Talk:Transistor

From Wikipedia, the free encyclopedia

WikiProject on Electronics This article is part of WikiProject Electronics, an attempt to provide a standard approach to writing articles about electronics on Wikipedia. If you would like to participate, you can choose to edit the article attached to this page, or visit the project page, where you can join the project and see a list of open tasks.
Start This article has been rated as Start-Class on the quality scale.
Mid This article has been rated as Mid-importance on the importance scale.

Contents

[edit] Archiving

Previous contents of Talk:Transistor became too big, so moved to new page Talk: Transistor 1 (this page will be page 2) - CPES 04:38, 26 Apr 2005 (UTC)

[edit] Oscar Heil

Is anyone aware of details of this claim about Dr Oscar Heil patenting the FET in germany in 1934?--Light current 17:03, 11 November 2005 (UTC)

  • Best I can do is this link JFETS: THE NEW FRONTIER wherein it states: "Field-effect transistors (FETs) have been around for a long time; in fact, they were invented, at least theoretically, before the bipolar transistors. The basic principle of the FET has been known since J.E. Lilienfeld’s US patent in 1930, and Oscar Heil described the possibility of controlling the resistance in a semiconducting material with an electric field in a British patent in 1935." DV8 2XL 18:28, 11 November 2005 (UTC)
I forgot to mention here that in June 2006 I made an article on Oskar Heil, including a copy of his patent drawing of a FET-like structure. Dicklyon 06:57, 27 August 2006 (UTC)

[edit] Pictures

Does anyone have a copyright free picture(s) of transistors we could use at the top of the page to replace the rather grey looking one.?--Light current 17:25, 9 November 2005 (UTC)

Here are some more: commons:Category:Transistors — Omegatron 18:02, 9 November 2005 (UTC)

Replace lead pic with better one. This is still not a very good pic and we need a better one if we are to submit article as a featured article.--Light current 18:28, 9 November 2005 (UTC)

I added one I just took. --agr 20:43, 11 November 2005 (UTC)

Thanks. Thats better!--Light current 20:58, 11 November 2005 (UTC)

A picture of the 'chip' inside the transistor would be nice too.

I just did a Google Search. There r bout 67,000 images. But not sure how many r copy right free. I found this [1] much better. Most of the images can be used on wikipedia. --Electron Kid 02:34, 29 January 2006 (UTC)

[edit] Band gap

"The band gap is highly temperature dependent, decreasing with increasing temperature." part erased: the energy gap is always the same. What changes with temperature is the probability to cross the gap. User: 129.13.186.1

The above correction is from tran history and makes a very good point. I have read that the band gap varies with temperature. It also says this on Wiki band gap, so this needs to be sorted. I have changed the "Semiconductor material" sub section to cover this aspect. CPES 12:38, 25 Apr 2005 (UTC)

The band gap varies with temperature, but very little, and this is an almost insignificant contributor to changing conductivity. As stated above, the most significant factor is the ability of the carriers to cross the gap. Phil Holmes 15:34, 19 August 2005 (UTC)

[edit] Problems with part of 12.3 - development

Consider the following paragraph: "Brattain started working on building such a device, and tantalizing hints of amplification continued to appear as the team worked on the problem. One day the system would work and the next it wouldn't. In one instance a non-working system started working when placed in water. The two eventually developed a new branch of quantum mechanics known as surface physics to account for the behaviour."

It seems unclear to simply put in the term 'team' without specifying the team members & where they're working, as I see no mention of these things in the preceding paragraphs.. Later, "The two eventually developed... " - which two? Brattain & somebody, or 2 completely new somebodies? I'd like someone to write this paragraph more clearly, please. -- tharkun860 July 5, 2005 01:46 (UTC)

[edit] Stuff here and there and re-arrangement needed

Two types of transistors? I believe we do have a gazillion types of transistors these days. Perhaps we should accurately say, "in the beginning, there are two basic types of transistors, BJT and FET".

Looks like this article need some re-arranging. The types were described above (Semiconductor type, power ratings, etc), and then towards the bottom, we see BJT, FET, etc. They can probably live happily in one section. I am translating this page into wiki:id, I'm lucky as no one will protest the way I'm arranging the sections, heheheh... Kortsleting 03:23, 18 August 2005 (UTC)

[edit] Ending of initial paragraph is very inaccurate

It says "... allows a precise amount of current to flow through it from the circuit's voltage supply." I would recommend to replace "precise" with something like "substantial" or "significant". Using "precise" is completely wrong because in fact every individual transistor is very imprecise, and it's parameters vary widely from sample to sample, and with temperature. An electronic circuit can be made precise if a certain combination of transistors is used (like OA, operational amplifier) having high amplification but with a deep negative feedback loop. - apredtechenski, Sept 06, 2005

Agree. Ive changed it!--Light current 00:39, 21 September 2005 (UTC)

[edit] Put this page on a diet!

Just been trying to tidy the article and make it flow, but there is just too much stuff here that really needs to be elsewhere. For instance the history of semiconductors should go to the semiconductor page if there is one, detailed description of BJTs is already covered on BJT page. There is too much emphasis on FETs at the expense of BJTs etc. Can we have comments please on how this page can be slimmed down and be made an easier read?--Light current 00:47, 21 September 2005 (UTC)

I copied the semiconductor history stuff (its still here as well tho') over to semiconductor device which was a short article anyway, and I think the stuff looks good over there and fits in well. Have a look. If everyones happy, then we can delete that stuff from her as a start in tidy up operations.--Light current 01:36, 21 September 2005 (UTC)

Since there were no objections, I have deleted the large section 'History of semiconductor devices'. It can now be found on the semiconductor device page.--Light current 20:31, 25 September 2005 (UTC)

Also, after we point out that transistors are far less microphonic than tubes, do we really need to the link to the Guitar Amplifiers page? I love tube driven single ended hi-fi amps, and all kinds of Guitar Amps too, but come on, is this relevant?

