- #4048 [WilliamsonAudio], 24-05-09 13:49
- #4047 [WilliamsonAudio], 24-05-09 08:10補上圖片
- #4046 [WilliamsonAudio], 24-05-09 08:10下面是臺灣張文俠的一段。
假設我們用一支p較高的真空管来推動300B,例如ECC83/12AX7,根擦真空管手冊得知ECC83/12AX7的屏
BErp在屏壓250V時篇62.5KS2。
-般三極的屏極負载電阻RL大多設定在屏内阻mp的3~7倍之間,我們取其中間值5倍爲屏極負載電阻,mp的
倍爲屏極負戴RL,即:
62.5K2x5=312.5K0
代入上式:
fc=159/RL{Cin+Cgp(1+A))
=159/0.3125{9415(1+3)]
=7.37KHz
----
如果屏阻用約100k, 頻應也有約20kHz -
- #4045 [WilliamsonAudio], 24-05-09 07:37#4043
我記得是12AX7 的頻應不高(沒有回輸),原因是內阻高和Miller effect.要用回輸來增加頻應。
所以後期的設計主流,就是用低內阻膽。當然沒有回輸設計也是主因之一。在無回輸情況之下,頻應也要足夠寬闊。在Low Tim 失真的石機設計。也是有這一條設計原則。
ARC, 開頭用12AX7.之后是6DJ8.現在是6H30
最後修改時間: 2024-05-09 07:42:59 - #4044 [johnnykmtang], 24-05-09 07:19In general, it is better to drive the Interstage transformer (IT) with low impedance. It is a lot easier to get better performance of the IT with lower inductance. Therefore, using the 6SL7 or 12AX7 isn't good with inductive load. The limitation is the IT.
Med/high gain triodes like D3a, C3g and 5842 which has low plate resistance would be nice to drive the IT or plate choke.
Johnny
最後修改時間: 2024-05-09 07:32:48 - #4043 [niceday009], 24-05-09 00:31Hi Wilson,
用12AX7 或6SL7 做differential stage 有一好處,一級电压放大已經達到合併机之增益,Pathos自己應該諗過呢個優點,但用AX7或SL7唔可以用PP牛load因為做唔到隻咁大电感牛比12AX7 load, 計過起碼要1200亨(60K(12AX7 內阻)x2.5÷(2x3.14x20)),只可以用CCS
Sim過12AX7 swing 到58V P to P(等於50W RMS output 幅度),失真其實好少,只有二次0.02%失真,三次order 以上harmonic 基本係冇乜
我完全相信你曾經講過,电压放大Triode 最好,电流放大石最好,失真已經證明到,而80年代其中一期MJ曾講過失真糸數係無負回輸下更有意義
最後修改時間: 2024-05-09 00:46:09 - #4042 [WilliamsonAudio], 24-05-08 22:57#4040
Hi Eric
Pathons 是用小九腳膽做推動。ECC83+ECC82
ECC83 就算用到舊裝金獅B759 或 Mullard 第一代炭黑屏。在冇回輸之下。開環的特性。是十分不理想的。冇回輸的設計,儘量避免用ECC83.
ECC82 的失真特性。也是比6SN7高。
ECC99 雖然係9 腳膽,但我見Lynn Olson 也推介用。應該是非常之綫性的膽。
最後修改時間: 2024-05-08 22:58:03 - #4041 [johnnykmtang], 24-05-08 02:52Hi Niceday,
Using Class A PP MOSFET stage works like a power buffer circuit. That means the output voltage is 1:1 between the gate input and output.
The electrical characteristic between Class A tube PP and MOSFET PP is a bit different. The tube PP would have a higher impedance than the MOSFET. The MOSFET class A would require a large heatsink.
I believe the MOSFET PP is something I would like to try too.
Johnny - #4040 [niceday009], 24-05-08 00:47Hi Johnny,
In term of distortion level, the differential stage with plate output transformer seems to be the lowest we can find
If we go for PP, the single differential high gain stage with either CCS or transformer connected, then coupled to Class A PP MOSFET stage should be the lowest distortion. Without using triode as final output device meaning no need to step down so many. We only need the differential stage to swing to exact output voltage for output power level, say 20 V RMS can produce 50W RMS. Much Much lower voltage swing compared to Triode power tube, much less voltage swing means lower distortion accordingly.
