If you have a related question, please click the "Ask a related question" button in the top right corner. The newly created question will be automatically linked to this question.

• TI Thinks Resolved

# TL431: TL431 Temperature Drift

Part Number: TL431

The temperature deviation of the reference is positive.  Do you know if the reference always moves in the same positive direction with increasing temp?  We use this part to set up a stable droop characteristic and a customer wants us to do a worse case analysis.  They pointed to the drift of our reference as a possible issue but I assured them that the drift from part to part goes in the same direction with increasing temp.  Just want to verify.

• Hi John,

As can be seen by figure 1 from the datasheet (below) that the output voltage does not necessarily rise with temperature.

The temperature variation is described as the difference between the min and max value obtained. The max value can appear anywhere, and on a typical part it can be around 25C.

Regards, Diego Lewis

• In reply to Diego Lewis:

Hi Diego,
The customer reply is given below. From Figure 1, it looks like the slope is a function of the reference voltage. If one reference is at the highest limit, it can see a positive slope with temperature while another device with a lower voltage, can have a negative slope with temperature. Can you please comment.
Thanks,
John

From the customer:
Brick (DC-DC power module) specified is over -40C to 100C baseplate. Interested in direction, not magnitude of reference voltage. As long as 2 bricks with 2 different references when heated have the reference voltage go in same direction is what I am after.
• In reply to John Wallace1:

Hi John,

The curves are provided as typical information only, so while they show the approximate behavior, they are not guaranteed.

That being said, the curves in figure 1 above show the severe limits. Most parts will look close to the 2.495V case. If there is a concern with the shape of the curve, I'd recommend using the B grade parts, as they will have a tighter tolerance and distribution.

Regards, Diego Lewis
• In reply to Diego Lewis:

Hi Diego,

Thanks but not quite yet. Ask your factory if two different TL431's were heated up to say 85C, would the expected drift in Vref for both be in the same direction? The magnitude is covered in your spec. I just need to know that both go in same direction, for example positive temp coefficients. If one drifted up and the other down, our current balancing falls out of whack and we do not meet Cisco's spec. Test data shows that the references go in same direction, but they are asking if that would be true if we sampled a 1000 parts.

Thanks,
John
• In reply to John Wallace1:

Hi John,

We don't have any data on the shape of the temperature variation on a large set of TL431s.

As I mentioned above, it is likely that most parts will have a similar shape (Vout decreases as temp increases past 25C), but we cannot guarantee the shape.

Regards, Diego