[edit] Opto FETs

I seem to remember that optically activated FETs are available. If so, this should be mentioned in the appropriate paragraph.--Light current 02:06, 4 October 2005 (UTC)

I don't know if discreet opto-FETs are available, but they're certainly available when packaged with an LED as an optical isolator.
Atlant 12:44, 4 October 2005 (UTC)

Yes thats what I meant. Thanks for confirming--Light current 23:31, 4 October 2005 (UTC)

Discrete photo fet at: http://www.linearsystems.com/datasheets/LS627.pdf I think there are a load for hs fibre rx too. But most are in optocouplers as said. Here is some description stuff: http://www.radio-electronics.com/info/data/semicond/phototransistor/photo_transistor.php - CPES 23:36, 27 April 2006 (UTC)

[edit] Proportional current?

Im worried about the phrase 'allows a proportional current to flow'. In fact the current is not proportional to the voltage in a BJT but prop to the exp of the voltage. So this lead para needs rewording. Any suggestions?--Light current 01:09, 5 October 2005 (UTC)

I rewrote that sentence. Snafflekid 05:28, 5 October 2005 (UTC)

The basic problem, which is currently not emphasised, is that BJTs and FETs operate in different ways. You push some current into the base of a BJT and more current flows thru the E/C terminals. On the other hand you put a voltage between the source and gate of a FET and it conducts a proportional current between the source and drain. In both cases this current flows regardless of the e/c or s/d voltage (provided that the e/c or s/d voltages are above saturation). Simple as that. CPES 02:05, 24 April 2006 (UTC) (80.177.169.33)

Not exactly. BJT transconductance (ratio of C-E current to B-E voltage) is exponential, whereas FET transconductance is parabolic. Neither one is really proporional in input voltage. Sure, BJT collector current is more or less proportional to base current (beta), but it's highly variable based on temperature, current, and even individual device. BJT circuits are usually not designed based on a specific value of beta, you can't depend on it. One a specific die, two devices will have the same beta, but it's hard to control. Conversely BJT circuits are designed based on transconductance, which you can depend on. FETs are a real problem, their transconductance can vary by a factor of two or more for a given device. Take a look at the specs for the venerable MPF102. IDss and Vgs(off) vary widely. As for pushing a current into G-S of an FET, you can use your imagination ;-) Madhu 22:36, 25 April 2006 (UTC)

[edit] History section??

We used to have a history section that described where the name came from, etc. What happened to it?

Wait, don't tell me. Someone declared it "useless trivia" and deleted it. — Omegatron 04:21, 13 October 2005 (UTC)

Yes it was I who moved history to semiconductor devices after discussion and agreement I believe!--Light current 00:30, 14 October 2005 (UTC)

Not exactly. Semiconductor and Semiconductor device were to be merged. I did cleanup on the two pages. The history section of Transistor was moved to Semiconductor device, because Transistor was getting too long, as I recall. And, the "history of transistor development" involves more than transistors. I think that the history makes a good fit over there. Maybe it would be good, however, to add a very brief history section to Transistor with a link to Semiconductor device for more info, since I bet a lot of people will look here first for a history. Snafflekid 18:49, 13 October 2005 (UTC)

[edit] Removed from page

All This stuff has been removed from the page because: A) This text is not about transistors B) page is about transitors and not transitor amplifiers. C) page is not about audiophiles or how transistors sound

Some argue that the larger number of electrons flowing in a vacuum tube behave with greater statistical accuracy, although this ignores the fact that vacuum tubes generally have a high-impedance control terminal (grid), and that discrete transistor circuits (as opposed to integrated circuits) can also be designed to use large currents.

Others detect a distinctive "warmth" to the sound. The "warmth" is actually distortion caused by the vacuum tubes, which some audiophiles find pleasing. This is "soft-saturation" which occurs when vacuum tubes are overdriven, causing poorly designed vacuum tube amplifiers to sound better than poorly designed transistor amplifiers. Tube amplifiers are also less prone to slew-rate limiting, which was a problem with early semiconductors and is still observed in low-cost transistor audio amplifiers.

Above speculative opinion gives no correlation between characteristics or measurements and perceived sound quality. See Tubes vs. Transistors - Is There An Audible Difference? and Transistors vs. Tubes - Brief Feature Comparison.

Vacuum tubes are also preferred in guitar amplifiers which are designed to be overdriven, because they have a different non-linear transfer characteristic than transistors, and create a different, more pleasing spectrum of harmonic distortion or "fuzz". Digital signal processing (DSP) can be used to achieve similar effects in the digital domain.

(single-ended transformer coupled to push-pull transformer coupled to push-pull capacitor coupled to push-pull direct coupled). DSP Recently, inroads have been made in digital signal processing (DSP). DSP is a technique that can (among other things) be used with A/D and D/A converters to allow a digital processor (along with a computer program) to manipulate analog signals.

I totally agree that this stuff does not belong here. I didn't have the energy to start a new page at the time. :) Snafflekid 02:11, 14 October 2005 (UTC)

If its to go anywhere, it should go to audio amplifiers or electronic amplifiers--Light current 00:50, 14 October 2005 (UTC)

  1. It most definitely "goes somewhere". Deleting large chunks of information is not helpful to anyone. Yeah, it's still in the history, but it should be in an article, and people generally don't go digging around in the history unless they are looking for something specific.
  2. Better places to put it would be Tubes vs transistors, Valve sound, or Audiophile.
  3. When something is moved from a general article into a more specific article like this, there should still be a short summary in the main article with a link, to keep the articles tied together. This is definitely relevant enough to the transistor article to get a sentence or two and a link. — Omegatron 01:58, 14 October 2005 (UTC)

I try to delete as rarely as possible. At bare minimum it is better to move it to an new page or even cut and paste the unwanted stuff onto a new wikipage and let someone else edit it IMHO. Snafflekid 02:09, 14 October 2005 (UTC)

Moving to the talk page is a very good alternative to deleting it, too. — Omegatron 02:15, 14 October 2005 (UTC)

Im not sure if you are telling me off here 'O'. Have I done something wrong IYO?--Light current 02:21, 14 October 2005 (UTC)

[edit] Unweildiness - is that a word?

