標題: doubt supply-independent biasing [打印本頁] 作者: ffsher100 時間: 2011-7-12 06:02 PM 標題: doubt supply-independent biasing i got a question from Razavi textbook in chapter 11.2 : Supply-Independent Biasing 1 _* F1 O& o @' n, M+ a[attach]13191[/attach]==>figure 1) a, |0 ~' O0 Q ]0 V% |
[attach]13192[/attach]==>figure 2 + k# j5 ^% h- ^7 o8 |& y0 i/ J7 V$ L
by figure2, it looks like vdd really has no effect on the figure 1 CKT. 2 P! t, j. d& C0 Q. }1 x1 o- Z. `8 Q- }! H0 J3 j
the textbook also introduce start-up circuit M5 into the design as following:0 V$ U* u z" ^! w c; K" x
[attach]13194[/attach]==>figure 3/ _4 K5 B" u1 s
when simulating the circuit with startup(M5), it seem to be effected by VDD.6 g& c1 F+ S. d5 ~
Current Iout, Iref vary with VDD. 1 E9 Y" L8 J, L; _In practice, figure 3 become Supply-dependent Biasing since startup must include. # h, z1 m! Q2 L% _how could this be? if using figure 3, supply- independent case never happen.$ l7 ?( n( s) y* s! F
7 G8 w9 U$ y) ?5 F0 z$ n5 bbelow figure, i sweep .dc vdd. ' H* N8 J5 z. n6 S( r" I5 H
[attach]13193[/attach]( ]$ F. u- o% N
As you see vdd ramp from 0v to 5v, current Iref and Iout are not constant if vdd>2.8v作者: lchuang 時間: 2011-7-27 05:21 PM
從你的結果來看，分三種case來檢查......... * k1 m4 O# U5 ]) x" E3 }' E P/ q
輸入電壓(分別為2.5V、2.8V and 4V)後，檢查你每一顆MOS的工作點，是否都為飽和區，2 M2 U; C6 l4 D: E
/ W5 y% P/ G6 J; B Q
並且記錄每一顆MOS的Vds(sat) 、Vds跟Vod 2 J! S( K. @% }" a7 S + C2 y0 m/ s( e5 h5 R: w(如果沒有Vod的話也沒關係......這是Vgs-Vth得來的，Vod這參數要H2006以後的版本才有) & s' ?* j- y- I- _) ]8 i e+ _ ) n4 [9 F& ?8 E* S* y3 Y' l, ^- m有了以上的資訊你才可以知道問題出在哪............ n3 k+ O/ y) o5 n ; @3 c% |- t9 z4 b" f6 pPS：既然M5是當作startup用的MOS，那麼W/L比就不用太大，以一般設計來說W=0.5u L=10u' u, ?8 `9 H. U$ x3 D/ F; {
. V/ b7 Z' v- _, v. ^一般你要模擬所謂的start-up MOS，以你上面電路為例，當你不加入M5這一顆MOS的話......$ @, t4 D' k8 Y% U/ u1 U
3 r/ v. M) K' _# p1 P9 @$ G
你可以在spice檔內下一個初始值的指令，6 C1 I ~+ Z2 G6 w! c; K% S9 t# \* Z