freq transistors will work better in audio since this is what they are used for.
There are plenty of comments about this topic on audio forums.  This NPN  BJT :  2SC1815SL   is an example of low noise audio transistor.
https://html.alldatasheet.com/html-pdf/110924/TOSHIBA/2SC1815L/295/1/2SC1815L.html    You can look at the noise figure in the data sheet.
The complementary PNP is 2SA1015L
The   "L"   is for low noise.
The transistors I mentioned are old transistors and maybe some newer surface mount transistors would be better. Another problem is that there are a lot of purposely falsely labeled transistor out there, That is another subject.
 

[coromonadalix]

Transistors are more than 3 legs

Silllicuim, Gemanium, Mosfets, bjt ujt ...
gains / HFE
voltages
capacitances in some cases
leakage
avalanche 

Simple ones, darlington  etc ...  Matching pairs

List goes on and on,  in some projects, simple substitution can work, in other cases they uses caracteristics ...

I have some projects who can only use one transistor kind, no substitutions is found at the moment


You will never get an simple answer

If it was so simple,  you would not need tons of brands, tons of part id's  etc ... 2sa 2sb 2sc 2sd  2n  mje 2sk 2sj  ......

You try to find the part who will make everything possible, NOT 

Designs take many factors into consideration, calculations, tryouts and errors

You change one part and everything may become messier, bad,  less good in specs, noisier etc ...


Audio is not Video, Data is not this or that    etc ...  applications depends of the needed use,  vice versa   ...

[Zero999]

Your question is vague and ill-defined, so you won't get a good answer.

There are three different types of transistor normally used in audio amplifiers:
BJT
J-FET
MOSFET

These can then be split up into NPN and PNP and N-channel/P-channel for J-FETs and MOSFETs.

Then there are the different voltage, current and power ratings. J-FETs are purely used for low level signals, BJTs exist in both power and signal types and MOSFETs do come in low power variants, but are mostly high power devices, used in class-D amplifiers nowadays.

You won't get any better answers, without a schematic or photograph of the device in question.

[M4trix]

Write down all the transistors in that bag and we'll explain which one suits best in what type of circuit.

[Bud]

. Or the person who has designed the amplifier has given me a list of transistors that can be used.
How do I select the best one.

Just pick one from his list. You will not get drastic difference from using either of them.
Most often performance of a circuit is a result of the design topology, not particular transistor.

[Alex Nikitin]

Your question is vague and ill-defined, so you won't get a good answer.

There are three different types of transistor normally used in audio amplifiers:
BJT
J-FET
MOSFET

These can then be split up into NPN and PNP and N-channel/P-channel for J-FETs and MOSFETs.

Then there are the different voltage, current and power ratings. J-FETs are purely used for low level signals, BJTs exist in both power and signal types and MOSFETs do come in low power variants, but are mostly high power devices, used in class-D amplifiers nowadays.

Also MOSFETs are divided also into enhancement mode (the vast majority) and depletion mode (N-channel only to my knowledge, and my personal favorite for many audio applications). There are power JFETs nowadays as well.

On the subject of this thread: yes, there is a difference between different type of devices in the same category (i.e. low power NPN BJT) and yes, it is usually measurable in a particular circuit and in many cases can be audible as well. If the circuit is not well designed or made in a pursuit of a particular philosophy  , than even devices of the same type, make and batch can easily show the differences and you have to resort to a selection/matching process.

Cheers

Alex

[TimFox]

Building analog circuits from discrete components is an interesting endeavour, and might produce better results than using integrated components, especially if you need higher voltages than, say, 30 V.
Modern integrated analog components (e.g., op-amps) are very well characterized and are usually easier for implementing full circuits, since all the really hard work has been done for you at the IC manufacturer.
A good reference for audio signal-level design, using both discrete semiconductors and integrated devices, is D Self, Small Signal Audio Design, 2nd ed, Focal Press 2015.
It is free of nonsense, and comparison between different components (e.g., capacitors) is illustrated with actual measurements.
Some of the circuits shown are historical (dating back to germanium transistors).
Chapter 3 is only discrete transistor circuitry, while Chapters 4 and 5 discuss op amps.
(Most of the rest of the book focuses on op amps.)

[Zero999]

Your question is vague and ill-defined, so you won't get a good answer.

There are three different types of transistor normally used in audio amplifiers:
BJT
J-FET
MOSFET

These can then be split up into NPN and PNP and N-channel/P-channel for J-FETs and MOSFETs.