Page is starting to get too big again with addition of lots of stuff thats not strictly relevant and that could go in other pages. Please be carefeul about whats added to this page.

"There is no need for haste." See Wikipedia:Article size. — Omegatron 01:50, 14 October 2005 (UTC)

32kB is just a suggestion, but a good suggestion. However, transistor has a very big field to cover. Snafflekid 02:04, 14 October 2005 (UTC)

[edit] Not Fair

sorry! I forgot bout the size of page. now, i have provided links to images. but dont u think that we are not doing justice with transistors; comparing the vastness of the capacitor page and this page :) .we must include many concepts like Ebers Moll Models, Π -model, h-model, etc. or if these are already present, dont u think that they should be merged? --Davy Jones 04:37, 16 October 2005 (UTC)

Heh, well you could create new pages for each of those concepts and link them into Transistor. BTW, did you ever find your locker? Snafflekid 06:37, 16 October 2005 (UTC)

[edit] Define your terms

The terms "base" "emitter" and "collector" are introduced without description.

[edit] Transistron

The Wikimedia Help Desk received advice from an expert in the field that the article in IEEE Spectrum How Europe missed the Transistor contains considerable errors which have been placed in the Wikipedia article. The IEEE apparently has been asked to correct the record:

Specifically, the informed source says:

"Your page about transistors contains an abstract of the Spectrum article and should be reviewed as follows: The correct term is "transistron". The French "Compagnie des Freins et Signaux Westinghouse" was not a subsidiary of Westinghouse Electric. The amplifier developed by Welker and Mataré was not a point contact device. It was based on the minority carrier injection process."

Capitalistroadster 07:33, 6 December 2005 (UTC)

The mistakes are not the IEEEs but mine; though the transistron (correct spelling) did have point contacts and all bipolar transistors rely on minority carriers. The French company must at least have been an affilate of the Westinghouse (railway) brake company. --Wtshymanski 03:47, 8 December 2005 (UTC)

[edit] Transistor biasing-The most basic question

Why is the base-emitter junvtion of the transistor forword biased and the collector-base junction reverse biased?Why not the other way round?

If the BJT is used in the forward active mode (which is almost always the case) the collector voltage is higher than the base voltage and the base voltage is higher than the emitter voltage. For an N-P-N you get the condition you mentioned. If the collector is pulled low enough it will become forward biased also, this is called saturation. Snafflekid 17:22, 8 December 2005 (UTC)

If you biased the B-E and C-B junctions as mentioned by the original (anonymous) poster, the collector and emitter terminals would be reversed (more or less). This does work, but not very well. In normal configuration, if the B-E junction is not forward biased, no current will flow into the base and no transistor action will result. Transistor action is not necessarily intuitive, that's part of the reason why BJT discovery was more experimental than theoretical. FETs are more intuitive, but fabrication was harder in the early days. Madhu 23:46, 8 December 2005 (UTC)

[edit] History?

If Lilienfeld patented the transistor in 1928, how come it was not used immediately? Military uses must be pretty obvious, etc. Also very hard to believe that no US electric industry giant realized its potential. Imagine where computing would be today if those twenty years from 1928 to 1947 were not wasted on vacuum tubes and clicking relays! We could already have true 3-law robots. 195.70.32.136 09:44, 23 December 2005 (UTC)

I don't know all the details of Lilienfelds work, but I don't think he made anything that actually worked. At the minimum, I don't think he had the theoretical foundation developed by Shockley. Lots of people had transistor like ideas, but at the end of the day, Shockley, Bardeen, and Brattain developed a real device and later refined it to a level that few could have predicted. Also, keep in mind that quantum mechanics was still in it's infancy in the early part of the century and the sophisticated techniques needed to grow high purity materials were not a trivial matter. Look at it this way: if you tried, you could build a home made vacuum tube today that works reasonably well. Building a transistor that works at all is much harder. Get a copy of Crystal Fire, it's a fantastic read.
As far as vacuum tubes go, they were cheaper and more reliable than transistors for decades. We still use vacuum tubes today, but not for small signal applications. Most TVs still use CRTs and microwave ovens use magnetrons. Many transmitters still use Klystrons and power companies probably continue to use Ignitrons. CRTs will probably be replaced completely in the coming years, but a cheap, one kilowatt, microwave transistor is a tall order. Madhu 23:05, 23 December 2005 (UTC)
I don't know all the details of Lilienfelds work, but I don't think he made anything that actually worked. That seems a bit amazing. You really think that Lilienfeld patented those very detailed devices, spending significant funding for three separate patents ...but without ever building any of them?!!! Really? If so, this Lilienfeld must be a lunatic. Ah, I found a reference that suggests you're wrong: semiconductor specialist Dr. Harry E. Stockman in a 1981 letter to Wireless World magazine: "(Lilienfeld) created his non-tube device around 1923, with one foot in Canada and the other in the USA, and the date of his Canadian patent application was October 1925. Later American patents followed, which should have been well known to the Bell Labs patent office. Lilienfeld demonstrated his remarkable tubeless radio receiver on many occasions, but God help a fellow who at that time threatened the reign of the tube." So it might seem amazing that Lilienfeld's transistor was ignored for twenty years, but only if you aren't aware of the stiff resistance we humans have to new ideas, and the large barriers we erect to slow them down. Look at the embarassing story of S. Ovshinsky's amorphous semiconductors, their total rejection by the USA sci-tech community, and the multi-billion dollar industry they spawned ...in Japan.--Wjbeaty 01:59, 28 October 2006 (UTC)
That's very interesting, if true, that Lilienfeld was able to making a working semiconductor amplifier device. Conventional received wisdom says no, like this recent book. It would be great if someone could turn up a verifiable source to the contrary. Dicklyon 03:07, 28 October 2006 (UTC)
These books and articles seem very strange to me, claiming knowledge about Lilienfeld's devices for which I've never seen evidence. From my past reading about Lilienfeld, the only evidence leads me to conclude that we don't know the details of Lilienfeld's accomplishments. If true, then nobody has any right to say that he never build a transistor, or to say that he build non-working devices. Absence of evidence is not evidence of absence. If evidence is lacking, we must conclude that we don't know,. We're not allowed to leap to unwarrented and unsupported facts such as statement that "his devices didn't work," etc.
Regarding Shockley and the theoretical basis: as I understand it, the Germanium transistor came first, the theoretical stuff second, and Shockley had little role in the initial discovery. The two guys who built that razor-slit-gold-foil-contact transistor had been ordered by Shockley to stop pursuing any such work since it obviously wouldn't lead to the FET they were trying to invent. They were proceeding anyway, while keeping their equipment on a rolling cart stored in a closet away from Shockley's eyes. Once their investigations resulted in a working device, their boss took over the project and developed the theory as well as the grown-junction improved transistors. (Sounds very similar to Townes' story about how the first Laser was developed, eh?) The moral: you can discover the transistor if you just dedicate significant time looking for it. The advanced theory is added later, and only gives improvements.
On the other hand, the Bell Labs group ran up against a major barrier in their efforts to create FETs: surface states which act to shield the conductive channel from the effects of the gate voltage. Perhaps these authors mentioning Lilienfeld conclude that, since the "surface states problem" is such a major barrier for Bell Labs, that Lilienfeld must have encountered the same problem. Perhaps they assume that Lilienfeld dishonestly pretended to have working devices, while in reality he failed in the same way that Bell Labs initially failed. If this is the basis of these authors ideas, then their ideas are pure speculation, and they're dishonest if they don't label them as such. Or maybe I see those authors as more intelligent than they actually are; maybe one of them decided to claim in writing that Lilienfeld had no working hardware ...and all the other authors just copied the first guy without thinking too much about it. But this is science, and Nature can cut through all the speculative BS: if we build a Lilienfeld transistor as detailed in his patents, and it gives useful gain, then all these authors are wrong. Here's a major possibility I saw in some article long ago: Lilienfeld's devices were not FETs at all, but instead were grown-junction NPN transistors created when the material of the gate-conductor diffused into the semiconductor layer and reversed its doping. This would suggest that Lilienfeld could have built a genuine transistor radio in 1925 ...but that he didn't have an accurate theoretical description of his devices. As with the Bell Labs transistor, the working hardware comes first, and only later do scientists extend physics theory in order to explain it. Or perhaps Lilienfeld behaved as an inventor rather than a physicist by getting "dollar signs in the eyes" and keeping his work secret rather than publishing all the details in physics journals. --Wjbeaty 22:02, 28 October 2006 (UTC)

