Opamp current biasing to PTAT

[aryajur]

Opamp Biasing current

Why would we want to have the tail current to bias an opamp to be PTAT?? The gain of a diffamp is gm ro , and
gm = Id/(Vgs-Vth) (Id is the tail current)
Vth has a negative temp coefficient so for gm it makes sense to have Id as PTAT to prevent gm from falling. But ro = 2/λId, that would decrease when we increase Id. Is there ome consideration for the temperature coefficient for λ if there is any?

 

[sunking]

Opamp Biasing current

which is more important? gm or ro?
of course gm.

 

[aryajur]

Re: Opamp Biasing current

Why is gm more important?

 

[Syukri]

Re: Opamp Biasing current

gm is the input gain....

we wanted all the input to be asrobd by the op-amp to have high current amplifying

 

[sixth]

Re: Opamp Biasing current

Why is gm more important?

Becuase the GBW is determined by Gm/C. For an opamp, the constant gbw is more important than the constant gain.

 

[rfsystem]

Re: Opamp Biasing current

If the MOS diff pair is saturation region

gm(difpair)=Ibias/(4*VDSAT)

but if the pair is in substhreshold region

gm(difpair)=Ibias/(4*m*VT)

where m is the subthreshold slope. So if subthreshold region is used the PTAT is a perfect compensation. For the saturation region I think there are modified current bias cells. These are regulating a current trough a MOS matched to the diffpair operating in linear mode such the voltage drop VDS is equal to a fraction of a bandgap voltage.

 

[vbhupendra]

Re: Opamp Biasing current

we use tail current source to minimize the input common mode variation.

 

[opamp741]

Re: Opamp Biasing current

If u are concern about ur UGB then using PTAT is a good idea.
at Slow corner (125 degree) u will have a higher Bias current.
which will cause u a better gm and hence UGB.
while is Fast or typical corner u can meet ur speed spec with
lower current that a PTAT can provide u.

 

[jinxingsun]

Opamp Biasing current

gm also determines the noise!

 

[qslazio]

Re: Opamp Biasing current

gm also determines the noise!

actually gm/Id determines the noise

 

[opamp741]

Re: Opamp Biasing current

gm also determines the noise!

actually gm/Id determines the noise

i am unable to understand ur statement. can u elaborate further.

 

[laglead]

Re: Opamp Biasing current

For an opam input stage, its current gain (gm) does matter in opamp compesation as well as bandwidth. So, its we should make its gm relatively stable with temperature variation.

 

[montage3000]

Opamp Biasing current

to make gm constant

Added after 2 minutes:

make gm does not change with temperature

 

[jordan76]

Re: Opamp Biasing current

......These are regulating a current trough a MOS matched to the diffpair operating in linear mode such the voltage drop VDS is equal to a fraction of a bandgap voltage.

Hi rfsystem,

Could you please elaborate more about the above? Thanks!

regards,
jordan76

 

[elbadry]

Re: Opamp Biasing current

A PTAT current would make gm constant FOR A BIPOLAR DIFF. PAIR ONLY

This is because for a bipolar device: gm = Ic/VT ( VT: thermal voltage and is directly proportional to absolute temperature)

Hence if I is proportional to absolute temperature, the term Ic/VT would "ideally" be constant with temperature...

 

[jeff_zx]

Re: Opamp Biasing current

A PTAT current would make gm constant FOR A BIPOLAR DIFF. PAIR ONLY

This is because for a bipolar device: gm = Ic/VT ( VT: thermal voltage and is directly proportional to absolute temperature)

Hence if I is proportional to absolute temperature, the term Ic/VT would "ideally" be constant with temperature...

but the gain gm×ro is still not constant for bjt?!