Graham Slee写的一篇关于运放和分立晶体管电路的文章

小白

交易区新帖推荐

Op-Amps versus Discrete Transistors    运放和分立晶体管
25 February 2013 - by Graham Slee


Transistors are cheap. On average, small signal transistors are 2 cents (1.5 pence, 2.5 euro-cents) each. Production quantities are $20 per thousand - a pocket money investment.   晶体管元件很便宜。平均来说小信号晶体管2分钱一个（1.5便士，2.5欧元里的分）。20元批量做一千个——这种投入是口袋里的零钱级别的。

Building hi-fi circuits using transistors is labour intensive but you sub it to a Chinese assembly shop and your circuit will often cost less than a single op-amp.  用晶体管元件来制造hifi线路是一项劳动密集型的活动，不过，你只要把活儿交给中国装配厂，然后你的电路成本比运放更便宜。

Bolt your Chinese made transistor boards in your Chinese made case and you can still announce your product is made in Britain, the USA, or wherever your company is.  把你的中国装配的电路板装到中国制造的机壳里，你仍然可以声称你的产品是大英帝国制造的，美国制造的，或者随便什么地方制造的。

Your Hi-Fi company is hungry for money so what explanation do you give your customers? You tell them how superior your discrete transistor circuits are - how they're superior to 'cheap' op-amps. You get your customers to look down their noses at op-amps and anybody who uses them - but in truth you're misleading people.  你的hifi公司对金钱很渴望，所以你怎么向你的顾客们解释呢？你告诉他们，你的分立元件的电路非常优越——比廉价的运放要优越。你要让你的顾客们瞧不起运放以及使用运放的厂家。但实质上这是在误导人。

If you know anything about electronics at all, you know that discrete transistors have a wide matching tolerance which means the performance of your products will vary widely - unless you use lashings of negative feedback... but your ad says zero-negative feedback...   假如你懂一点电子，你应该知道分立的晶体管元件有很宽的匹配误差，这意味着你的产品性能会有一个较宽的波动范围——除非你用大量的负反馈。但你的广告里说“零负反馈”。

So you hand select the product you send for review. Your customers get what's left over...   所以你精选出一台来，发给杂志去评测。余下的卖给你的顾客。

Dear customer, are you still convinced discrete circuits are superior to op-amps?   亲爱的顾客们，你还觉得分立元件比运放优越吗？

In 1965 I had my first adventure in electronics - I realised how exact things had to be to work, never mind how well they worked. I was only 10 years old.   我是在1965年初次进入无线电电子之门的——我认识到工作时元件必须多么精确，不管它们的表现有多么好。那时候我才10岁。

In a couple of year’s time my electronics experiences will have spanned 50 years! A lot of the companies tempting you for your money are operated by people with little or no experience in electronics (often a Chinese man does it all for them), yet they make incredible claims that sound so convincing - perhaps they're gifted in that way? 再过2年我的电子学方面的经验就要跨过50年门槛了。有很多瞄准你口袋里钱包的公司，其背后的人士，几乎根本不懂电子，有时就是一个中国人帮他们把活儿全包了。但这些不懂的家伙说起话来却似乎很有说服力——也许那就是他们的天赋吧。

In these nearly 50 years I've witnessed the virtual end of the valve and the beginning of the 'tinny' germanium transistor. The improvement the silicon bipolar transistor made, the commercial use of field-effect-transistors (FETs and MOSFETs) and the introduction of operational amplifiers (op-amps).  在这仅50年里，我亲眼目睹了电子管时代的终结、小小晶体管的起步、二极管带来的进步、场效应管的商业应用、以及运放的引入。

You may find it surprising to know that the first ever op-amp was a circuit board containing quite a few valves (tubes). Silicon bipolar transistors later replaced the valves and the circuit boards of op-amps became smaller.  也许你会觉得很惊讶，最早的运放是一块包含了几个电子管的线路板。之后二极管取代了电子管，然后运放的线路板变得越来越小。

The performance breakthrough happened when mankind found how to etch and dope those transistors, passive components and interconnections onto a wafer of silicon. The modern hermetically sealed chip op-amp was born and with precision laser trimming repeatability of results was assured - a far cry from the matching difficulties you’d have with discrete transistors.  技术突破发生在人们掌握了如何把晶体管、被动元件和连线蚀刻和掺进硅晶片的技术时。于是现代的封装运放诞生了，精密的激光技术保证了产品制造的可靠性和一致性，分立元件线路时遭遇的匹配困难没有了。

It also meant that wider bandwidths could be accommodated due to the reduction of circuit capacitance - much more stable (and hence reliable) than boards made with discrete transistors. 同时，线路电容量的减少也意味着可以适应更宽的频带。稳定性和可靠性比分立晶体管线路更好。

The op-amp gave real electronic designers the opportunity of making better circuits and in the case of high fidelity equipment, the awesome possibilities of incredible sound quality and amazing stereo imagery.  运放给了电子设计师制作更好线路的机会，对hi-fi来说，这意味着更好的音质、更佳的立体声结像。

It is often a case that those who can - do, and those who cannot go into sales. There are a lot of 'cannots' in hi-fi judging by my experience.  于是那些做不好、做不来的家伙，开始进入销售领域了。从我的经验来看，在hi-fi圈里这样的人很多。