[edit] Page size

The page is getting big. I suggest we put the more technical BJT and FET stuff (equations, models) on the respective device pages. - mako 07:32, 30 December 2005 (UTC)

I had the same thought. This page gets bloated regularly. Snafflekid 15:44, 1 January 2006 (UTC)

I moved the BJT stuff off to the BJT page. I'm not familiar with the equation in the FET section, though; where does it belong? - mako 05:35, 22 January 2006 (UTC)

[edit] Vandalism

This user deleted the entire page on the transistor. I fixed the page. Perhaps someone could look into this to see if this has occured before, or is occurring regularly.

11:18, 24 January 2006 82.211.102.159 - This user deleted the entire contents of the page.

Vandalism on Wikipedia happens all the time. A large number of Wikipedians spend an awful lot of time fixing it, whether it's the blanking of an entire article, the insertion of "Xxx is so Gay!", or the subtle changing of a numerical value from accurate to false. Welcome to the club!
One thing you can always do is look at the edit history of an article. Click on the vandal's IP address (it's almost always an anonymous IP user that's doing the vandalism) and you'll get a list of their contributions. Click on any that interest you, especially recent ones where they are shown as holding the top edit. Then, if necessary, see WP:VIP on the procedures to report them.
You may also want to see Wikipedia:Recent changes patrol.
IMNSHO, vandalism is one of the largest problems faced by WIkipedia.
Atlant 13:38, 24 January 2006 (UTC)

[edit] Vacuum tubes

The article mentions the advantages of transistors over vacuum tubes, but not the other way around. I believe there still are applications where vacuum tubes are preferred over transistors, like high fidelity audio. At least until recently. I don't know if there has been such advancement in transistor technology that maybe vacuum tubes are not used anymore at al where they can be replaced by transistors.(MrDeBeuker 19:40, 16 February 2006 (UTC))

You understand that the concept that vacuum tubes are better for audio than transistors is basically a religious belief, right? That is, it's supported by faith but basically no evidence any more. See Audiophile and High-end audio.
On the other hand, tubes probably still are better if you're trying to design equipment that will survive EMP; that's about the only place I can think of (besides highly-specialized tubes such as CRTs, magnetrons, klystrons, and TWTs) where tubes still have any edge at all over solid-state electronics.
Atlant 22:12, 16 February 2006 (UTC)
Well, there's sometimes talk about harmonics, but it does boil down to audiophilery. - mako 23:54, 16 February 2006 (UTC)
The big difference is that tube amps tend to "soft clip" as you drive them to their limit whereas transistor amps tend to hard clip; that is, the transistor amps operate linearly right up to the point where they don't but in tube amps, the gain rolls off as you drive near the limits, rounding off the clipping waveform. In this way, when they're both driven into the clipping region, tube amps produce a lot fewer harmonics. But why drive either into clipping, ehh? :-)
And then there's microphonics...
Atlant 00:34, 17 February 2006 (UTC)
You might be interested in Valve sound. There has been quite a bit of discussion on the topic. Madhu 02:21, 18 February 2006 (UTC)
It is more "preference" than "faith" if one prefers the sound of tube audio. If I prefer chocolate to vanilla, I don't see how that is "faith." The harmonic distortion modes are different. Then there's the old saying "Transistors hiss, tubes hum." There's also the notion that recordings in the tube era were doctored to sound good when played with tube amplifiers through big bass reflex speakers. When played through more "accurate" equipment, they might therefore not sound as good.Edison 16:32, 17 August 2006 (UTC)
The differences in harmonic distortion content are unrelated to vacuum tubes vs transistors; they depend on topology. Tubes just tend to be used in asymmetrical class A more often than transistors. Likewise with "hiss" noise levels and ground loop hum.
All of these differences are a non-issue, anyway, if the sound is not run into clipping and the amplifier is designed well enough to have inaudible distortion/noise/hum, which can be done with either device. — Omegatron 17:41, 17 August 2006 (UTC)
IMHO it's a "preference" if someone says they prefer Coke and consistently can tell the difference between Coke and Pepsi in a blind taste test. It's "faith" if someone always orders Coke and says it tastes better, but prefers the taste of Pepsi in a blind taste test if they are told that it is Coke.
I'm speaking, by the way, as someone who always believed I could tell the difference between Coke and Pepsi until one day at McDonald's I said to myself "This Coke tastes a little funny," and went on drinking it until my daughter said "Dad, you've got my root beer." Dpbsmith (talk) 18:05, 17 August 2006 (UTC)

