Talk:Sziklai pair

Pushing and Pulling
Does anyone understand all this talk about pushing and pulling? I think the writer needs to re-think their words!--Kisss256 06:57, 31 March 2010 (UTC)

Yes, and you're right, the idea that the reason to use CFP pairs on both sides of the rail is to improve matching is false. Douglas Self's book Audio Power Amplifier Design goes into great length about why one would choose CFP's (Sziklai) pairs for both rails of a pull pull amp and it ultimately comes down to the fact that no matter how you slice it, if you don't need parallel output devices, CFP stages are significantly more linear, more power efficient, easier to stabilize thermally and produce fewer crossover harmonics. If you need parallel devices, some of those things go away and you're stuck with emitter followers if you want the best distortion performance. There's also the fact that CFP's are a little harder to stabilize. It's a small consideration that an be worked out in a simulation quite easily but it does scare beginners away from them. I've been using them in my < 100W amps for years and the fact that I can make a very low distortion amp that idles under 2W and can cooled very easily (without sticking your drivers/bias servo/diodes to the heatsink) keeps bringing me back to them. 142.68.70.169 (talk) 23:05, 26 March 2016 (UTC)

Pronunciation
Does anybody know the pronunciation of `Sziklai' ? `zIk-lay'? `tseek-lie'? Or something else? Thanks for help. —Preceding unsigned comment added by Tosendo (talk • contribs) 21:42, 2 December 2007 (UTC)
 * If I'm not mistaken, I guess it's something like seek-loy but better ask somebody who knows hungarian.89.136.39.53 (talk) 22:18, 6 March 2009 (UTC)Apass
 * On thinking again, I guess it's more like sick-loy (with a short e)89.136.41.31 (talk) 12:11, 11 March 2009 (UTC)Apass

Is it really true that PNP transistors were historically more scarce?
I have quite a few books from the times when germanium transistors were in their golden days (well, at least in russia and in the eastern europe) and I can see that PNP were the kings of the day, very few circuit diagrams were using germanium NPN transistors. Anyway, even if this was not the case for germanium transistors, I still think that a citation is needed for this affirmation: To reduce the number of PNP transistors, which historically have been more scarce than NPN transistors.89.136.39.53 (talk) 22:32, 6 March 2009 (UTC)Apass


 * Yes, I always thought it was the other way around. Definitely needs a cite.
 * EDIT: re-reading it, I think the author meant power PNP's we scarce (or just less common). -Roger (talk) 03:33, 7 March 2009 (UTC)
 * I don't know about that. I have a brief catalog page and I could find only PNP power germanium transistors (but like I said, it's only a brief catalog). The thing is that since the majority of the low power transistors seemed to be PNPs, there is no reason for the power transistors to prefer NPNs.89.136.39.53 (talk) 08:48, 7 March 2009 (UTC)Apass

From memory, while here in the West, while PNP transistors were more common in the Germanium era, as soon as the take over of Silicon Planar transistors came about, this changed. Scarce is probably the wrong word, anyway; PNP power transistors were more expensive and in less common use than NPN ones, and the selection was less good, so for power transistors, scarce might be justifiable. However, for general-purpose small-signal transistors, there was probably a similar availability of type numbers in PNP and NPN and the cost was very similar (given that packaging was a significant part of the cost for small-signal discrete transistors, the purchase price costs are going to be similar). I think 'in less common use' would be better, but you could argue for 'more expensive' or 'lower performance' or 'more difficult to obtain' if what was being discussed was just silicon power transistors. —Preceding unsigned comment added by 212.84.126.62 (talk) 10:03, 7 May 2009 (UTC)
 * I agree that expensive is less confusing than scarce. That being said, the reason why they are more expensive is because they (or at least their constituents) are scarce. From what I remember, because the hole mobility is so much lower than the electron mobility, production of a PNP transistor to match the performance of an NPN transistor requires a greater amount of resources. These additional resources translate to a greater sensitivity to scarcity. Hence, scarce makes sense, but expensive is just as accurate and is less confusing. I've changed the reference to scarce. All of that being said, is the reason for the scarcity due to the asymmetric mobilities? &mdash;TedPavlic (talk) 18:45, 7 May 2009 (UTC)