Then there are the different voltage, current and power ratings. J-FETs are purely used for low level signals, BJTs exist in both power and signal types and MOSFETs do come in low power variants, but are mostly high power devices, used in class-D amplifiers nowadays.

Also MOSFETs are divided also into enhancement mode (the vast majority) and depletion mode (N-channel only to my knowledge, and my personal favorite for many audio applications). There are power JFETs nowadays as well.

On the subject of this thread: yes, there is a difference between different type of devices in the same category (i.e. low power NPN BJT) and yes, it is usually measurable in a particular circuit and in many cases can be audible as well. If the circuit is not well designed or made in a pursuit of a particular philosophy  , than even devices of the same type, make and batch can easily show the differences and you have to resort to a selection/matching process.

Cheers

Alex

That's true, but note that J-FETs can also be biased into the enhancement region. The gate impedance drops, but it's not an issue, so long as it doesn't load the circuit driving it and the gate current isn't exceeded.

[Bud]

Gate impedance and depletion/enhancement mode is not what OP wants. He wants simple answer, so give him one.

[Someone]

That's true, but note that J-FETs can also be biased into the enhancement region. The gate impedance drops, but it's not an issue, so long as it doesn't load the circuit driving it and the gate current isn't exceeded.

At which point we also discuss "reverse active" region of bipolar transistors ?

[TimFox]

Besides obvious parameters such as maximum voltage and power dissipation, an important question for choosing a BJT is whether you want high beta or lower beta.
Unfortunately, this otherwise useful discussion neglects that question.
https://wiki.analog.com/university/courses/electronics/text/choosing-transistors

[CatalinaWOW]

Perhaps the best way to answer the OPs question would be to translate it into his terms and then look at the answers here in those terms.  Question:  I have a shelf full of motors and engines.  All I'm interested in is fuel economy and small vehicles.  Which engine would be best?

And our answers;

It's not simple.  There are gas and diesel and electric.

No It's worse than that. There are reed valves and overhead cams and hydraulic lifters and two and four valves per cylinder...

[vk6zgo]

. Or the person who has designed the amplifier has given me a list of transistors that can be used.
How do I select the best one.

Just pick one from his list. You will not get drastic difference from using either of them.
Most often performance of a circuit is a result of the design topology, not particular transistor.

Yes, at one point, the essence of circuit design was to reduce the reliance upon device characteristics to a minimum.
Today, not so much.

[vk6zgo]

That's true, but note that J-FETs can also be biased into the enhancement region. The gate impedance drops, but it's not an issue, so long as it doesn't load the circuit driving it and the gate current isn't exceeded.

At which point we also discuss "reverse active" region of bipolar transistors ?

Back in the dear, dead days beyond recall, when Philips & Mullard were virtually the "only game in town", such things as true "bipolar" transistors existed, which were electrically symmetrical.
They were used in various rather esoteric circuits.
That's why I prefer the term "BJT".

[chinoy]

Ask me a question which you think is complex about the design of a two stroke engine maybe I could give you a simple answer in simple english. That even a newbie could understand. Ive spent the last 20 years of my life answering such questions for my forum members.

Lets narrow it down to TO92 BJT NPN transistors only.
I dont know why everybody feels they could answer the question better if they had the complete circuit. So here goes.
These are the two amps Ive built.
If you scroll down you will see the transistors he has listed as optional.
https://vaspelectronics.com/400w-hifi-mono-amplifier

And the other is this one.
http://sound-au.com/project3a.htm

Complete Schematics and transistor options listed.

What I have self learned from studying old amps that sound amazing is.
1. The more stages you have the better.
2. A low Pf value is critical.
3. Low gain low speed transistors seem to sound and work better.
I also scrounge junk yards for high end Amps just to pull parts that you can no longer find new and which work so much better than just using a regular BC547.
I have a cheap china component tester a 100 Mhz Scope, DatsV3, A USB DAC and a UMIK-1. On the s/w side I have ARTA and REW. To do spectrum analysis and measure THD + Noise.

[JohanH]

That's narrowing down your question considerably. It sounds like you already know in general what works or not. Maybe ask the question in diyaudio.com forums.

Btw, I remember your forum from way back. I still have an RD350 and H1 500 in the shed that I haven't ridden in years. I used to read lots of 2-stroke literature back then (Jennings, Blair, Bell, Cameron, etc. books I have), even wrote a basic 2-stroke exhaust tuning program. I'm a little disappointed the 2-stroke advancements took so long to make it into production and now electric motors are taking over. But that's off topic...

[tooki]

Lets narrow it down to TO92 BJT NPN transistors only.