[edit] Physical layout of a transistor on IC

Please note also that all of bipolar PNP and NPN and JFET do not have graphics or scheme displaying physical layout of the transistor. I can try to do something according to that what is in my University textbook, but I will not do it w/o consulting others.

A simple, basic layout, even 3D, cannot hurt the matter anyway. -- Mtodorov 69 13:42, 10 March 2006 (UTC)

[edit] Transistor observed in 1910s ???

According to Bell System Memorial there were accounts in British magazines from the 1910s about Russian ship board operators achieving gain from "cat's whisker" diodes with two whiskers. Worth a follow up and adding to the history section?

[edit] SETs

Single-electron transistors are not mentioned at all in the article and neither have a page of their own. Could someone fix this? --Khokkanen 20:57, 22 March 2006 (UTC)

I added a mention to this article. We have an article on the SET. --Heron 22:05, 22 March 2006 (UTC)

[edit] Transistor Curve Tracer Images

Tran page is much leaner and cleaner now thanks to the good work of the electronics standardisation team and others. So what is the relevance the images of the tran curve tracer: suggest this should be deleted? Also, it would be nice if we had some images of tran packages showing surface mount thu to TO3 to give an indication of package size range. CPES 00:52, 24 April 2006 (UTC) (80.177.169.33)

[edit] Tran Advantages

"though this is usually a desirable part of the sound of guitar amplifiers" There are many valve characteristics that contribute to the "valve sound" both for guitar amps and hi fi amps (valve amplifier topology also has a major influence). But the modulation of the valve characteristics by vibration is not a significant or consistent one. In some layouts the valve amp can be physically isolated from the speakers but this does not affect the sound. In fact valve manufacturers go to great lengths to minimise vibration effect and they often do this in different ways giving different characteristics. (Russian EL34s compared to Mullard EL34s). Suggest this statement be deleted. CPES 01:22, 24 April 2006 (UTC) (80.177.169.33)

Also statement "and can take advantage of the higher electron mobility in a near vacuum" is not really a main advantage but more of a tube design aspect which is not visible to the user as an advantage. Valves are not faster than trans (not anymore anyway). This aspect does not fit with rugged aspect either. KIS also applies. Suggest delete. - CPES 22:25, 27 April 2006 (UTC) (80.177.169.33)

[edit] Tran Images & Symbols

Currently there are two images of thru hole trans, both essentially showing the same thing. The original image (with tape measure) is best, because it shows a progressive size range and indicates scale. Suggest image presently in intro be removed and "tape measure" tran image be moved to intro. An image of surface mount trans would compliment the thru hole image: there once was one but it seems to have got lost in the major revision of tran page. (nice to see the better tran symbols image has been reinstated: but still no IGFET symbols in it) - CPES 20:21, 26 April 2006 (UTC) (80.177.169.33)

Any chance someone could check norm to put correct symbils into page. Quick check of other language versions showed pretty much every use different symbols.

[edit] Bipolar Tran Name

Currently the BJT para says that they are so called because they use both majority and minority carriers. Minority carriers are unwanted leakage (ref Amos). Majority carriers (holes & electrons) are the bits you want. This is what is said about the name from some place on the interweb: "because the controlled current must go through two types of semiconductor material: P and N." I think this is right. - CPES 22:13, 27 April 2006 (UTC) (80.177.169.33)


I don't really understand what you say here. In a bipolar transistor, conduction is done using both type of carriers, electron and holes. Which is minority and which is majority only depends on the type of semiconductor material. I can't see any relation to the leakage current. Have I missed something? - CyrilB 18:48, 22 May 2006 (UTC)
It's too simplified. Actually, see, Bipolar_junction_transistor#Basics_of_transistor_operation, or how it's explained in Russian Wikipedia: Bipolar transistor:
In active mode, voltage is applied to transistor so, that its emitter junction is shifted in forward direction (is open) and collector junction is shifted in reverse direction. For clear explanation, let's consider NPN transistor; all the reasonings are valid for the case of PNP transistor with substituting the word "electrons" for "holes" (and conversely) and also reversing signs of all voltages. In NPN transistor, electrons, the majority carriers in emitter, pass through open emitter-base junction in the region of base. Part of this electrons recombinates with majority carriers in base (holes), part of electrons diffuses back into emitter. However, due to the base is made very thin and (very?) low doped, the main part of electrons, injected from emitter diffuses into region of collector. Strong electric field of reversly shifted collector junction captures electrons (note that they are minority carriers in the base, so the junction is open for them) and carries them in collector. So, collector current practically equals emitter current, except for small loss for recombination in the base, which produces base current (Ie=Ib+Ic). Coefficient α, relating emitter current and collector current (Ic=α Ie) is called "transfer coefficient of emitter current"(????!). The value of coefficient α equals 0.9 — 0.999; the more the coefficient the better the transistor. This coefficient poorly depends on voltages at collector-base and at base-emitter. So in the wide range of operating voltages collector current is directly proportional to base current, the coefficient of this proportionality being equal to β=α/(1-α)=(10-1000). Thus, changing low current of base the considerably larger collector current may be managed.
ellol 22:20, 22 May 2006 (UTC)

[edit] Semi conductor vs conductor

This original section is so wrong it ought to be simply deleted, but I'll leave it here for the moment. It attempts to be a hand-waving simplification. Is there a semiconductor or material physics section that this article coul simply point to.

