請教各位大大，PPA和M3改進成雙通道，具體改法是怎么樣的？

[maxyeah]

爬了不少文，可是都沒有找到完整的解答

^-^ thanks a lot

[Shortie]

把Ground Channel全部棄置不用改用正負雙電源

[maxyeah]

把Ground Channel全部棄置不用改用正負雙電源

這樣輸出地不就同樣干擾了輸入地嗎？

[liuchengdar]

M3

MOSFET & 水塘的電由雙電源穩壓拉出

穩壓 + - G 分別接到TLE2426

耳機地 可由穩壓G直通

[amb]

Sorry, but I consider this a degradation of the original design. I've done a lot of testing of 3-channel active ground topology, and in every case it outperforms 2-channel with passive ground.

[isas]

幫AMB大翻譯一下
把Ground Channel砍掉是把M3降級的做法
AMB大做過很多測試
每一項測試三聲道active ground都是勝過2聲道passive ground

[voyager]

嗯....可是AMB大, 一些網友改完後認為解析和低頻有變好耶?

為啥會這樣?

還是說少了地聲道會有其他負面影響?( 雜訊變多....之類的)

如果像Mark Yang大這樣的裝法會不會比單電源更好?

http://tw.myblog.yahoo.com/MKY-Audio/article?mid=378

[EvilHunter]

比較好 但是成本高出一些 (半邊穩壓)
TLE2426電流輸出能力滿低的

active ground的好處...M3和beta22網頁都有講到吧

其實我沒裝過M3....有點落伍的感覺 XDD

[amb]

The M³, β22 and Mini³ websites all describes the benefit of an active ground channel (under "Tech highlights").

Basically, the ground output channel sources/sinks the return current from the headphone and redirects the current to the tightly regulated supply rails rather than dumping the current to signal ground and polluting it. Since the amplifier has good PSRR, any small fluctuations in rail voltage due to the current flow will not affect the amplifier as much as the same fluctuation on signal ground. Also, in class A amplifiers such as M³ and β22, the total power supply current draw is constant, so there is no fluctuation at all.

Measurements show that an active ground amplifier does not suffer from increased stereo crosstalk as the load impedance is reduced like a passive ground amps does. Such crosstalk occurs because of signal ground pollution. I have built a 2-channel version of the M³ ("the β18"). β18 is almost identical to M³, except it uses a dual-rail (non-tracking) power supply with LM317/337 regulators, and no ground channel. You can see the difference in their stereo crosstalk performance (both with 33Ω load):

M³ (3-ch active ground):


β18 (2-ch passive ground):


Signal ground pollution affects not only stereo crosstalk, but considering that real headphone loads are reactive rather than purely resistive, the output current could become phase-shifted to the output voltage. When the return current is dumped to signal ground with a phase shift, the result is a form of dynamic distortion.

The TLE2426 rail splitter in the M³ and Mini³ is not asked to supply much current at all (only in the μA range) -- it simply provides a virtual ground voltage reference, and has the benefit of acting like a tracking dual-rail power supply (because the ground is automatically fixed to the halfway point between the supply rails). It is an effective and lower cost way to get dual-rails with tracking characteristics.

As for why some people think that a M³ modified to 2-channel is "better", I can't explain. It's probably psychological, because whether real or imagined, modifications are supposed to improve things, right?

[狂人]

Hello AMB, I'm wondering what would the transient response be with and without active ground. I've always been wondering what's the difference that active ground is going to have with on the transients when compared to regular ground.


Thanks.

[liuchengdar]

這個問題實在僵持不下~

但樓上V大所提供的Mark Yang改法
雙電源提供更大的電流輸出與更精準的分壓
這點絕對是2426無法比擬的

不如小弟把手上待售的m3也改一改
再把之前售出的m2借回來弄個pk文吧

[isas]

上面那段技術性的小弟看不懂

AMB大是說
M3上面那顆TLE2426是拿來做電位參照用
所以電流負載不大
AMB大無法解釋為何那樣改會有改善 可能是汗水加分?

小弟想偷偷問一下
Beta22用Passive Ground和Active Ground會差很多嗎?
這陣仗怎麼看都大的好可怕....

[amb]

Hello AMB, I'm wondering what would the transient response be with and without active ground. I've always been wondering what's the difference that active ground is going to have with on the transients when compared to regular ground.

The active ground channel's speed is the same as the left and right channels, and does not limit overall transient response. In fact, since the active ground does not rely on the speed of rail capacitors to quickly swing current, it theoretically provides better transient response. In some respects active ground acts like a balanced amp, except that the ground side does not swing any voltage, just current. At any rate, I could not measure any tangible difference in transient response between the 2-channel passive ground β18 amp compared to the 3-channel active ground M³.