So why do we still need power tube?MOSFET can do the work more efficiently even without a transformer
I actually quite like the Italy Pathos circuit, attached is their power amp utilizing single differential stage and cathode follower then connected to class A MOSFET stage
最後修改時間: 2024-05-08 01:06:49 - #4039 [johnnykmtang], 24-05-07 01:10#4026, Hi Niceday,
My guess the problem with the Marantz T1 is with those input and inter-stage transformers. Those UTC transformers were not designed for wide bandwidth audiophile application. As you said, the CCS at the common cathode is the best performer of this balance PP circuit.
I have built balance differential PP line and power amp. Both sound so good and exceptional low background noises.
SE amps are charming but the PP is great if it is designed correctly.
Johnny - #4038 [niceday009], 24-05-06 22:02点至可以用齊50同6B4G 胆屏耗作power output 係一個話題,係乜情況先致可令兩款胆有效用齊兩隻18W(18x2 =36W) 屏耗? 訊號入6B4G 先放大3.9倍,放大咗3.9倍此放大电压再入50 grid 再放大因為Ground Grid同相,我諗唔到点樣Cascode 兩隻功率胆可以同時有效地用盡屏耗???
或係我想多咗,9W output 只係由50胆出
最後修改時間: 2024-05-06 22:25:46 - #4037 [WilliamsonAudio], 24-05-05 12:24正常50膽只有 3-4 W
- #4036 [WilliamsonAudio], 24-05-05 12:22這個50+6B4G 有成9W 多。
比正常50膽多了一倍power - #4035 [WilliamsonAudio], 24-05-05 12:20diyAudio 有人提到可以有柵流,也就是 A2 mode
https://www.diyaudio.com/community/threads/power-tubes-in-cascode-for-output-section.213608/ - #4034 [niceday009], 24-05-05 02:33#4033
靜態兩隻胆一定相同idle current 因為兩隻係cascode, 只能夠最好處係可以 driver tube 屏出直接能推最後功率胆,但功率胆栅係接fixed 电压,咁樣是否仲可以A2?
最後修改時間: 2024-05-05 02:34:28 - #4033 [WilliamsonAudio], 24-05-04 23:51#4023
作者是這樣設計。
用50ma
6B4G Ep=210V,Ip=50ma
50 Ep=400V,Ip=50ma - #4032 [WilliamsonAudio], 24-05-04 22:57這個Patent 就是用 cascode 去取得一個低失真的推動電壓,然後再推。power stage.
電路有點複雜。我是上年下載的。但還沒有細心研究。 - #4031 [WilliamsonAudio], 24-05-04 22:44#4029
我有時間。就試下Sim
我見有高手用這個方法。去取得低失真的電壓。
而且有例子用6SN7, 但電壓用到750V - #4030 [WilliamsonAudio], 24-05-04 22:22wireless world 1956 年有coscode AF amp.
是我最早見到的文章提到cascode, 他的例子是用7
F7,特性等同6SN7
N 年之后Glass Audio 有人用這個理念設計了一台Power amp,憑記憶是用6DJ8
https://keith-snook.info/wireless-world-articles/Wireless-World-1956/Cascode%20A.F.%20Amplifier%20-%20Long-tailed%20cascode%20pair.pdf - #4029 [niceday009], 24-05-04 22:01Wilson,
有時間可能sim下佢至知distortion level, cascode 上面隻ground grid輸入阻抗比common cathode 低好多可能20K左右,用6dj8 2.7K plate resistance 可以搞但6SN7做cascode 可能失真好大, 12ax7就唔使諗
最後修改時間: 2024-05-04 22:04:12 - 下一頁 (1 of 203)
- 返回 ...
https://pearl-hifi.com/06_Lit_Archive/02_PEARL_Arch/Vol_08/Sec_31/2071_The_Miller_Effect.pdf
12AX7 -3dB =12.4kHz
但我用LTSPICE 去 Sim,
又沒有那麼低。