Help understanding transistors.

[pucho812]

O.k. brainacs I am in need of a back to school lesson on transistors.
from basic to advanced I am lost at the moment.
1.what are they designed for? my understanding is it is a gainstage.
2.what is the best way to find replacements for models not made anymore?
3.pnp Vs npn?

 

[rodabod]

[quote author="pucho812"]
1.what are they designed for? my understanding is it is a gainstage.[/quote]

2.what is the best way to find replacements for models not made anymore?

3.pnp Vs npn?

Someone like PRR will be able to give a better description than myself. Your best option is to get an electronics book and start at the transistors chapter - it will probably explain it from the ground up.

1. Transistors are used for many applications. Primarily switching and amplifying. The transistor can be used like a sort of electronically controlled tap (using a water analogy). I've gone off a bit here.

Basically you can use a transistor to control the flow of current. If we use a small voltage (eg. a mic signal) to control a large available current, then we have amplification.

2. There are equivalency data sheets availbale for this, otherwise, you can look at some of the specs. eg Beta, current rating, NPN/PNP, etc.

3. This evolves from the physical construction of the materials within transistors. I always forget the implications, but remember that their configuration in a similar circuit will be different.

Have a look through a book if you can or Google "transistor theory" to learn more. My description is a bit poor.

 

[NewYorkDave]

C'mon, Pucho, don't be lazy. Google brings up a ton of links that explain transistor theory... and I suspect that at least some of them might even be accurate! :wink:

Here's just one of many links:

http://www.tpub.com/content/neets/14179/css/14179_61.htm

 

[pucho812]

true dave. but just cause it's on the web doesn't mean it is completely accurate. thanks fot the link.

 

[pucho812]

dave that link was awesome. :thumb:

 

[NewYorkDave]

Just 'cause it's in a book doesn't mean it's accurate, either. Anyone who's looked at a TAB Publishing book knows what I'm talking about :wink:

 

[bcarso]

[quote author="NewYorkDave"]Just 'cause it's in a book doesn't mean it's accurate, either. Anyone who's looked at a TAB Publishing book knows what I'm talking about :wink:[/quote]

With PRR's one notable exception, boy is that ever the truth.

 

[Larrchild]

Tubes use heat to make electrons to go thru them.

Transistors take 3 pieces of sand, glue them together, and a small signal going into the middle piece of sand makes a bigger signal flow thru the outside pieces. no heat.

npn and pnp are which side of the sand they glue to.

 

[SSLtech]

[quote author="Larrchild"]no heat.[/quote]

Yeah...?

-Really?

-Try telling that to the Studer capstan motor drive amplifier that just roasted a hole in it's circuit board! :green:

May I suggest: "comparatively, almost no heat."

Keef

 

[Larrchild]

Well, yeah, thats so true. I'll try to clarify that. Unlike a tube, heat is not an initial ingredient for conduction, but is a byproduct of it.

hows that?
Sorry bout that hole in the board sir=(

 

[Family Hoof]

Pucho,
Here's my oversimplified take on things. Hope these explanations hlep and not confuse you.

[quote author="pucho812"]1.what are they designed for? my understanding is it is a gainstage.[/quote]
Yes, amplification was the original intention and most any transistor will amplify when operated within the correct range. However, at either extreme of its operating ranges are "cutoff" where no signal passes through, and "saturation" where the output cannot get any higher no mater what you keep throwing at it. The latter is pretty self-explanatory. By exploiting quickly changing between these two conditions a transistor can become a switch instead of an amplifier.

2.what is the best way to find replacements for models not made anymore?

There are three things that I try... 1) Go to www.alldatasheet.com and get the date sheet for the part in question. I look at the characteristics I'm most concerned with such as hfe, bandwidth, and the various voltage ratings. Go to major semiconductor manufacturer website like On Semi, Fairchild, or Central Semi and search through their lines to find something close enough. Also on these website are cross-reference search engines where you can just type in the part # and see if they suggest a replacement from their line. Always double check the results by comparing data sheets. 2) NTE makes generic replacement semiconductors mean't to substitute for a whole bunch of stuff. Their search engine will turn up a result for almost part number that you throw at it. Sometimes they're dead wrong but sometimes you can look at the data sheet for their replacement and it will give you some ball park hints as to what a suitable replacement would look like or what your part did in the first place. 3) Sometimes you can also find dedicated transistor substitute websites and search engines just by google searching for your part number or for "semiconductor cross reference" or something like that.

3.pnp Vs npn?

