Designer’s Guide Community

Dipankar

Community Member

Offline

Posts: 59

tackling noise in OPAMP : any preferred architecture ???
Apr 15^th, 2010, 8:29pm

Dear All,

Can someone share some experience regarding low noise OPAMP design . Basic things I follow is

1 make gm of the differential pair higher compared to the load device 2. If possble use a little bigger device to lower the flicker.
3. use PMOS differential pair.
4. where possible use resistor load for the second stage.

(I don't like to increase miller cap to improve noise (voltage) performance at the cost of BW or power. )

But is there any study or thumb rule regarding architecture selection - under same dc power consumption which one is noisier than other.

Back to top

With Thanks and Regards,
Dipankar.

IP Logged

Mayank Community Fellow Offline Posts: 334	Re: tackling noise in OPAMP : any preferred architecture ??? Reply #1 - Apr 17^th, 2010, 9:28pm A correction -- Make Gm/Id of the load device smaller than input one. QUESTION -- how will a pMOS input transistor reduce noise ?? addition -- you can filter Tail Current Source noise with a // cap or a series inductor.
Back to top	IP Logged

nobody Community Member Offline Posts: 75	Re: tackling noise in OPAMP : any preferred architecture ??? Reply #2 - Apr 17^th, 2010, 10:42pm Pmos input pairs are used for low flicker noise,I guess.
Back to top	IP Logged

analog_chip New Member Offline Posts: 2 Bangalore	Re: tackling noise in OPAMP : any preferred architecture ??? Reply #3 - Apr 19^th, 2010, 3:34am Is there any provision for using BJT or let me know the maximum supply current that you are looking for from the architecture and also the input bias current. What is the noise value you are considering for design?
Back to top	Regards, analog_chip IP Logged

Dipankar Community Member Offline Posts: 59	Re: tackling noise in OPAMP : any preferred architecture ??? Reply #4 - Apr 19^th, 2010, 9:11am No BJT available. spec : input referred noise ~ 6 nV / sqrt(Hz) @ 2MHz , total dc current ~ 300 uA, vdd = 1.8 , CL ~ 0.1- 0.2 p . GBW > 300 MHz , PM > 60.
Back to top	With Thanks and Regards, Dipankar. IP Logged

loose-electron

Senior Fellow

Offline

Best Design Tool =
Capable Designers

Posts: 1638
San Diego California

Re: tackling noise in OPAMP : any preferred architecture ???
Reply #5 - Apr 19^th, 2010, 4:26pm

Why don't yopu take a systematic approach and do a noise analysis of the device and find out what are the noise contributors?

- Oh and one comment - a parallel capacitor on the tail current source?

On the source itself kills you common mode rejection of the amplifier = bad thing.

On the bias of the source is ok.

Back to top

Jerry Twomey
www.effectiveelectrons.com
Read My Electronic Design Column Here
Contract IC-PCB-System Design - Analog, Mixed Signal, RF & Medical

IP Logged

Mayank

Community Fellow

Offline

Posts: 334

Re: tackling noise in OPAMP : any preferred architecture ???
Reply #6 - Apr 21^st, 2010, 7:51am

Hi,

@ Jerry : Yes Sir, CMRR will degrade at higher frequencies...But if cap is chosen keeping in mind the CMRR / CMFB LoopGain BW, then it should not be a problem i guess.

Anyways, A cap at Bias OR Passive RC filtering will bring down noise injection from Bias cktry...which is widely used now...But it wont bring down Noise injection from Tail Current Source.

Also, can somebody explain wot "nobody" said Tongue

& I Quote:

Pmos input pairs are used for low flicker noise,I guess.

For same Gm, pMOS will require larger sizes than nMOS, thus reducing the flicker noise ?? ...Or is there something deeper coming into play here ??

thanks,
Mayank.

Back to top

IP Logged

HdrChopper

Community Fellow

Offline

Posts: 493

Re: tackling noise in OPAMP : any preferred architecture ???
Reply #7 - Apr 21^st, 2010, 6:32pm

Some additional comments/suggestions

Dipankar wrote on Apr 15^th, 2010, 8:29pm:

1 �make gm of the differential pair higher compared to the load device 2. If possble use a little bigger device to lower the flicker.
3. use PMOS differential pair.
4. where possible use resistor load for the second stage.

(I don't like to increase miller cap �to improve noise (voltage) performance at the cost of BW or power. )

1. However, larger gm means larger BW, thus integrated noise is NOT improved...the only way is by increasing miller cap (at least for two stage amps) although you do not like it.

2. Chopp the amplifier so you get rid of the flicker
3. IF chopper is not possible then PMOS ok
4. Use resistor load for first stage, where noise is set. Use max possible tail current and minimum resistor load.

Best
Tosei

Back to top

Keep it simple

IP Logged