Two Stage Miller OTA gain?

[The_Dutchman]

Hi,

I am designing a 2-stage miller OTA like the one in the attachment.
I need 72dB DC-gain, currently I'm stuck on 57.8dB.

I've calculated an expression for the gain:
Av(DC) = gm1(Ro1//Ro3)(Ro5//Ro6)[gm5 - gm8]
What do you think of this expression?
My instinct would say:
Av(DC) = gm1(Ro1//Ro3)gm5(Ro5//Ro6)
However the last doesn't take into account the small signal current flowing through M8-M9-M6

To increase the gain I increased the lengths of M1 M3 and M6 and M5 so Ro increases and thus the gain. But the width also drastically needs to increase for the current flowing (420µA in first stage, 6.1mA in second stage) and I reach the maximum W for a single MOST.
If I decrase Vgs-Vt I get closer to linear region requiring an even bigger W/L ratio for the same current.

Can anyone help me with this?

 

[erikl]

... I reach the maximum W for a single MOST.

In such case you can always use several MOSTs in parallel (m>1) or use fingering (f>1).