Hello all, I am designing a bandgap circuit using two-stage amplifier but the output I got is about 1.1v ! ,could anyone help me to identify the problem? attached the circuit schematic and vref plot vs temp and vdd variations.

You're using 1.8 V transistors with a 5 V VDD. Not a good idea. The output reference at 1.1 V is ok.

Many thanks.

@Luis Henrique Rodovalho, why should you never connect gates directly to the power supply?

For layout considerations and antenna rules. There are special rules to connect transistors directly to the die pads and very long metal wires, which is normally the case for the power lines. For analog inputs, there are resistors between the transistor gates and pads, so it's not a big problem. Some design rules even dictate exclusive guard rings for transistor gates connected to pads. It's better to solve this problem at the block schematic level than correct it top level after the layout is finished. http://mantravlsi.blogspot.com/2015/01/a-tie-high-tie-low-circuit.html

@Eslam Morsie Your voltage supply should be restricted to 1.8V as you are using 1v8 transistors. Check the voltage difference at the differential amplifier inputs. Any offset voltage is multiplied by ~(30k+5k)/5k=7 in the output bandgap reference.

I would also check the diode model to see if it makes sense. I.e. drop a constant current into your diodes and plot VBE versus temperature. In addition, I would start with an ideal amplifier for debug (high gain e source).

Is there an ideal op-amp in xschem? I searched for it and didn't find it.

@Eslam Morsie, to emulate an ideal opamp, you can use a voltage-controlled voltage source and set the gain to a larger value, e.g. 100,000.

Ok thanks sir @Boris Murmann

@Eslam Morsie I have an analoglib library https://gitlab.com/edalibs/analoglib.git which I use in xschem and it contains an ideal opamp called opamp2. It's dominant pole compensated with a gain and bandwidth of your choosing. Its output will also saturate at the supply voltages for more realistic behavior.

Ok thanks sir @Tom