LT3080 output offset voltage

[marekpasek]

Hi,
Recently I am considering building a power supply. When I was browsing the datasheet of lt3080 I encountered one specification which I am not sure how to deal with. I am talking about output offset voltage. The datasheet tells that the offset voltage typically varies from minus 6mV to plus 6mV. Well, I dont actually know what that means. Is it a set constant of one particular unit which stays the same no matter what voltage I set (I set 1V, output is 1.006V and I set 10V, output is 10.006V)? Or is it a constantly varying offset which is dependent on the set voltage (I set 1V, output is 1.004V and I set 10V, output is 9.997V)?
Another way ,can it be compensated in software just by adding/substracting the value of offset I measured on one particular unit?
Btw. I know that the offset is negligible anyway, but I am sort of perfectionist and if I was seeing the offset constantly on my power supply's voltage reading it would be teasing me incredibly. 
Thanks

[Audioguru]

The LT3080 is a variable voltage regulator, not an opamp. YOU set the output voltage and the "set pin" has a small offset voltage range that is amplified if your output voltage is high. But the parts you use to set the output voltage probably have a much wider range unless you adjust them with a potentiometer.

The output voltage of a voltage regulator changes a little when its temperature changes and also the output voltage drops a little more when the load current increases.

[Cerebus]

Actually, if you look at the simplified block diagram for the LT3080 (page 1 of the data sheet) it is practically an op amp plus an emitter follower with 100% negative feedback that sets the gain to one.

So in this case, yes, the offset voltage is specifying just the same thing as it would with an op amp. So, if you supply a 1V input to the SET pin then you'll get out a fixed voltage that is between 0.994 V and 1.006 V. It will vary from part to part, but will stay the same for the same part.

The offset will change a little bit with temperature (+/- 500 uV, graphs on page 5 of the data sheet), and with load current (0 to -750 uV), and a negligible amount with output voltage (less than 100 uV over the full range) and will drift a little with time (not specified on the data sheet but probably in the order of  1 to 10 uV per month with a diminishing drift with time, it eventually settles down).

[David Hess]

Is it a set constant of one particular unit which stays the same no matter what voltage I set (I set 1V, output is 1.006V and I set 10V, output is 10.006V)?

It is a fixed value which varies slightly with temperature.

Or is it a constantly varying offset which is dependent on the set voltage (I set 1V, output is 1.004V and I set 10V, output is 9.997V)?

That is common mode rejection.  The offset also varies with the power supply voltage which is power supply rejection.

Another way ,can it be compensated in software just by adding/substracting the value of offset I measured on one particular unit?

Sure.

Btw. I know that the offset is negligible anyway, but I am sort of perfectionist and if I was seeing the offset constantly on my power supply's voltage reading it would be teasing me incredibly. 

An external error amplifier could be used to compare the output of the LT3080 to a reference removing the effect of its offset voltage.