Which opamp on 1.8 V supply in 0.15 has higher gain?

[ccw27]

Opamp for 1.8 V supply

Which of the amplifier below has higher gain? What range of gain can I expect using this amplifier on 1.8V supply in 0.15 process, 70 dB? I am trying to design a Opamp for my bandgap and need one that has a high gain. I am trying to avoid the traditional 2 stage opamp since it requires compensation and hence higher area. I am open to other suggestions for my opamp.

Thanks

 

[rambus_ddr]

Re: Opamp for 1.8 V supply

I think the first can get the higher gain since the cascode output, the second is not easy to increase the gain since the gain mainly acquire from the output stage. But the first output swing is small, but i think it meet our application in bandgap error opamp.
The second reason is the bandgap Vbe is low, so it need a Pmos input.

 

[Mline7]

Re: Opamp for 1.8 V supply

A PMOS Input pair seems reasonable in terms of common mode Input, unless you're using a NMOS input pair in subthreshold (weak Inversion).

I advise you to investigate on folded cascode OPA, which combines the advantages of cascoding and low voltage operation on a single stage.

 

[chinito]

Re: Opamp for 1.8 V supply

I think you're better off with 2nd type since you have low voltage and can't afford cacode output stage as you will lose a lot of headroom. You can do things to increase gain, dealing with voltage headroom with the cascode output stage is more demanding and will require a lot of iteration.

 

[ccw27]

Re: Opamp for 1.8 V supply

What should I use to bias the opamp in the bandgap design? PTAT, Bandgap or constant gm?

Thanks

 

[jordan76]

Re: Opamp for 1.8 V supply

I am very interested in this topic, to be honest.

For your application of bandgap, the opamp needs to have the following features at least:
1.very high gain(>80dB better)
2.very low offset voltage Vos(<several mV better)
3.stable over full operating range(voltage,precess,temperature)
4.fast settling response(<1uS better)
...
I think the bias for the opamp is not so critical, simple voltage divider current source should be ok for the opamp to reduce silicon area.

If I am wrong,please correct me. Thanks.

regards,
jordan76

 

[leebluer]

Re: Opamp for 1.8 V supply

I prefer you to use the 2nd structure to realize the amplifer in 1.8V Vdd. I think it is more suitable to the bandgap. Firstly you should know the minimum Vdd which the amplifer can work under with two different structures, then you would get some clue on choice.

 

[wee_liang]

Re: Opamp for 1.8 V supply

2nd option is better. 1st option burns too much current (remember you still need to create bias voltages for the cascodes).

In fact, I'd suggest you to use a simple diff pair with active loads as the opamp for a start. You'd be surprise the bandgap performance may not be that bad.

Remember: keeping circuit simple is almost always the best thing in analog circuit design.

 

[eda4you]

Re: Opamp for 1.8 V supply

If power consumption isn't of first order priority, use the first design. The cascode will provide you high gain. Maybe it is better to use n-input pairs since the output coltage is normally buffered with a "big" capacitor". (n mos has higher noise figures).

The biasing isn't a dominant design criteria, since it is more important to have high gain. However, for small offset voltages you have to adjust the bias performance to reach your specs.