Continue to Site

Welcome to EDAboard.com

Welcome to our site! EDAboard.com is an international Electronics Discussion Forum focused on EDA software, circuits, schematics, books, theory, papers, asic, pld, 8051, DSP, Network, RF, Analog Design, PCB, Service Manuals... and a whole lot more! To participate you need to register. Registration is free. Click here to register now.

[help] urgent!!! about switched capacitor amplifier design.

Status
Not open for further replies.

lilac

Junior Member level 1
Joined
Jan 26, 2005
Messages
18
Helped
1
Reputation
2
Reaction score
0
Trophy points
1,281
Activity points
229
Hello, everyone!
I want to design a 12bit switched capacitor amplifier with following specs: 12 bit , closed-loop gain is 10, sampling freqency is 10MHz, supply voltage is 3.3V. and power consumption is not very restrict.
I have several questions:
1) To a switched capacitor amplifier, 12 bit means what?
2) Which type of opamps can match this request? and how to decide the specs of the opamps in order to make the amplifier work well under 10MHz sampling freqency?
3) For sampling and hold circuit, which circuit structure is the best choice. Can I use the noninverting structure to realize it ?
 

High gain(about 90dB), high Unit Gain(100MHz or so)
 

    lilac

    Points: 2
    Helpful Answer Positive Rating
Re: [help] urgent!!! about switched capacitor amplifier desi

lilac said:
Hello, everyone!
I want to design a 12bit switched capacitor amplifier with following specs: 12 bit , closed-loop gain is 10, sampling freqency is 10MHz, supply voltage is 3.3V. and power consumption is not very restrict.
I have several questions:
1) To a switched capacitor amplifier, 12 bit means what?
2) Which type of opamps can match this request? and how to decide the specs of the opamps in order to make the amplifier work well under 10MHz sampling freqency?
3) For sampling and hold circuit, which circuit structure is the best choice. Can I use the noninverting structure to realize it ?
define Vout as the output of SC, T=1/f=100ns,
1. 12 bit accuracy: ΔVout(t)/Vout(∞) < 2^(-12) after t=T/2(i.e. 50ns)
2.closed-loop gain is 10, it means β=0.1, defined unity-gain-bandwidth as UGB=A0*ω0
from (1), ΔVout(T/2)/Vout(∞)=exp(-β*UGB*T/2) < 2^(-12),
===> UGB > 1.664G
===> if A0= 1000 (i.e. 60dB), then ω0 > 1.664M rad/s ===> f0 > 264 kHz/s
3. for S/H circuit, β can be reduced to 1, so the spec. of OP can be loosen.
which circuit structure should be choosed ? it depends on ur application.
 

    lilac

    Points: 2
    Helpful Answer Positive Rating
Re: [help] urgent!!! about switched capacitor amplifier desi

Thanks sunking and Btrend!


2) ===> UGB > 1.664G
===> if A0= 1000 (i.e. 60dB), then ω0 > 1.664M rad/s ===> f0 > 264 kHz/s

/***According to what you said, does the amp( A0=10000 (80dB), f0 > 26.4 kHz/s) can meet the request?

3). for S/H circuit, β can be reduced to 1, so the spec. of OP can be loosen.
/***Sorry, I don't very understand what's mean. Can you explain it more clearly?

which circuit structure should be choosed ? it depends on ur application.
/***In fact, when I use a novinverting sampling structure and an amp (A0=10000 (80dB), f0 = 28 kHz/s), the amplifier can only work well at most at 2MHz sampling freq. I don't know whether it is due to the limitation of sampling cor amps circuit. if due to the forth, is there any high-speed sampling tech. to realize 10MHz sampling freq.? [/i]
 

Re: [help] urgent!!! about switched capacitor amplifier desi

lilac said:
Thanks sunking and Btrend!


2) ===> UGB > 1.664G
===> if A0= 1000 (i.e. 60dB), then ω0 > 1.664M rad/s ===> f0 > 264 kHz/s

/***According to what you said, does the amp( A0=10000 (80dB), f0 > 26.4 kHz/s) can meet the request?

