β22放大器制作全攻略 要的就是完成度和原汁原味 与各位DIY分享 全文更新完毕

rabbitgirl

ygw70

这样的好帖不顶对不起楼主。

haoxin

这样的好帖不顶不行啊

swsw4321

电路太一般了,这种形式的电路音质就这样了.

燃烧我的罪恶

首先感谢millwood热心提出的宝贵意见

我菜菜的翻译了下有关技术的，方便更多人看。

原帖由 millwood 于 2008-12-15 00:08 发表



how you want to modify it depends a lot on what you want with this amp and how you approach audio in general.
一般来讲 有很多种方法来修改和提高这个放大器

just some obvious design considerations, based on the first schematic in post #2:

下面是一些关于二楼图纸的一些明显的设计考虑

1) you should never use a mosfet without gate stoppers. R8=R26=11 - 470ohm based on your mosfet and for this particular pair, you can go lower but not having one is asking for trouble.
使用MOSFET必须使用gate stoppers,在这个制作中 mosfet 使用R8=R26=11 - 470ohm，除非比较殊的，用很小的但是自找麻烦  

2) R14/R21: unless you want to use this as a power amp, use higher R14/R21 to mminimize distortion andn provide better thermal stability. 4.7 - 47ohm would be a good starting point.

对于R14/R21，除非你想将这个做为一个大功率放大器，使用更大的R14/R21来降低失真，并提高热稳定性，4.7～47ohm 比较合适。

3) C6/C7/C12/C13 are on the small side. but keep them if your amp isn't oscillating.

C6/C7/C12/C13取值小了，如果放大器不自激就可以保持不变

4) Q12/Q13: the use of BJTs to compensate mosfet output devices tends to result in overcompensation. I would use a Vgs multiplier here -  replace those transistors with the same transistors as your output device.
Q12/Q13是使用的三极管来补偿输出用的的MOSFET，但容易导致过渡补偿，我会在这里使用Vgs multiplier在这里，用你输出级的MOSFET来替换这些器件

5) R35/R34: this is not a CFB amp and it doesn't benefit from absurdly small feedback resistors. not to mention those astronomical gain numbers for a headphone amp.
对于R35/R34 该机并不是一个CFB amp，如此小的 反馈电阻是荒谬的 得不到好处。对于耳机放大器来说是不需要的


6) C19: don't use C19 unless you have exhausted anything else.

不要使用C19，除非你不得不用

7) R18: too large. it doesn't nothing but make the amp susciptible to interference and introduces thermal noise. for perfect offset cancellation, R18=R34//R35 but since you are using jfets, that may not be as big of an issue here. I would make sure that it is less than 47k.

R18过于大，容易让放大器受到干扰和引入噪声，最佳的大小应该是R18=R34//R35 ，但是这里使用的JEFET，所以可能不够大，我确认这里至少要使用47K

8) R9/R10/R19/R24: a little bit high, especially for jfets but shouldn't be a big deal.

对于R9/R10/R19/R24 稍微大了点，虽然这是可以针对JFET

9) C2/C16: make sure you use a small and fast capacitor. I typically use polyester caps here.
C2/C16 确保使用的是小容量和充放快速的电容，我推荐使用聚酯电容



...

燃烧我的罪恶

millwood

CFB amp 全名是什么啊 不清楚。

关于技术和声音的设计理念 每个人理解和审美情趣都不相同

您可能不知道 我曾经是PASS的FANS

我做过PASS 的 ZEN4 和 ALEPH 5 嘻嘻

只能说追求不多 各有各的特点和风格

millwood

mosfets have significant advantages over bjts for high power applications.

however, they are far less complimentary than bjts and they produce more distortions of higher order harmonics.

in a headphone amp, you need less than .3w so a bjt (bd139/140 for example) would be a far better choice than those mosfets, especially one with Ids of 17amp (irfz24).

millwood

a better designed amp is the Gilmore headphone amp which can also be used to drive speaker phones too.

Other than the servo, it doesn't have many of the issues of b22.

millwood

燃烧我的罪恶 , cfb means current feedback. most opamps are "voltage feedback types". modern high speed opamps tend to be "current feedback" opamps.

"只能说追求不多 各有各的特点和风格": that's correct. Pass focuses on making the best sounding amp. b22 was designed to have the best specs.

hz_zhc

上面这个K.G的线路成熟稳定，一装就成。没那么多其它的问题，非常好声。晚点贴个我做好的机。

tyyj

原帖由 swsw4321 于 2008-12-15 19:21 发表
电路太一般了,这种形式的电路音质就这样了.

你可以把你的高端电路摆出来。。。像那位说英文的老兄的态度是值得学习的，虽然说得不是很正确，或者各人的理解不一致，但也是正常的心得交流，也能让我们互相得到提高。如果你真有压箱之作，也可以让我们先一饱眼福，有机会也能一饱耳福。

tyyj

millwood，看出来了，你是走美声路线的。

tyyj

技术路线的走向也代表了各人声音的取向不同。。。。。发烧也二十多年了。。一开始也是走美声路线，怎么顺耳怎么调电路。。。但是随着走入大城市，经常接触到现场音乐，就慢慢改变了原来的思路，到底声音应该怎么样。。现场给了我很大震撼。。。音乐的真实自然，才是它魅力所在。

millwood

the KG amp is a sound and proven design. the biggest issues with it are a) limited range for the servo; and b) stability.

1) the original design requires very precisely matched transistors (<1%). Otherwise, the servo will not be able to null the output DC. to cure that,  you will need to change one of the 500ohm resistors in the constant current source into a 1k pot and use it to trim the DC offset to as close to zero as possible.

After that, you may want to lower the 15K resistor on the output of the servo to expand the servo's range. It can be as low as being shorted.

2) the amp doesn't have any miller capacitor so it is likely to oscillate with high speed transistors. I would start with two 220p capacitors and lower them until the oscillation goes away.

3) you will need to groundn the input to reduce interference. a 22k resistor there is a good starting point.

gaobingic

好！好好！好好好！。。。能把PCBLayout上传就更好了！大家DIY！好好！

millwood

音乐的真实自然，才是它魅力所在。

that's absolutely true.

it is impossible to reproduce live music, in my experience.

in my earlier years, I have designed and built a lot of those "techno" amps with tricked out circuitry that will measure far better than anything Pass has to offer. decades later, I am coming back to simple amps because they sound better.

millwood

原帖由 tyyj 于 2008-12-15 20:54 发表
millwood，看出来了，你是走美声路线的。

yes and no. the only reason I am into amplification is for good sound.

however, I cannot stand a technically incorrect amp (amps that oscillate, or opamps without supply bypassing, etc.).

jt1000027dsl

高手太多,顶顶

燃烧我的罪恶

原帖由 millwood 于 2008-12-15 20:34 发表
燃烧我的罪恶 , cfb means current feedback. most opamps are "voltage feedback types". modern high speed opamps tend to be "current feedback" opamps.

"只能说追求不多 各有各的特点和风格": that's  ...

current feedback
晕 一时没反应过来是这个缩写。。。

燃烧我的罪恶

原帖由 millwood 于 2008-12-15 20:28 发表
mosfets have significant advantages over bjts for high power applications.

however, they are far less complimentary than bjts and they produce more distortions of higher order harmonics.

in ...

恩 139 140这对管做为输出级推耳机很不错
以前做过OP+DB BUF之类的结构
已经很好听