This simply refers to the polarity of the device. NPN came first and was the equivalent of a tube (and still tend to have slightly better specs to this day). PNP is the opposite polairty and is useful because you can get complementary pairs of NPN+PNP devices and form (arrguably) more efficient, interesting, and complex circuits. Tubes only had one polarity so you had to be more creative.

 

[PRR]

>> no heat.

> Studer capstan motor drive amplifier that just roasted a hole in its circuit board!

Any power system can be made to overheat. Linear power amps are pretty sure to run hot. One small advantage of tubes is: between heater power and very poor conduction, they run hot at best and not much hotter when abused. Transistors need no (extra) heat to get started, and will happily eat all the power they can suck until they soften and quit or burst.

> Tubes use heat to make electrons to go thru them. - Transistors take 3 pieces of sand, glue them together, and a small signal going into the middle piece of sand makes a bigger signal flow thru the outside pieces.

It is hard to make electrons move from metal to vacuum. Heat is one way to make them move. The threshold temperature for a small usable flow is far above room temperature.

I think conventional transistor action also depends on heat. But if you glue that de-oxified sand together tight enough, the threshold temperature is far below room temperature. Ambient thermal energy plus a small voltage will flow big current. So there is heat needed, but anyplace you want to use a transistor probably has ample heat laying around.

Like: to get a steam engine started, you have to raise the temperature of some key parts to 100 deg C. To start a car engine on natural gasoline, the engine can be at 0 deg C or less (much less with refined gasoline blended for winter use).

> lesson on transistors

Travelers in St Louis want to go to San Francisco to work. There is an infinite supply of them. They can't cross by themselves. They need a Guide. One Guide can escort many Travelers.

The travelers stand in a place called "emitter". When Guides run out of money and are thrown out of the bar, they go to a place called "base", which is next to "emitter". Each Guide grabs 20 to 200 travelers and escorts them from Emitter to Collector, in San Fransisco.

The number of travelers who pass from Emitter to Collector is limited by the number of Guides you put in the Base. You can have a thin trickle or a massive flood of travelers.

The work done (pay to the guides) is the distance (2,345 miles, or 40 volts) times the number of travelers (or Amps). Obviously you could imagine a lower-pay low-power situation where Guides only escorted Travelers through, say, a small museum instead of across a great desert.

One monster flaw with this analogy: the Guides actually don't go to San Francisco, they only go Base to Emitter. And if you take that direction as given (it is ambivalent) then the travelers must go Collector to Base. The Guides take a short-cut and still manage to escort a large number of Travelers across a long distance.

And if that don't confuse you....

 

[Family Hoof]

PRR,

Do you teach? You'd be absolutely fabulous at it. Bravo!

 

[pucho812]

lesson from the furu himself PRR thanks :guinness: :guinness: :guinness:

 

[Larrchild]

http://www.bellsystemmemorial.com/belllabs_transistor.html

some cool history on the transistors' birth.
Pretty much what the furu said.

check out alternate names they were batting around, We could have ended up calling it an Ioatron.!

 

[PRR]

> some cool history on the transistors' birth.

Yeah, well, if you read pre-1950 transistor lit you will get confused. All they could actually make those days was Point Contact transistors. These have Alpha greater than 1, while all our transistors have alpha of 0.99. This meant a Point Contact transistor in a Grounded Base connection (the best way to stabilize the DC bias) could give output current greater than input current (which does not happen in tubes or junction transistors). But "production" yield on Point Contacts was 1% to 10%, they had huge noise, pathetic power handling, and failed a lot.

Modern transistors are all Junction. I think the theory goes back to before 1947, but they could not make crystal pure enough until about 1950. The rise of practical transistor systems starts around 1952-1954 when junction transistors were being produced in (small) quantity.

Even then, reliability was not great. Somehow the next step did not come from Ms Reliable Ma Bell, but from IBM for mass-produced "indoors" transistors and a little instrument lab in Texas who packaged transistors tough enough to drop down oil wells. The IBM process was well written-up and copied by Raytheon, RCA, GE, and a few others I forget. TI pioneered the plastic-blob on which all current packages are based.

If you look at voltages 1950-1965, you will see base-emitter drops much smaller than we expect today. Silicon transistors have a factor of 2 in the equation (and I have not found a good reason). Germanium follows theory, and base-emitter voltages are half of Silicon. But Germanium leaks so bad that sometimes it is hard to keep it OFF, rather than having to force the transistor ON as we do in Silicon. Kids today may not get taught Shea's Stability Criteria (known by several names), but it was essential when working in Germanium.