3). for S/H circuit, β can be reduced to 1, so the spec. of OP can be loosen.
/***Sorry, I don't very understand what's mean. Can you explain it more clearly?

which circuit structure should be choosed ? it depends on ur application.
/***In fact, when I use a novinverting sampling structure and an amp (A0=10000 (80dB), f0 = 28 kHz/s), the amplifier can only work well at most at 2MHz sampling freq. I don't know whether it is due to the limitation of sampling cor amps circuit. if due to the forth, is there any high-speed sampling tech. to realize 10MHz sampling freq.? [/i]
1. A0=80dB , f0=26.4kHz should be o.k.
2. in S/H with no gain, β=1, so exp(-β*UGB*T/2) < 2^(-12)
reduced to exp(-1*UGB*T/2) < 2^(-12) ===> ur UGB requirement can be reduced 1/10.
3. did u include load cap (Cload) at the simulation of ur f0, in single pole behavior the first pole should be at gm/Cload, which is depended on Cload
 

    lilac

    Points: 2
    Helpful Answer Positive Rating
Re: [help] urgent!!! about switched capacitor amplifier desi

I believe you need to check the specification first. 1.6G GBP is a large number. Which tech you are using? do not do impossible job
 

    lilac

    Points: 2
    Helpful Answer Positive Rating
Re: [help] urgent!!! about switched capacitor amplifier desi

dear Btrend,
according to your euation given,
you can decide what should be the UGB of the opamp for a given accuracy and sampling time.but how to decide what should be the open loop gain should be.
 

    lilac

    Points: 2
    Helpful Answer Positive Rating
Re: [help] urgent!!! about switched capacitor amplifier desi

open loop gain=Aβ=UGB*β/w0
 

Re: [help] urgent!!! about switched capacitor amplifier desi

but dear Btrend,
from the above equation,once you get the UGB,to get the value of ω0,you have already assumed A0=1000. i want to know how did you assume that A0 should be 1000.
 

Re: [help] urgent!!! about switched capacitor amplifier desi

1.6G UGB is possible for 0.18um process
 

Re: [help] urgent!!! about switched capacitor amplifier desi

Dear Btrend pl reply.its urgent.
 

Re: [help] urgent!!! about switched capacitor amplifier desi

avinash said:
but dear Btrend,
from the above equation,once you get the UGB,to get the value of ω0,you have already assumed A0=1000. i want to know how did you assume that A0 should be 1000.
that's a good question. in the above derivation I choose A0=1000 for explanation only. There is a theorem behind the choice of A0. cause A0 will affect the accuracy of the assumption of virtual short in ur SC circuit.
I want to say first, the following derivation is based on continuous time system, I'm not sure whether it is suited for discrete time system (such as SC ckt.), I had read some documents about the non-ideal effect of opamp such as finite gain, finite bandwidth, finite slew rate,...., but I had not understood totally in the current stage.
In negative feedback theorem, the difference between "+" and "-" is Vi/(1+Aβ).
so there will be a gain error 1/(1+Aβ) after the SC ckt.
for example, if ur SC is used as a multiplier by 2, then ur "2" will reduce to "1.998"
if A=2000, β=0.5.

hope this help u.
 

Re: [help] urgent!!! about switched capacitor amplifier desi

the determination of A is depended on ur application,
in Nyquist rate:
if ur spec is 12bit == 4096 == 72.25dB, ur spec. of Aβ should be large than 72.25dB to overcome the gain error (1/(1+Aβ)), that is 20log(1+Aβ) > 72.25dB
in oversampling sigma-delta application:
cause "oversampling sigma-delta " will increase the SNR by (N+0.5)*6 dB for every doubling of OSR, so the requirement of Aβ can be reduced to a lower value depended on ur choice of N, and OSR
 

    lilac

    Points: 2
    Helpful Answer Positive Rating
Status
Not open for further replies.

Similar threads

Part and Inventory Search

Welcome to EDABoard.com

Sponsor

Back
Top