AD8599 Operational Amplifier is running hot.

We've measured temperatures of 90C-100C on the case at room temperature. The part itself can withstand a 150C junction temperature, but there are two of these op-amps heating up our entire board. Other devices aren't running optimally with all of the heat around.

We need to run the op-amp from +/-12V; in order to have the proper voltage range for the outputs.

The each part is a dual package with the first op-amp driving the other, and the second op-amp is driving the gate of a FET. The FET drive is analog (as in holding the gate voltage to vary the channel resistance); so it's not switching. The signal levels are very small.

I'm thinking that there must be a part with similar noise/drift specifications, the same SOIC 8 footprint and a lower power dissipation.

???????
What is the power dissipation method that's cooking this part?
Is that dissipation method something inherent to its design (i.e. it's a mostly Bipolar design) that another op-amp (i.e. FET-based) would have lower power?

According to datasheet, supply current and thermal data, you'll expect about 30K temperature rise. I presume you are working at normal ambient temperature, so there must be a different reason why the device is running hot, preferably some high frequent self oscillation of your circuit.

FvM said:

According to datasheet, supply current and thermal data, you'll expect about 30K temperature rise. I presume you are working at normal ambient temperature, so there must be a different reason why the device is running hot, preferably some high frequent self oscillation of your circuit.

Click to expand...

There are no measurable oscillations in the input or output signals. That was one of the first things that we looked for.

I'm wondering if we meet the assumed criteria for ground plane size, thermal vias and such that the data sheet specifies.

There are many times that I've read a data sheet that specifies a power rating or temperature rise, and then the small print says something to the effect of "6 square inches of circuitboard" is required for specified performance.

That's a non-starter with modern circuit densities. We have two of these parts in less than one square inch.

I perhaps need to review that as well.

The TI OPA2209D is pin for pin compatible. It's the same package, and it specifies a maximum supply current of 2.5mA; whereas the AD8599 has a typical supply current spec of 4.8mA.

I'm getting samples to try out. Going just by that specification, the power could be cut in half.