Fully Differential Folded Cascode OTA: 50dB through 400Mhz !

folded cascode ota

Hi forum.
For my exam in Analog Design I was assigned to Design a Folded-Cascode OTA with the following Specs in 0,35um tech:

CLoad=5pF
Open LoopGain = 50 dB
Gain Bandwidth product =400Mhz
Phase Margin =45
Input Offset Voltage = -+15mV
Peak2Peak Output Range = 0,8V
Iref= 10uA
Vdd=3,3V

So.
Nothing else
The fact is I don't know where to start.
Please don't tell me to study/search in books.
I have searched Razawi,Allen the gm/Id method, Sedra,
I have surfed this forum to see if someone had the same issue.

The fact is following:

Do you think I have all the data needed or something is missing.
Cannot find the starting point because something I think is missing...
I know I am doing something wrong..but what?

Can you please provide a small flow-chart to help me?
Or at least what to do first?
My teacher is on holiday and noone can help me at school.
Any help would be VERY appreciated.

Greets from Italy


Peter

folded cascode parasitic pole

1. The offset voltage the first things. The main sources of the one mainly input transistor. So, the first choose WxL.
2. There is no input voltage range. Depending on one you have to choose wether input NMOS or PMOS to use.
3. Also flicker noise and thermal noise are mainly depends on the input transistor. You don´t have data about it. So, it is not important,
3. Determine gm of the input pair gm= GBWP x Cload and tail current. In fact you have to obtain one-pole system. However, be aware about parasitic pole of a cascode.
4. If you need equal fronts of the step signal you have to make the same current in the folded transistor.
5. Choose the folded cascoded transistors and biases to satisfy output voltage range. Current sources in a folded cascod should with smal gm and strong inversion. Cascod transistors with minimal length and might be in a weak inversion.
6. If you don´t have enough gain use gain boosting

fully differential folded cascode ota

Thanks alot.
I am beginning this way.
Your help was Very appreciated

study of folded cascode

I would recomend start from single-ended amplifier to exclude the problem associated with CMFB. Then you can easy transform the amplifier to a fully-differential one.