AD5165 digi pot - is the graph believable?

On the front page of the data sheet it tells me that the temperature coefficient of the resistance in the digi pot is 35ppm/C and in the table on page 3 it says the Voltage Divider Temperature Coefficient is typically 15ppm/C.

 

No problems so far and the device looks OK but because of my application I have to look for the graph that shows me the temperature coefficient versus wiper position and, from experience the tempco with low codes (wiper close to one end) usually gets seriously worse. What I see is this graph on page 9: -

 

It looks too good to believe so my question is this: -

 

IS THERE AN ERROR IN THE DATA SHEET?

 

For instance, if I look at a slightly different device the AD5280 the front page tells me it is 30ppm/C and in the table on page 3 it says the Voltage Divider Temperature Coefficient is typically 5ppm/C (seriously better than the AD5165) but the graph shows this: -

In other words it is significantly worse at low wiper codes.

 

Can anyone help me with this?

Parents
  • Achim,

    You can't be looking at the same graphs when you said: -

    If you compare the 200k-curve of the AD5280 with the (only) 100k-curve of the  AD5165 the differences are not that big anymore


    When I look at the 200k graph of the AD5280, I see a peak in the tempco at about 15 ppm/degC. When I look at the same graph for the 100k AD5165 is a see a peak of about 6ppm/degC.

    Are you telling me that this is not the case?

    Andy

Reply
  • Achim,

    You can't be looking at the same graphs when you said: -

    If you compare the 200k-curve of the AD5280 with the (only) 100k-curve of the  AD5165 the differences are not that big anymore


    When I look at the 200k graph of the AD5280, I see a peak in the tempco at about 15 ppm/degC. When I look at the same graph for the 100k AD5165 is a see a peak of about 6ppm/degC.

    Are you telling me that this is not the case?

    Andy

Children
No Data