[edit] Semi-Conductor vs. Conductor

On an atomic level, the difference between a conductor and a semi-conductor is in the number of electrons. Every atom has several orbitals, or paths that an electron follows, going around the nucleus of the atom. Each orbital has a certain maximum capacity of electrons that can follow that path. When an orbital that can only hold five electrons has four electrons and an electric charge gives it two more, it will try to pass one of them on to the next orbital or the next atom. A conductor atom has all of its orbitals full, making it pass on charges easily and without the loss of any electrons. A semi-conductor does not have all of its orbitals full which means that unless a certain minimum charge is sent through it, it will not pass the electrons to the next atom.

My alternative simplification would be that...

On an atomic level, the difference between insulator, semi-conductor and conductor is the energy required to promote an electron into the conduction band. In conductors, no energy is required, each atom is already contributng at least one electron to a "sea" of electrons which slosh about freely in the material. In insulators, a very high energy is required, typically the material has all electrons strongly bound to the atoms and unavailable to move. In semi-conductors, the band gap of the material is such that thermal energy is sufficient that some electrons become available for conduction. The inclusion of impurities or doping with atoms with a different valency provides sites which donate extra electrons or holes (a lack of an electron where one would be expected) to modify the conductivity of the material.

NeilUK 09:23, 6 June 2006 (UTC)

That's a good substitute. I've corrected one typo.
Atlant 13:38, 6 June 2006 (UTC)

[edit] Cost

It looks like this page has been trimmed, so I didn't want to add to it without discussion. I'd like to see a bit of information relating to cost. A quick Internet browse shows that from about 1955, the bulk price has dropped from about $5.50 per transistor to about $1.00 per 200 billion transistors. With that drop in price, there have been enormous gains in various efficiencies (switching speed, power consumption, frequency response, consistency of manufacture, etc.). Fracture98 06:45, 27 August 2006 (UTC)

I note that it may be removed becuase of monetary indifference. You can justify the entry by using How much is that?. We are using this website in RWC to make historical pricing relevant. Other additional odd references. meatclerk 17:48, 1 September 2006 (UTC)

[edit] Single molecule transistors

See http://www.photonics.com/content/news/2006/September/1/84283.aspx

[edit] Multiple emitter

Perhaps a section should be added on multiple emitter transistors, an important topic in electronics.

Ordinary Person 22:24, 1 October 2006 (UTC)

[edit] etymology

the article currently says transistor means both transfer resistor and transfer varistor. the more correct one appears to be the latter. it might be worthwhile to note in the article that many people believe transistor means transfer resistor and it isn't far from the truth since a varistor is a type of variable resistor. but might as well have wikipedia be accurate here. --Rmalloy 02:33, 6 October 2006 (UTC)

Actually, neither is quite right. Crystal Fire says
Brattain explained the problem and asked for advice. "Pierce knew that the point-contact amplifier was the dual of a vacuum tube, circuit-wise," he recalled. An electrical engineer, Pierce recognized the vacuum tube is a voltage-driven device, in which an input voltage controls the output current; in a point-contact amplifier, by contrast, an input current signal controls the output current. After thinking it over quietly for a moment, he observed that the relevant parameter of a vacuum tube was its "trans-conductance." Next, he mentioned the electrical dual of this property, or "trans-resistance." Then he put everything together, suddenly uttering a brand new word: "transistor." "Pierce, that is it!" exclaimed Brattain.
John Robinson Pierce also wrote an article on the naming of the transistor for the Proceedings of the IEEE, but I can't find mine right now; here's the abstract. He is also quoted on this PBS web page:
"The way I provided the name, was to think of what the device did. And at that time, it was supposed to be the dual of the vacuum tube. The vacuum tube had transconductance, so the transistor would have 'transresistance.' And the name should fit in with the names of other devices, such as varistor and thermistor. And. . . I suggested the name 'transistor.'"
I knew JRP for a long time before I knew he was the guy who named the transistor. It wasn't something he talked about, as he was always too busy with new ideas. Dicklyon 03:21, 6 October 2006 (UTC)
On the other hand, the BTL transistor naming ballot memo of May 28, 1948, which I believe is reproduced in JRP's article in Proc. IEEE, and which I happen to have a scan of from some place on the web, says
Transistor. This is an abbreviated combination of the words "transconductance" or "transfer" and "varistor". The device logically belongs in the varistor family, and has the transconductance or transfer impedance of a device having gain, so that this combination is descriptive.
So people who cast a vote for that name with this memo probably did think it meant something like transfer varistor. But if we're going to say that, we should probably tell a more complete story. Dicklyon 03:38, 6 October 2006 (UTC)
Here you can find the transistor name memo/ballot Dicklyon 03:41, 6 October 2006 (UTC)
Oops, just noticed it's already linked and quoted in the article. I added the JRP quote and PBS link after it. Dicklyon 05:45, 6 October 2006 (UTC)

[edit] Types of transistors

This section is becoming more a laundry list than an informative and useful bit of prose. I don't see much utility in enumerating every major variation of transistor. I think it ought to be trimmed way down and just cover the three basic classes; FET, bipolar, and experimental. -- mattb @ 2006-10-20T03:06Z

Good point. KIS is always best. CPES 23:28, 25 October 2006 (UTC)

[edit] Photo transistors

Now that the weird "photo fet" has been pointed out, we really need a section on photo-transistors to explain a bit about how these things work, and why the "photo fet" is a sort of bastard stepchild of a photo-transistor (OK, maybe not in those terms). Here's one explanation. Saying the light makes the FET conduct is bit too misleading, in my opinion. Dicklyon 15:58, 30 October 2006 (UTC)

Here's another that explains the photo-fet as a combination photo-diode and fet. Dicklyon 16:00, 30 October 2006 (UTC)

