What is important in using op amp for biasing?

Hi all,

do we need to care about bandwith, max i/p and op/ swing, ICMR if we just use our op amp to fix the voltage only. i.e biasing.

Thanks in advance

op amp for biasing

if you are trying to maintain the voltage constant then ideally you are operating in DC... i dont think you would have worry about these things...

btw what opamp are you using and what circuit and application r u tryin which requires an opamp for biasing???

Re: op amp for biasing

A.Anand Srinivasan said:

if you are trying to maintain the voltage constant then ideally you are operating in DC... i dont think you would have worry about these things...

btw what opamp are you using and what circuit and application r u tryin which requires an opamp for biasing???

Click to expand...

Hi Anand,

Thanks a lot for ur reply.

I am using simple diff amp. I just want to makesure one of the node voltage in my circuit is always at bandgap voltage.

op amp for biasing

oh i see... so you are going to use the opamp to make a bandgap reference right??... there are lots of material available on that... better check out razavi..

Re: op amp for biasing

A.Anand Srinivasan said:

oh i see... so you are going to use the opamp to make a bandgap reference right??... there are lots of material available on that... better check out razavi..

Click to expand...

yes... thanks for the help....

so, can I say for DC biasing, only gain of the opamp is important....

op amp for biasing

as long as you operate it in close loop you wont be into any trouble....

op amp for biasing

The op-amp is used in the bandgap reference circuit. If this is the case than u have take special care of input offset voltage . For a 350nm technology (1.2v) , the typical output offset should be less than 20mv. All these things will work provided DC bias to the circuit is done properly. As the the bias is provided by the resistive network , the bias voltage is well maintained .
For further detrails see Grey & Mayer book

op amp for biasing

And never forget stability simulations whenever having an op-amp in closed loop....

Re: op amp for biasing

charkyriak said:

And never forget stability simulations whenever having an op-amp in closed loop....

Click to expand...

Why are u suggesting stability simulation for closed loop... since he is making bandgap reference he is ideally going to be operating in DC so i dont see any problem with stability.... Am i missing something here??

op amp for biasing

Closed loop stability is a vital parameter for this opamp. The opamp has to be stable at least in the operating frequency range(At least 10 times the operating frequency range) .

op amp for biasing

dude i dont understand why are talking about operating frequency here... the opamp is being used to generate a voltage reference to bias a particular node.... it is a bandgap reference.... so the opamp is ideally working at DC...

Re: op amp for biasing

Obviously an OP in a bandgap reference is operated in a close loop. Thus it can get unstable. This may be simply handled by a generous compensation at first sight.

However, there are several properties of a reference that involve AC parameters for the internal OP to my opinion
- stability with capacitive load
- PSRR over frequency
- transient response to load changes
- noise spectral density

op amp for biasing

k FvM i understand what you are saying.... the op becomes unstable only when there is a 180 degree phase shift in the feedback and i understand your suggestion of internal compensation... the reason i'm confused is that 180 degree phase shift in the feedback will happen only at high frequency given the fact that the resistor forming the output impedance of the op is gonna be very low giving a very low RC constant and how is it possible for the occurence of high frequencies in a bandgap reference which is operating at DC...

Re: op amp for biasing

The power supply is a rich source of all frequencies. See how your circuit settles if you apply a small step (about 20mV) at the supply. This is directly related to the stability of the loop. In practice there are more ways for high frequency noise to enter the circuit which is why you need good stable loop even though your signal is dc. If you bring the poles together within the GBW you can actually see your BGR osillating.
Stability is a must for any closed loop system. Won't you compensate an LDO even though it is supposed to process only DC?

op amp for biasing

A.Anand Srinivasan has a point, the band gap reference is actually operating in DC so how come that it will be affected by the frequency related stability issues.. if your voltage source is a rich source of all frequencies, it is not a good voltage source... and the problem will not the op-amp close loop stability, that is the regulation of the power supply, which can be resolved in a separate way, which will involve the proper filtering and regulation as what we all know...

Re: op amp for biasing

lsimeon said:

A.Anand Srinivasan has a point, the band gap reference is actually operating in DC so how come that it will be affected by the frequency related stability issues.. if your voltage source is a rich source of all frequencies, it is not a good voltage source... and the problem will not the op-amp close loop stability, that is the regulation of the power supply, which can be resolved in a separate way, which will involve the proper filtering and regulation as what we all know...

Click to expand...

The design requires the amp to provide gain near dc., but where are you limiting the amp's amplifying nature?., where are you telling it to amplify only the signal and reject the rest. So what is going to prevent the amp from amplifying the noise that it sees from everywhere?
Good Voltage source? You can have a "good" voltage source outside the chip., but at the BGR, it is going to be plagued by a variety of loads that spit electrons at whatever rate they like. Designers need to be aware of this and create sensible designs that are not sensitive to these.
Even if the supply is clean, the start-up transient itself land your circuit in oscillations. To prevent it you have to ramp it very slowly...at a rate you say that your signals are.., but it's awkward to pass the restriction to the user (at that rate the start-up also may not work)
You assume that the amp's PSRR is going to save you. It's valid only if the amp is stable. Every property of the negative feedback can be taken for granted "iff" the loop is stable.

Re: op amp for biasing

saro_k_82 said:

The design requires the amp to provide gain near dc., but where are you limiting the amp's amplifying nature?., where are you telling it to amplify only the signal and reject the rest. So what is going to prevent the amp from amplifying the noise that it sees from everywhere?
Good Voltage source? You can have a "good" voltage source outside the chip., but at the BGR, it is going to be plagued by a variety of loads that spit electrons at whatever rate they like. Designers need to be aware of this and create sensible designs that are not sensitive to these.

Click to expand...

can you rephrase your questions? i cant understand what you are trying to point out...

Re: op amp for biasing

There is no nice supply as such inside the chip. This is possible only if you have all the circuits consuming constant currents all the time and the current loop area is zero!(or the current itself is zero!). So all circuits need to tolerate the supply noise to a certain degree.

Assume that the supply is clean and your circuit works from -inf so that you don't have even start-up transients.
Every device you have in the circuit is noisy. Higher your bandwidth., more noise you accumulate in to your circuit. The poles and zeros in the loop decide the way in which these noise are shaped. If you move the poles together, you may end-up having a high-Q circuit for noise at UGB.
The problem is that you assume that you are dealing with only dc signals, but fail to see where that is validated. There is a big spectrum that you could have made use of., but you chose to operate near dc., thats fine as long as you understand this and have a very low frequency dominant pole to roll-off the OL gain beyond your signal range. If you inject 1MHz signal somewhere in the loop, how do you see the signal getting rejected or die-down?

You need to phrase your questions as to where exactly you have troubles understanding this.

Thanks,
Saro

Re: op amp for biasing

very nice explanation saro, but have you tried to prove your answer practically or experimentally? your answer is interesting but i found it very bookish...

Re: op amp for biasing

What saro_k_82 says is totally real. Experimental and practical 100%.

If you are using an opamp in closed loop and it is not stable for the entire frequency range you are assuming the risk of having oscillations.

Imagine that your opamp is unstable at 300MHz. During start-up phase of your circuitry (and you always have to start things up ) you will see transient signals with harmonics at 300MHz that can trigger oscillation. See that you don't even need a signal at 300MHz, just harmonics that would be amplified in phase, fed back and again amplified, etc.

You better consider stability for your opamp and OF COURSE for your bandgap. 8)