雙電源提供更大的電流輸出與更精準的分壓
這點絕對是2426無法比擬的

I already explained that the TLE2426 in these amps do not supply current, so its output current capacity is irrelevant.

The rail-splitting accuracy of TLE2426 is very good (better than ±1% at 40V and 25°C), often superior to independent positive and negative voltage regulators. But that is also largely irrelevant because the amplifier can tolerate a voltage mismatch between its positive and negative supply rails with no ill effects. What is important, however, is the fact that the TLE2426's virtual ground reference "floats" with changes across the two rails, so the system acts like it has a tracking dual-rail power supply, improving CMRR.

Of course, a dual-rail tracking regulated power supply such as σ22 also has this benefit, but is more complex and costly.

小弟想偷偷問一下
Beta22用Passive Ground和Active Ground會差很多嗎?
這陣仗怎麼看都大的好可怕....

There is a difference of course, enough so that I recommend the 3-channel active ground configuration for headphone amplifier application, because high quality headphone listening is by its nature highly resolving of fine details. With a reference-class amplifier like β22, this is what it's all about.

[isas]

小弟想偷偷問一下
Beta22用Passive Ground和Active Ground會差很多嗎?
這陣仗怎麼看都大的好可怕....

There is a difference of course, enough so that I recommend the 3-channel active ground configuration for headphone amplifier application, because high quality headphone listening is by its nature highly resolving of fine details. With a reference-class amplifier like β22, this is what it's all about.

小弟另外想請問
如果三聲道Beta22只用來做耳擴用途
電壓設定在24V
Sigma 22那兩對並聯的MOSFET可不可以拆掉一對?
小弟查了一下那兩隻MOSFET最大輸出電流達17A跟14A.....
如果可以 射極電阻需要降低阻值嗎?

[voyager]

小弟菜英文
有大大可以代為翻譯一下AMB大在TLE2426的解說嗎?
用翻譯軟體只大概看懂一半


AMB大是說用雙電源表現不會比TLE2426好嗎?

[isas]

小弟菜英文
有大大可以代為翻譯一下AMB大在TLE2426的解說嗎?
用翻譯軟體只大概看懂一半
AMB大是說用雙電源表現不會比TLE2426好嗎?

AMB大的意思是說TLE2426性能已經非常棒(誤差只有1%)
像Beta22這類雙電源也不錯
但是 1.貴 2.複雜

UCC穩壓如果是買零件包
U大好像有特別配對某些零件
+-電壓出來也是很準
不過問題點也在要花比較多工夫和錢來裝這個就是

[rx119tw]

雙電源提供更大的電流輸出與更精準的分壓
這點絕對是2426無法比擬的

I already explained that the TLE2426 in these amps do not supply current, so its output current capacity is irrelevant.

The rail-splitting accuracy of TLE2426 is very good (better than ±1% at 40V and 25°C), often superior to independent positive and negative voltage regulators. But that is also largely irrelevant because the amplifier can tolerate a voltage mismatch between its positive and negative supply rails with no ill effects. What is important, however, is the fact that the TLE2426's virtual ground reference "floats" with changes across the two rails, so the system acts like it has a tracking dual-rail power supply, improving CMRR.

Of course, a dual-rail tracking regulated power supply such as σ22 also has this benefit, but is more complex and costly.

小弟理解的內容不知道有沒有錯
來說看看 請前輩指教
---------
首先 tle2426可以提供更為精準的+-電壓
首先是因為一般的雙電源的穩壓
縱使調整成相同電壓
也可能會因為溫度升高 導致越來越不精確
(因為每個晶體的溫度係數都不一樣)
而tle2426是浮動式的
他會根據電壓的改變 自動來調整正負電壓的誤差
而其40V 25°C 的狀況下
誤差應該為+-1%之內

但正負電壓的準確是不是真的如此重要 也可以在討論
因為amp本身對於正負電壓之不匹配
具有一定程度之忍受力
------------
以上 有錯請指教 感謝~~~

[voyager]

嗯嗯.....原來如此

那如果TLE2426這麼好用, 為何Beta22要改用雙電源呢?

是因為單電源的輸出無法滿足Beta22嗎?

[rx119tw]

嗯嗯.....原來如此

那如果TLE2426這麼好用, 為何Beta22要改用雙電源呢?

是因為單電源的輸出無法滿足Beta22嗎?

據amb大所提
beta22之電源σ22
是dual-rail tracking regulated power supply
應該就是具有與tle2426一樣
可以控制+-電壓之精確性的power

[EvilHunter]

嗯嗯.....原來如此

那如果TLE2426這麼好用, 為何Beta22要改用雙電源呢?