Gracious that second book looks pretty old... The wording makes it seem like it was written before the interaction of photons with semiconductors was well understood. Anyway, as I've seen the term used, "phototransistor" can include a lot of fairly different devices. The only real constant is that there's some sort of quantum interaction between photons and EHPs going on in a semiconductor. Beyond that you could be using a heterojunction system, a MOS system, a traditional bipolar system, etc. That's the reason I objected to the phrasing of the added text; it suggested to me that phototransistors all use photon-related RG current in exactly the same way, which isn't true. -- mattb @ 2006-10-30T16:16Z

Yeah, 1989; ancient history; oops, no, I made a light-sensing NMOS chip in 1980; Chamberlain did in 1968; of course, the photosensitive parts were the isolation diodes, not the transistors. My point is that in the photo-fet, the light doesn't make the transistor conduct, or not in a way analogous to the inherent transistor action on photocurrent that you get in a BJT. A few words to explain the two, and their difference, could be useful. Dicklyon 16:22, 30 October 2006 (UTC)

[edit] Reason for reversion

The following two paras were reverted out of the tran page with the edit comment below:

Photo BJT

"Bipolar transistors can be made to conduct by light (photons) as well as current. Devices designed for this purpose have a transparent window in the package and are called phototransistors."

Photo FET

"Like bipolar transistors, FETs can be made to conduct by light (photons) as well as voltage. Devices designed for this purpose have a transparent window in the package and are called phototransistors."

Edit Comment

"phototransistors don't "conduct" photons (or voltage, for that matter); the photons cause RG current... Plus the effect isn't limited to bipolar transistors"

I can't understand the edit comment (putting aside the photo FET issue for the moment). Neither paragraph states that phototransistors "conduct" photons or voltage and neither paragraph says that the effect is limited to bipolar transistors... or am I missing something? CPES 02:56, 31 October 2006 (UTC)

I didn't notice the "by" in there when I reverted. My bad. Anyway, see Dick's comments one section above. -- mattb @ 2006-10-31T03:04Z
Oh, I see. Its been driving me mad trying to figure it out. Thanks CPES 03:13, 31 October 2006 (UTC)

[edit] Revised wording: photo BJT and photo FET

Well put (I think it is important to cover opto aspects) CPES 10:42, 1 November 2006 (UTC)

I've never heard "Photo BJT". Let's don't use it without a good source. Dicklyon 21:43, 1 November 2006 (UTC)
Photo BJT in this case is just used to differentiate between photo aspects of BJT and photo aspects of FET(photo transistor could be interpreted to cover both BJT and FET). I meant that it is important to cover the opto aspects on the tran page for both BJTs and FETs, rather than say nothing, and that the revised wording is all that is needed. I was not suggesting that any new information on opto be added to the tran page. CPES 08:50, 2 November 2006 (UTC)

[edit] Interesting! Evidence of a Bell Labs' coverup

In 1981 the semiconductor physicist H. E. Stockman said "Lilienfeld demonstrated his remarkable tubeless radio receiver on many occasions, but God help a fellow who at that time threatened the reign of the tube." See Bell Labs Memorial: Who really invented the transistor?, starting at "Oscillating Crystals".

Here's a paper which details some history of the laboratory testing of Lilienfeld's patent claims by others: The Other Transistor: early history of the MOSFET See pp235-236

Briefly:

In 1964 a physicist V. Bottom asked in Physics Today magazine whether these transistors worked, and J. B. Johnson of Bell Labs responded saying that he'd tested them and they didn't work. This probably is the origin of the story that Lilienfeld never had any working hardware.

Then in 1995 R. G. Arns found a 1948 legal deposition by Johnson which said the opposite: that Bell Labs back then had a project to test Lilienfeld's transistors, and before Johnson took over the project, Shockely and Pearson had built a variation of Lilienfeld's aluminum oxide MOSFET from his patent and found only an 11% modulation index, but that "useful power output is substantial"! To me it appears as if Johnson, being with Bell Labs, perhaps had an agenda to promote his own company's discovery while misleading the physics community about Lilienfeld's. After Shockley/Pearson's success, Johnson had tested the other two Lilienfeld patents and was unable to replicate them ...so Johnson was only dishonest in his covering up the fact that Bell Labs well knew that Lilienfeld had something real. Between these times B. Crawford in 1991 built successful but unstable Lilienfeld MOSFETs and saw evidence that Lilienfeld had done the same. In 1995 J. Ross built stable Lilienfeld MOSFETs. In addition to all this, a 1934 patent by Oskar Heil exists for another MOSFET.

The author makes a very telling statement about the honesty of Shockley et. al.: "Published scientific, technical, and historical papers by these Bell scientists never mention either Lilienfeld’s or Heil’s prior work." --Wjbeaty 03:13, 1 November 2006 (UTC)

I had added that Arns ref to the John B. Johnson article a while back. It's an interesting story, but doesn't really get close to being conclusive about whether Lilienfeld ever made it work. As for the coverup, that was the Bell way with lots of tech advances, not specific to Shockley. I can provide a copy if anyone wants and doesn't have the access. Dicklyon 04:47, 1 November 2006 (UTC)

[edit] Pictures

Can we get some close-up pictures, like on [2]? We could ask manufacturers, as they did. — Omegatron 22:25, 8 November 2006 (UTC)

I guess we could, though I have my doubts about the utility of such things... Most optical photographs of functional transistors (including the ones you linked, I think) are just going to show the top metal layers or contacts. SEM photographs might be slightly more interesting. -- mattb @ 2006-11-09T00:07Z
P.S. - I can take plenty of optical microphotos of various forms of transistors... Getting an SEM image might be more difficult, though. -- mattb @ 2006-11-09T00:10Z
Most optical photographs of functional transistors (including the ones you linked, I think) are just going to show the top metal layers or contacts.
So what? Most people can never see even that. — Omegatron 16:45, 8 December 2006 (UTC)