The fact that it’s a TO-92 tells you one thing, and one thing only, that is of any practical consequence: that the expected heat dissipated in the transistor isn’t expected to exceed the thermal characteristics of the TO-92 package. (Assuming the engineer did power calculations…) That’s it. It doesn’t tell you anything else about the device. If you happen to already know that it’s an NPN BJT, you now know that it’s an NPN BJT with power dissipation under around 600mW.

I suppose you can also make some (completely pointless) assumptions about its maximum voltage (in that it must be something below a voltage that would arc between the leads, so it’s not going to be a 10kV part), and its maximum current (in that it’s not going to be designed to pass 50A through those thin legs). But it can’t tell you anything useful about its voltage and current. Nor will it tell you anything at all about its other characteristics.

I dont know why everybody feels they could answer the question better if they had the complete circuit.

Because to select a transistor, you need to know what function it’s going to perform, and what its operating environment (by which I mean the circuit it’s operating in) is — what voltages and currents will it handle? What frequencies will it see? What kind of gain does it need? Is it operating in the linear region or in saturation? Etc etc etc. And you need to know what your design priorities are, because selecting for one characteristic may mean sacrificing another.

[Tation]

But it can’t tell you anything useful about its voltage and current. Nor will it tell you anything at all about its other characteristics.

It will give a clue about its Cbc being smaller than the Cbc of transistors with bigger packages. The Cbc in the 2nd stage has impact on overall amplifier distortion. So preferring TO-92 packaged transistors on such place, better than others with bigger packages, has some foundation. Suppose that the OP is aware about the books from Douglas Self on the matter.

[tooki]

Somehow I doubt the OP is at that point. 

[tggzzz]

I dont know why everybody feels they could answer the question better if they had the complete circuit.

That seems to be a common point in the threads you start, e.g. the one on measuring voltages with multiple instruments. https://www.eevblog.com/forum/chat/how-do-you-calibrate-your-gear/msg5148885/#msg5148885

You are a beginner who doesn't know what they don't know. Other people have significant experience, and do know what is not in your question.  Hence they ask questions in order to provide better answers.

Listen to the questions, and answer them. That way you will learn something.

Alternatively, if you don't want to provide the necessary information, why ask the question?

[Someone]

Ask me a question which you think is complex about the design of a two stroke engine maybe I could give you a simple answer in simple english. That even a newbie could understand. Ive spent the last 20 years of my life answering such questions for my forum members.

Let's narrow it down to TO92 BJT NPN transistors only.

Let's narrow it down to cam shafts 320mm long.

A boring parameter that has little to do with the action/function of the device.

[chinoy]

Btw, I remember your forum from way back. I still have an RD350 and H1 500 in the shed that I haven't ridden in years. I used to read lots of 2-stroke literature back then (Jennings, Blair, Bell, Cameron, etc. books I have), even wrote a basic 2-stroke exhaust tuning program. I'm a little disappointed the 2-stroke advancements took so long to make it into production and now electric motors are taking over. But that's off topic...

Small world eh.
The old forum grew too big when we crossed 100,000 members. I could not afford the bills. Esp as we refused to run adds or accept sponsors. So I moved it to Face Book. Time permitting I hope to fire up the old forum one last time. But there is a lot of work involved as the forum s/w version is now beyond the point of easy migration.

They have been saying that two strokes are dead for the past 20 years. Yet based on our sales for our products even I am surprised at how well they are doing. I have a RZ350 and a few RDs that I need to rebuild and start riding again. After my 2nd heart attack Drs said to avoid 2 wheelers. But I plan to freshen up the engines with some fresh oil seals and ride them.

Sorry for going OT.

And thanks for all the helpful answers.
What I have learned since I made this post is that
1. You cant measure Noise and THD on a cheap scope with FFT. Much better to use software.
2, In order to use software I need to get an external USB sound card.

As a kid in 79 I blew my dads pride and joy an Akai amp. He flew jets for the IAF and the best of the best instructors were sent to Iraq to train their pilots on Mig 21s. It was during this tenure that he purchased his amp. So its not like he could go out and get another or fix it. Im trying to recreate that transistor sound as an AMP for him.

I have built a few quasi complementary amps, som chip amps LM3886 / TDA7294 / STK and then the P3A and vasp class A/B.
In most cases I buy just the PCB from a local vendor and do the rest following their schematic. And recommended mods.
I have been in audio for all of 1`year. So forgive the noob questions. And Id be happy to answer any engine related posts in the Everything else section if I can.  (Mods can delete this post if its TMI. ).