Early transistors (signal and power) were PNP only. The popularity of the Sziklai connection came about because of the need to match the characteristics of the output transistors to keep distortion figures down. In the early days of NPN power transistors, their PNP "equivalents" weren't a good complementary match at all, and so the classic complementary pair configuration invariably ran a bit lobsided. One way around this was to use all NPN (or PNP, but it's almost always NPN) output transistors and a Sziklai connection. Modern NPN/PNP complentary pairs are a much closer match than they used to be, but they're still a long way from being perfectly complementary. Midgemagnet (talk) 01:33, 18 May 2009 (UTC)
 * Ok. Please see the recent changes, which attempt to synthesize this whole discussion. &mdash;TedPavlic (talk) 20:18, 18 May 2009 (UTC)

Matching transistor output characteristics

In the article it states that this structure is used "when the designer wants to use devices of the same type (e.g., NPN), instead of complementary types, which rarely match accurately." I assume that refers to the power transistors in the complimentary pair. It then says "the designer uses Sziklai pairs for both the upper push pair and the lower pull pair." That would defeat the goal of trying to use output devices of the same type, as you'd still have to use power output transistors of opposite polarity in each Sziklai pair. If you want to match NPN to NPN you have to use a standard Darlington as the upper pair and a Sziklai as the lower. At least that's the way I've always seen it done. For example, the power amplifier circuits here: http://www.angelfire.com/planet/funwithtransistors/Amp-400_W.html

"Sziklai Pair" or "Sziklai pair"?
The title and body are inconsistent. Should it be "Sziklai Pair" or "Sziklai pair"? --Mortense (talk) 22:18, 23 September 2021 (UTC)


 * Fixed - somewhat belatedly. 81.154.179.214 (talk) 09:27, 6 February 2022 (UTC)

Is "Sziklai pair" a case of misattribution?
The earliest mention of Sziklai that I can find in connection with the circuit shown in the main article is in Horowitz and Hill. All through my engineering education at MIT in the 1980s, the connection was simply referred to as a complementary Darlington, or something similar, and Sziklai's name didn't become prominently attached to the circuit until later. Earlier versions of the wikipedia article referenced two patents by Sziklai in support of the nomenclature, but conspicuously, neither of those patents show or describe the circuit at all. I've looked reasonably hard, and have been unable to find a primary source for the nomenclature, which makes me suspicious that H&H might have made a spurious attribution, and that the popularity of that wonderful book caused the misattribution to propagate without any further checking for correctness. It would strengthen the article greatly if an actual primary reference could be found and cited. Otherwise, perhaps the article should be rewritten to acknowledge the possibility of a misattribution.
 * I found a Google Books snippet dated 1985, "Semiconductor Circuit Approximations" by Albert Paul Malvino, page 259 that mentions Sziklai pairs. Of course it's only a snippet so I can't see the whole thing but this pre-dates at least the second edition of Horowitz adn Hill. I'd check IEEE's site but I've found that no matter how many societies you join, the reference you want is never in your subscription. --Wtshymanski (talk) 21:28, 20 May 2022 (UTC)


 * George Sziklai shows an application of the circuit in his 1953 paper |"Symmetrical properties of Transistors and Their Applications": a "two-stage direct-coupled Class-B amplifier" in figure 8. In other words, two Sziklai pairs drive a loudspeaker. Each stage is a Sziklai pair, one a PNP pair, and the other a NPN pair. The paper came about one year after Darlington's patent. 181.225.181.225 (talk) 23:53, 17 December 2022 (UTC)

General question about citations.
I apologize if this is the wrong forum for this topic. I wasn't sure that there's a specific page for things like this.

Anyway, I was wondering, in regards to citations from a book, how are you to determine where exactly in the book the citation comes from? It seems a page number should be included to facilitate easy look-up.

In my particular case, I wanted to look up the first citation from The Art of Electronics by Horrowitz and Hill, since I have a copy. But with no page number or additional information included, this wasn't as straight-forward as expected.

I know you could use the index in the book to find certain topics. But that's not always going to be a viable solution.

So, I'm just wondering if I'm missing something fundamental here. Or if my observations are correct. I'm also hoping to get some advice on how to better consult book-based citations such as the example provided.

Thanks in advance for your time and advice. Cheers! VoidHalo (talk) 23:44, 8 November 2023 (UTC)