One of the false sales claims from those 'who cannot', is that op-amps are ready-made amplifier stages whose performance is decided by the op-amp manufacturer. This is the farthest from the truth one can get!   那些做不出好东西的人，所宣称的错误观念之一是，运放是预先做好的放大级，其音频性能是运放制造厂所决定的。其实这个观念和真相差了十万八千里。

They base the above argument on high school teaching which in recent years has seen the op-amp on the syllabus. High school subjects are nowhere near advanced and are extremely basic.  他们的这个错误观念是基于高中教科书的，而高中教科书是非常非常基础的东西。

You can hazard a guess at the ages of these so called 'brilliant designer's' from what I just said. Their employers may be older and more experienced... but in sales!  你可以根据我前面所说的猜出那些所谓“天才设计师”的年龄。他们公司确实有些年纪更大、更有经验的人士——但是在销售部门！

There is more to designing with op-amps than most post graduates would have the time for. The amount of advanced educational material from op-amp manufacturers is colossal.  用运放来设计器材，这其中的内涵，大多数的学校毕业生没有时间去掌握。运放制造厂家们有着大量的进阶技术资料。

Apart from being superior in performance to an open (or even potted) board full of discrete transistors, the op-amp can be utilised in many different ways to those most students would know. The understanding of how op-amps work requires an advanced understanding of how discrete transistors work and that's just for the first course!  运放的性能优于分立元件的线路板，除此之外，运放可以以多种方式使用。要理解运放的工作方式，首先必须理解分立晶体管是如何工作的，而这还只是第一课！

Often op-amps can be 'tamed' to have the same flat open-loop frequency response of valves (something I use in a number of my Ultra-Linear designs).  运放经常可以被“驯服”为拥有类似电子管的平直开环频响（我在Solo超线性耳放中利用了这一特性）。

Most hi-fi designers have no knowledge that single op-amps have more than two inputs, let alone being able to tell the difference between noise gain and signal gain and its effect on high frequency stability - no wonder they hide behind their simplistic transistor designs.  大多数hi-fi设计师并不了解单片的运放有多于两组输入，更不知道“噪声增益”和“信号增益”之间的差异，以及这一差异对高频稳定性的影响——于是他们躲回简单的分立晶体管元件设计背后。

But are they simplistic really? Many discrete transistor designs I've seen are direct copies of op-amp schematics! Op-amps are so bad they copy them???  然而，真的很简单吗？很多我见过的分立晶体管的设计是照抄了运放的线路。运放如果很糟糕，他们为什么要抄它？？？

I'll leave you with this thought: one major criticism levelled at op-amps is their internal circuit complexity. Out of the most 'complex' op-amp circuits I have ever seen, there are only five or six stages in the 'signal path' but most are four.  针对运放的一个主要批评意见是运放内部线路很复杂。然而我所见过的最复杂的运放线路，在信号通道上只有五级或六级，大多数是四级。

The simplest discrete transistor stage I have seen that can be considered to be compatible with input and output requirements has three stages, but mostly they're four, five or six!  我所见过的最简单的能满足输入和输出要求的分立元件的放大线路，有三级，但大多数是四级、五级或六级！

The rest of the op-amp features these 'failed' designers cannot understand are the op-amp's power supply and power conditioning circuitry. Yes, they come with their own precision voltage references and power supply noise rejection circuits!  运放线路中所包含的其余功能，是供电和稳压线路，但那些“不懂的设计师们”不知道这个。是的，运放有自身的精密电压标准和供电降噪线路！

So what is the op-amp versus discrete transistor argument all about? Arrogance maybe?  所以，运放和分立元件的这个争辩的本质是什么呢？也许是傲慢？

[ 本帖最后由 小白 于 2013-5-12 09:09 编辑 ]

小白

挺有意思的，有空了我慢慢翻译一下。读得懂的，就读读吧。也可讨论下这个问题。

wangjianqiang69

还是白版翻译一下吧，英语都还给老师了

小白

大致意思就是很多厂家在那里宣传分立晶体管元件的设计更优越，很多烧友在这种宣传的影响下看不起运放设计，认为分立元件的设计天生更好，其实这个看法是没有根据的。运放和分立元件这两种线路形式，没有高下之分。

headfan

有段看得笑喷了

落下一颗糖

中国人又被吐槽了

KEPER

表示老头很幽默

8492

单纯从元件匹配来说的话是，但是功放等不仅仅是原件匹配。

老丹还死活认为MOSFET做不出好声音呢。

palchiu

一堆hifi厂中枪

蛋疼人士

都有谁中枪了？

紫气东来

来学习

激光鼠

偶对运放深恶痛绝，一台放大器里面只要用了运放就很难出好声。你可以说是没用好运放，但一个能用好运放的设计师，就如Slee说的那样，必然对晶体管的工作原理有着深刻的理解，可以用分立电路做出声音更好的放大器。

xiaobudian

鼠版亲身经历才是真相哈~

yukuai11

等翻译，白版翻得比原文更清楚

小白

我这里备注一下，Graham Slee本人并不排斥分立元件，比如他的Novo就是分立元件的，他还跟我说过Solo耳放也完全可以用分立元件来做。他的观点是分立元件并不具有针对运放的“天生优势”。用运放、用分立元件，都可以做出很好的器材。