Sansen讀書會...

賴永諭

剛剛看了一下Ch1...,好像有些許錯誤的地方哩...
Slide 0129: gmb 電流方向(Bluck 接地時S->D,if Vbs>0, 電流由D->S)
Slide 0134&0136: 應該是VGST(VGS-Vt)~0.2 V
! i$ j9 w- y# r! V2 s& g$ Z$ ?5 m$ c 歡迎大家踴躍討論一下...剛剛開始閱讀

賴永諭

: N  I- p; i2 B
slide 1524 and 1525
+ M  h9 P. C. d( |) o
2 r( ^2 R3 g  p& U* P, g! {/ |個人分析iout/ic = 1/(gm3*ro1*2)

賴永諭

賴永諭 發表於 2022-8-5 12:03 PM
slide 0444
8 q; M" U$ Q  t; mNoise of a current mirror with series R:
前面的R小於1/gm時, 為什麼iout noise 隨的電阻增加 ...

更新一下~~
(1)因為當R變小, 電阻的雜訊電流就不會流到輸出端了
(2)電阻的雜訊電流到輸出端分析8KT/(R2+1/gm2), 當R2夠大時, 則整個8KT/R2 雜訊電流全部流到輸出端
所以前面的R小於1/gm時iout noise 隨的電阻增加, 後面隨電阻增加減少

賴永諭

slide 1756：

; o) n+ D; N/ }8 T7 B, L# }1.SC積分器的BW=(Feedback factor) X Gm/Ceff
2.Feedback factor=Cf/(Cf+Cs+Cp)=Beta
- g. u6 o' f# u6 [3.Ceff=(Cf串(Cs並Cp)): Cf回授電容，Cs取樣電容，Cp 輸入開關電容求和節點處的寄生電容
4.Integration期間的GBW=BetaXGm/Ceff
9 f3 G: E0 ^7 e8 t. w5.Integration期間的SR=Io/Ceff(Io opamp 輸出電流）
, h2 r7 y0 H& q9 h. N8 P( g2 i; r

賴永諭

slide 218
- d" @* R' e% B4 z# `8 \+ u, ~1. First order Sigma-Delta patterns in the output spectrum
=> Idle tones(pattern noise, limit cycles)
2. Second order Sigma-delta 輸入太大時還是會不穩定

賴永諭

* o. R# P' P. Fslide 215-216
1. Signal 的輸入雜訊直接到輸出
. l: S( `2 p% Z' v8 {2. DAC 的輸出雜訊也直接到輸出
3. Quantizer noise 才有noise shaping 到輸出

賴永諭

silde 1557：
7 _9 g5 s( B' a' Z% n1. Input offset: Bipolar is better than MOS
KT/C = 26mV << MOS Vov
" J6 u- J! u, x9 C4 @  f2. Mismatch 造成Randon offset
3. 電路不對稱所造成Systematic offset

賴永諭

slide 1020

（1) The closed-loop gain can be modified by changing RS, leaving the closed-loop bandwidth unchanged.
; f2 W" A* X( x% D3 i（2) For a given RF, frequency compensation can be optimized.
（3)Suitable for high frequency applications.

賴永諭

slide 1010

" S. e8 @- S1 vFor a CFB amplifier the open-loop gain is ZT and its feedback factor is 1/RF, making the loop-gain TCFB = ZT/RF.
2 W( V2 L- X7 H* SThe signal bandwidth is determined by RF and not by the circuit gain.
The circuit gain is independently set with RS.
8 l! r& l3 Z" v: @3 b5 ]' d3 LThe signal bandwidth remains stable for all gain settings .
Even for unity-gain operation, an RF resistor is required.
. }: p* H! ^$ _The best practice when designing with CFB op-amps is to use the RF values given in the datasheet and to adjust the desired gain level through RS.

賴永諭

' k. q" p. y9 \4 t2 G+ C: d

賴永諭 發表於 2024-6-6 11:16 AM
Two stage ring oscillator analysis文件上傳~~

" H% ]: C5 E7 O+ K' o# X1 @% FTwo stage ring oscillator analysis文件上傳~~

賴永諭

Two stage ring oscillator analysis文件上傳~~

賴永諭

Update 文件~
! \8 N6 r% l( I4 E% @5 A1 w9 w* VTwo stage ring oscillator analysis
2 z- ]0 d; Z" N% A1 Q" a. ^

賴永諭

Design of Crystal Oscillators
slide 2255補充資料：Two stage ring oscillator analysis

3 w& ]; Z7 |, w- _) Z  ~https://drive.google.com/file/d/ ... Hq/view?usp=sharing
+ @# n0 w  X, k

賴永諭

RIP

In Memoriam : Willy Sansen (1943-2024)

$ X$ h4 y* k  g4 ]& |https://bbs.eetop.cn/thread-968254-1-1.html

賴永諭

賴永諭 發表於 2024-3-4 04:01 PM
0 a. ^. l4 ^% B6 _7 v$ Islide 2222
) @( C% g. l6 d: rVerify that RB>1/(RsC3^2Ws^2)

gmA~ 4RsCL^2Ws^2
and 4WsCL^2/QCS


( X' L6 T1 c+ s1 Y' \  T

賴永諭

slide 2264 Calculation of gmA
. B9 z. W; O9 k" g! o" z- v2 _$ G
gmmax 應該是C1WC1/2C3

=>ther/2=gmax/(Ceff x ws)^2可推算出來

賴永諭

slide 2222
! l  N) Q  r' h4 w4 ~1 l6 j( yVerify that RB>1/(RsC3^2Ws^2)

=>說明如下, cmos 負電阻~gmA/Ws^1C1C2
用RB 來biasing 應該要大於1/gmA 電阻
# `, F2 I: C" f" T1 D! E所以RB應該大於1/(RsC1^2Ws^2)

賴永諭

slide 2222
Verify that RB>1/(RsC3^2Ws^2)
/ ~, t. T' [- h" r$ u
=>說明如下, cmos 負電阻~gmA/Ws^1C1C2
用RB 來biasing 應該要大於1/gmA 電阻
) v* |3 E1 j/ Z# ~+ k/ z所以RB應該大於1/(RsC1^2Ws^2)
  l9 R7 c1 h- t% }' b8 w# M7 L: M

賴永諭

slide 2222
Verify that RB>1/(RsC3^2Ws^2)

=>說明如下, cmos 負電阻~gmA/Ws^1C1C2
% U5 H3 A  P. D: t2 y用RB 來biasing 應該要大於1/gmA 電阻
+ u3 k( D/ v. @所以RB應該大於1/(RsC1^2Ws^2)
5 C) s- m$ d: O( C

賴永諭

slide 2222
/ S1 U+ M7 N; Y% aVerify that RB>1/(RsC3^2Ws^2)

1 z- M1 P1 n" e$ _; l=>說明如下, cmos 負電阻~gmA/Ws^1C1C2
用RB 來biasing 應該要大於1/gmA 電阻
所以RB應該大於1/(RsC1^2Ws^2)