Who can tell me the difference between off-set and bias?

In a analog design biasing is required to keep the transistors in design in strong saturation. Biasing also helps the circuit to operate around a operating point. A small signal input has to be applied around this point only. When a dc bias is applied to the input transistor , it forces the rest of transistor to maintain the predesigned overdrives. The ouput node is also expected to have the same biuas voltage( ICMR curve will give the curve between input and output ). In other word we can say that for zero input , output should be zero. Typicaly output is of the order of few mv. This deviation is called as the output offset
For details refer to Boyce,Baker

im assuming that we are talking about op-amp...
bias is the external source that you are applying to op-amp for it to properly operate while the offset is actually used for you to set your differential gain to be zero...

hi,
offset for input=0 out not zero
bias need for operate circiut

lssdpoly said:

Who can tell me the difference between off-set and bias?

Click to expand...

Good question. There are, indeed, cases in which "offset" is used instead of "bias".
For example, when a real BJT is compared with an "ideal BJT" which can be approximated by a special analog IC called "current conveyor (CCII)".
In articles comparing both parts you can read that "one of the advantages of the CCII is the lack of any offset voltage at the input terminals", which means that the bias voltage between both terminals is practically zero.

Summary: There is no clear distinction between both expressions. Offset is a difference of two voltages, which - in some cases - is used to bias an active device.
But in other cases (like for opamps) it is simply a deviation from the ideal case as mentioned by the former contributions.

IMHO, bias is part of the design. You need to take in account it when you design a circuit in order to properly set transistors in the correct region.
Offset is something unwanted and usually not predictable (due to mismatch, gradients, process variations) which may affect circuit behavior.

dalraist said:

Offset is something unwanted and usually not predictable (due to mismatch, gradients, process variations) which may affect circuit behavior.

Click to expand...

I agree - as far as, for example, opamps are concerned. In this case, offset is an unwanted voltage difference between ideal and real.

However, in general this is not true ! Offset is something like a difference - and not necessarily restricted to voltages. Think of a "frequency offset" for instance.

This difference sometimes is unwanted (as mentined above), but sometimes it is simply a result of electrical/physical properties (one example was mentioned in my first reply). Here is a second example: The control input for an OTA (gm steering) has an "inherent" voltage of app. one or two pn-junctions - and is called "OFFSET".

Both biasing and offset are unwanted, and both can be stochastic in nature. However we have to live with it.

If you look to the inputs of an opamp:
a biasing current of max 5nA would means that both inputs draw a 0nA to 5nA current. This is a common mode current. A pcb designer can cope with it by making the impedances of both inputs equal.
This biasing current is for instance caused by the input of an amplifier (base current, gate leakage, chopper amplifier), or protection esd diode leakage.

An offset current of max 5nA would means that the inputs draw a differential current of -5nA to 5nA. A PCB designer cannot really design it out. He has to cope with it or trim each part.
This biasing current is for instance caused by MISMATCH in the input of an amplifier (base current mismatch, gate leakage mismatch), or protection esd diode leakage mismatch. In a well defined process mismatch data is available so a designer can do something.

drDOC said:

Both biasing and offset are unwanted, and both can be stochastic in nature. However we have to live with it.

Click to expand...

Hi drDOC,
I cannot see that "biasing is unwanted".
If I think of a simple BJT which is powered only with a positive voltage, it is a very good property of the device that it can be biased with a certain voltage.
Only than it can be used to amplify both half waves of a sinusoidal signal.
Or did I misunderstand your arguments ?

offset is the dc voltage level which is added along with an AC signal ....and it is depicted when the output signal level is a bit raised from normal zero.......bias is a voltage level or current which is essential for the BJT or MOSFET transistors to turn on...or reman in the desired linear or amplification region....for proper working

I cannot see that "biasing is unwanted".

Click to expand...

Yeah we are talking about two different things. I am talking about bias currents and offset currents which go into the input of an amplifier, while an ideal opamp or nullor has zero input current.

In your simple BJT case this BJT has an input current of Ic/beta. So when the beta of your BJT spreads the input bias current also spreads. Anyway if you compare your single transistor BJT amplifier with an ideal amplifier you'll notice that this input current is non ideal behavior.

Hm. Lots of factoids in the answers without a clear picture. Let me have a go at it.

Bias: This is required to put the transistors into their operating region. On top of the bias we apply a signal. Without the bias, the transistor cannot operate (at least not for linear amplifiers, class E and digital is a different story). For instance on a single transistor amplifier the gate bias voltage sets the operating point while the signal is modulated on top of the bias. (lets say 1V bias and around that you can apply a sine wave of 100mV amplitude which gets amplified).

Offset: Usually this refers to an unwanted property. Lets say an ideal comparator with two inputs changes state if one input is fixed and another input crosses the same voltage. In the presence of an offset voltage, the change in the output will not occur at V1=V2 but at V1=V2+Voff.

I believe you may have been reading the datasheet of an opamp with bipolar inputs. In this case they refer to the input currents caused by the base emitter diode as the input bias currents.

In a analog design biasing is required to keep the transistors in design in strong saturation

Click to expand...

This is incorrect. I have designed lots of amplifiers with transistors in subthreshold and transistors for compensation in linear region. All of these are biased. For some reason people seem to mistake strong inversion with saturation a lot but that's a different issue.

bias is the external source that you are applying to op-amp for it to properly operate while the offset is actually used for you to set your differential gain to be zero

Click to expand...

Wrong. The differential gain will always be the same, you set the output voltage to zero (or mid point for single supply amplifiers). Vout=Adiff(V1-V2+Voff). In that equation Adiff stays constant.

dalraist: Very good answer. I would like to add that offset usually consists of stochastic (random) offset and deterministic (systematic) offset. The latter can be eliminated with perfectly symmetrical designs and layouts, but believe me a lot of thought goes into this, especially for high current stuff. Random offset can be minimized (device size, layout) but not eliminated except by system design (autozero, chopper, CDS).

drDOC: I read your posts too late, I could have saved myself a lot of typing

hey "agunther" that was a very long bookish answer huh... in your equation Vout=Adiff(V1-V2+Voff) what do you think is the function of Voff there?