Hi,
 I am trying to do a simple reference which should follow Vt. I just did a pmos diode connected and a resistor. I expected as I change the temperature, the threshold decreases and the voltage generated should be vgs-vth and it should change as per vth change, but I do not see that change. Is that I am missing something, this is in DSM in 65nm.

Is there any way to generate the voltage following the Vth to compensate for the temperature.

Thanks,
SBR

Hi,
 I am sending the fig. as per your request. I am working in 65nm and I see the mobility variation ~50% across temperature. I need to generate the bias which controls to generate the current which should vary to compensate for the mobility reduction across temperature to provide a constant delay cell.
The bias is generated for the constant delay cell.

Thanks and Regards,
SBR

It really sounds like you are trying to follow g_m, not V_t since you are trying to compensate for mobility.  There is a very simple constant g_m circuit that is found in a number of text books.  One that I think of off the top of my head is the Johns and Martin book, "Analog Integrated Circuit Design".

I ran a quick search on google, and found a paper that describes the technique.  www.eecg.toronto.edu/~kphang/papers/iscas2004_nicolson_opamp.pdf.  I have not used the additional techniques presented by the paper, but it does give a nice overview of the traditional way to bias amplifiers.  One difference is that I always implement the resistor on chip, so I don't ever see the stability problems that they were talking about.

I realize that everything that I have talked about is generated currents that will produce a constant g_m, and you asked for a voltage.  Do you really need a voltage?  If so, and if you have a low tempco resistor available, you can just drive the current into a resistor.