是因為單電源的輸出無法滿足Beta22嗎?

正負電壓的準確度只是其次...
M3的TLE2426只要餵三顆OP的input bias current
beta22線路中有其他點要接地
TLE2426的20mA很容易就不夠了

sigma22的負電源 會跟著正電源的大小跑(反向)
(寨主的獨臂刀王也有類似的架構)
對後端amp來說 PSRR應該會比較好一些
不過不會像TLE2426那麼準
有電阻和電晶體差動的誤差(應該還不小...)

[liuchengdar]

今天長了不少見識~~~

[amb]

小弟另外想請問
如果三聲道Beta22只用來做耳擴用途
電壓設定在24V
Sigma 22那兩對並聯的MOSFET可不可以拆掉一對?
小弟查了一下那兩隻MOSFET最大輸出電流達17A跟14A.....
如果可以 射極電阻需要降低阻值嗎?

It depends on how much current your load will draw, and the power transformer specifications. The σ22 was designed to supply up to four β22 boards, each biased to 160mA (for a total of 200mA per board), using onboard heatsinks. So the total draw is about 0.8A - 1A.

Assuming the pre-regulated voltage is 10V higher than the regulated voltage, then the σ22 must dissipate 1A * 10V = 10W in heat per rail. That's too much heat for one MOSFET on a board-mounted heatsink. That's why there are two MOSFETs in parallel, to help share the dissipation. Even 5W on each MOSFET will be hot.

If your load will draw only a small amount of current, then yes, you can use only one MOSFET per rail. The source resistors do not need to change.

[cklee]

The active ground channel's speed is the same as the left and right channels, and does not limit overall transient response. In fact, since the active ground does not rely on the speed of rail capacitors to quickly swing current, it theoretically provides better transient response. In some respects active ground acts like a balanced amp, except that the ground side does not swing any voltage, just current. At any rate, I could not measure any tangible difference in transient response between the 2-channel passive ground β18 amp compared to the 3-channel active ground M³.

I still doubt about how will the g-channel's response be when the both left/right channels want to draw high output current and exceed the g-channel's current supply limit.

Using single speed g-ch to supply double speed current(left channel + right channel)?

[maxyeah]

這個討論很好，特意提供完整對話型翻譯（各位brothers原諒我可能亂碼，我在使用簡體中文系統，看不到請post me

messages）:

maxyeah:
請教各位大大，PPA和M3改進成雙通道，具體改法是怎么樣的？
爬了不少文，可是都沒有找到完整的解答

^-^ thanks a lot
__________________

shortie:
把Ground Channel(翻譯：地通道、地頻道)全部棄置不用改用正負雙電源

__________________

maxyeah：
這樣輸出地不就同樣干擾了輸入地嗎？

__________________

liuchengdar：
M3

MOSFET & 水塘的電由雙電源穩壓拉出

穩壓 + - G 分別接到TLE2426

耳機地 可由穩壓G直通
__________________

amb：
很抱歉，但是我還是覺得這種改變是原版的倒退。我做過大量有源三通道路線的測試，他們每樣都表現得比二通道無源地通道

版本優秀。

—————————

isas

（省略）

——————————
voyager

嗯....可是AMB大, 一些網友改完後認為解析和低頻有變好耶?

為啥會這樣?

還是說少了地聲道會有其他負面影響?( 雜訊變多....之類的)

如果像Mark Yang大這樣的裝法會不會比單電源更好?

http://tw.myblog.yahoo.com/MKY-Audio/article?mid=378

——————————
EvilHunter

比較好 但是成本高出一些 (半邊穩壓)
TLE2426電流輸出能力滿低的

active ground的好處...M3和beta22網頁都有講到吧

其實我沒裝過M3....有點落伍的感覺 XDD

——————————
amb：
m3，B22和mini3網站都有講述有源地通道的好處（在他們的“技術亮點”里面）
基本上，地通道吸收和發放來自耳機的回歸電流，疏導這些電流通往已經穩壓得很妥當的電源端，而不像以往那樣，向信號地

傾灌。由于放大器芯片有很好的抗電源擾動性，電源端電壓由于電流灌注引起任何小變動不會像電流傾注在信號地通道影響那

么嚴重。同時，在甲類狀態下的放大器，比如是m3和B22，他們的總供電電流是

[maxyeah]

要上课了，我最于地通道也有自己独到的意见想发表，而且不短，今天太累了，下次我再说.....感谢各位朋友的讨论，从中让我增进不少....

thanks amb....
could you give me your msn?
my email-box is dead now