[edit] Substantial gain

Someone asked in the article "what doees substantial gain mean"? I believe the answer is that it means gain that is greater than 1 by an amount sufficient to convince the experimenters that they are observing actual active amplification. It can be voltage ratio or a current ratio or a power ratio, but in general gain is the ratio of output to input, and you don't get gain greater than 1 from electronic components that are passive, or not active. Dicklyon 21:26, 12 November 2006 (UTC)

Perhaps the word "substantial" should be removed then. It's ambiguous and understandably leaves the reader wondering, "how substantial"? -- mattb @ 2006-11-12T21:29Z
What you put in its place to indicate "enough gain that we were sure it was working"? Dicklyon 21:35, 12 November 2006 (UTC)
How about "above-unity gain"? A number of things would work better than "substantial". -- mattb @ 2006-11-12T21:37Z
Or you might want to quote John B. Johnson who used the term "substantial" in decribing his experiment on making a Lilienfeld device. See the section on field effect transistor. Dicklyon 21:38, 12 November 2006 (UTC)
Whatever, if you really want to keep the word I won't make any fuss. I personally believe it is not the best word to use here because it can connotate a large quantity rather than simply meaning considerable/significant/not trivial (though these, too sometimes have the "large" connotation) depending on how the reader takes it. However, I'm not currently on a mission to rewrite this article (my current mission is Computer if you want to take a look), so I'm not too interested in diction debates. :) -- mattb @ 2006-11-12T21:58Z

[edit] Makers

Matt, good job deleting the spam-magnet list of makers. I was tempted, but not as bold. Dicklyon 05:03, 30 November 2006 (UTC)

Thanks for the word of support... I had the evil idea of adding my own lab to the list if someone re-added it. (just kidding) -- mattb @ 2006-11-30T05:42Z

[edit] Needs Simpler Explanations

Not everyone who comes to this article is as technically informed as the authors. It's important that this and other technical and scientific articles include basic information and explanations for the less informed. After all, that's what they're here for---to get informed. Tmangray 08:10, 23 December 2006 (UTC)

Exactly what is unclear? Is the lead paragraph OK? If not, try to fix it. If something later is unclear, then please point it out or try to fix it. The complexity is bound to increase beyond what nontechnical readers can handle at some point, but if you point out at what point it needs simplification, perhaps it can be improved. And of course we will all agree, I think, that it is inevitable than in an article on a technical topic "Not everyone who comes to this article is as technically informed as the authors." That doesn't mean we're not sympathetic to the problem. Dicklyon 08:46, 23 December 2006 (UTC)
The intro looks fine to me. — Omegatron 06:19, 25 December 2006 (UTC)
I thought so, too. But then I looked again, with an eye to whether you and I are more part of the problem than of the solution. Perhaps so. So I attempted a simplification, at least by moving a few of the less "simple" concepts to a second paragraph. See what you think; hopefully I didn't lose anything. Dicklyon 07:05, 25 December 2006 (UTC)
The OP's concern is a lot easier to voice than to solve. I've tried for a while to think of a one-liner that simply and adequately explains what a transistor is and/or does. I don't really think there is such a thing. Transistors can be simple devices, but the principles that govern their operation are not readily accessible to the lay man (high school level electromagnetics are a minimum requirement, basic quantum mechanics and solid-state physics are preferable). Complicating things further is the fact that devices which work in very different ways (say, a BJT and a FET) are all lumped into the category of "transistor". The lead of this article as it currently stands is decent, but I'm not too sure if I like the whole "small signal controlling larger signal" general assertion. I think that this is usually true from an application point of view, but I can think of counterexamples (e.g. follower stages in amplifiers). It's on the right track, but I think it could be improved some. I'll look back through some of my semiconductor physics textbooks to see if they offer any decent terse "definition" of a transistor (I'm doubtful). -- mattb @ 2006-12-25T22:12Z
No such luck... All of my texts avoid actually defining a transistor and just start talking about how they work. I looked at a copy of Britannica, and it adopts a purely functional definition of the transistor. Rather than attempting to explain at the outset "what transistors do", it starts with "what they are used for". I think this might be a very good approach. Pursuant to this methodology, I think an adaptation of the first sentence in the second paragraph might be good. Let's try to think of the most general categories of application transistors can be used for... Off the top of my head, I'm coming up with signal amplification and control (switching, voltage regulating). I'm a little hesitant about including oscillators so explicitly, since an oscillator is much more a certain circuit topology than a device. I realize that this is somewhat an artificial distinction that could just as well be applied to any use of a transistor, but I think it's much easier to make the argument that a transistor can, on its own, behave as an amplifier or a switch than an oscillator. Anyway, let me know what you think. -- mattb @ 2006-12-25T22:34Z
I think the concept of a small signal controlling a larger one is fundamental, and it certainly does apply to a follower, which is a current amplifier. Plenty of good simple definitions are available to draw on if we need more inspiration: [3]. Dicklyon 02:06, 26 December 2006 (UTC)
It's a controlled valve. — Omegatron 07:43, 26 December 2006 (UTC)
It's a solid-state alternative to a controlled valve (using the British meaning of valve). Dicklyon 16:16, 26 December 2006 (UTC)
You could go that route, but I don't think it's a particularly useful comparison for introductory purposes. Someone who doesn't have a good grasp of transistors will not likely understand valves (tubes) well enough to benefit from the one-liner. If we're talking about the hydraulic analogy of circuits, yick... Let's not go there, please. I'm still not entirely sold on the "small signal controlling larger signal" statement. I certainly see the logic behind saying that, but I'm not convinced that it's the best way of summarizing transistor operation. I don't think it neatly covers the switching role of transistors, and I still think it's awkward to apply it to unity or sub-unity amp stages. I feel that a terse statement to the effect that "transistors are semiconductor devices used for electrical signal amplification and control" would be a nice way to cover all the bases. If you folks really want to stick with the small signal/large signal explanation, however, I guess that's okay too. -- mattb @ 2006-12-27T21:12Z
Retrieved from "http://en.wikipedia.org../../../t/r/a/Talk%7ETransistor_a9af.html"