OPA1652: Operating Junction Temperature for OPA1652

Hi Niels,

Thank you for your post. I think the recommended thermal information you're looking for is found in Table 6.4 of the datasheet on page 6. See image below:

-Tamara 

Tamara,

my question was:
What is the recommended operating junction temperature range for the OPA1652?

In table 6.4 I see a lot of thermal resistances, but no information about the maximum recommended operating junction temperature of the OPA1652.

I am sure that this is just because I am not able to interpret the information from table 6.4 correctly.

Please be so kind and explain to me how I can get the maximum recommended operating junction temperature of the OPA1652 from the information shown in table 6.4.
A appreciate you help, and look forward to your answer.

Regards, Niels

Hello Tamara,

From my understanding this means TI guarantees that the OPA1652 will perform as specified in the datasheet as long as the junction temperature stays below 200°C.

Thank you,
Niels

Hi Tamara,

I am not a native english speaker. But fro mmy understanding there is a "slight" difference between
"you should be fine"
and
"TI guarantees that the OPA1652 will perform as specified in the datasheet"

Do you agree, or are the two statements identical to you, respectively IT?

Regards, Niels

Hi Niels,

Thank you for your reply. I'd like to help clarify this.

We specify a recommended ambient temperature operating range as shown on the datasheet. We don't specify a recommended junction temperature range, that is not something we typically test with our devices and we don't have data on this. We do, however, mention junction temperature in the "absolute maximum ratings" but we also state the following regarding "absolute maximum ratings":

"Stresses beyond those listed under Absolute Maximum Ratings may cause permanent damage to the device. These are stress ratings only, which do not imply functional operation of the device at these or any other conditions beyond those indicated under Recommended Operating Conditions. Exposure to absolute-maximum-rated conditions for extended periods may affect device reliability."

This is why I cannot definitively guarantee this spec. 

To Kai's point, though, why is the circuit getting very hot? Maybe we can evaluate this design. 

-Tamara

Hi Tamara and Kai,

here is our OPA1652 schematic (we have 16 of those on-board). The OPA1652 are all powered with +/-15V.

And the following shows the OPA1652 temperature measured by a thermal camera:

As you can see on the temperature-picture, we have 16 OPA1652 on-board (the Schematic shows only one of the 16 OPA162) and they get quite hot (81.3°C, which is 56.3°C above ambient), the picture is taken at 25°C ambient temperature.

I am looking forward to your suggestions how we can improve the schematic to reduce the OPA1652 temperature.

Regards, Niels

Hi Niels,

please see figure 4 of datasheet. You cannot produce 10V output voltage at 400kHz with the OPA1652:

niels1.TSC

Kai

Kai,

you are wrong. We applied a +/-10V 500kHz sine wave to the OPA1652 when we made this thermal image.

:-)

But we did not expect OPA1652 to reproduce the sine wave exactly. This was just to get a feeling for the max. power dissipation of OPA1652.

With all inputs shorted to ground, we got the following thermal figure:

As you can see, we are already 30°C above ambient.

With a +/-10V 100kHz sine wave applied to the OPA1652, the following temperatures occur:

This is already 45°C above ambient.

If I scale this linear to 85°C ambient temperature, I get a OPA1652 top case temperature of 130°C, with junction temperature still a few degree higher.

Do you agree that this is quite hot?
I wonder if running the OPA1652 at those temperatures is covered by the data sheet specifications.

Regards, Niels

Hi Niels,

the WSON package has a thermal pad. Have you soldered this thermal pad to a solid copper plane which is connected to the negative supply voltage (V-)?

Kai

Hi Kai,
yes, we have soldered the thermal pad and this pad is connected to the VEE copper plane. Due to space constraints, we could use only spend a single via as Thermal-Pad-to-Copper-plane connection per OPA1652.
Regards, Niels

Kai,

this is how the pcb looks like:

There is a ground layer directly below. A lot of traces on other layers prevented the placement of additional thermal vias.

So we are back to the question: "How hot is allowed for the OPA1652"?

Regards, Niels

Hi Kai,
there is a negative supply plane 5 layers below (12 layer pcb). The thermal vias connect to this copper plane.
Your intentions are good, but we have a fixed pcb size that does not leave any room for a wider spacing between the OPA1652, additional vias or a negative supply plane "flowing" around the OPA1652.
Perhaps we can lower the temperature by forced air cooling. But again, I need a target temperature to determine if the air flow is sufficient.

What is the maximum (junction or case) temperature for the OPA1652 to perform as specified in the datasheet?

Regards, Niels

Hi Niels,

the long-term junction temperature should be as low as possible for a high reliability application. But during short periods higher junction temperatures can be allowed. The increase in failure rate due to increased junction temperature is a statistical process and can be estimated by the help of Arrhenius equation. The acceleration factor AF, by which the failure rate increases is:

AF = exp (Ea/k x (1/Tu - 1/Ts))

In this link you can find the failure rate data of OPA1652:

www.ti.com/.../estimator.tsp

Assuming an activation energy of Ea = 0.7eV and taking k = 8.63 x 10^-5eV/K, Tu = 273K + 55K = 328K, Ts = 273K + 200K = 473K yields an acceleration factor of AF = 1960. So, the failure rate will increase by a factor of about 1960, when the junction temperature is 200°C instead of 55°C.

Wafer tests have shown a mean time between failure (MTBF) of 5,350,000,000 hours (60% confidence level). Taking into account the acceleration factor of 1960 translates this to a MTBF of 2,730,000 hours at 200°C junction temperature. So, at 200°C junction temperature the MTBF is about 312 years.

A similar calculation shows that at a junction temperature of 150°C the MTBF is about 2367 years.

Datasheet of OPA1652 specifies 200°C absolute maximum junction temperature. I think that the long-term junction temperature should be well below this limit. I would recommend a maximum long-term junction temperature of 150°C.

And improve your layout! You can at least add more thermal vias. Using only ONE thermal via is no good design practise.

Kai

Kai,

thank you for your very helpful answer.
And yes, we will try to improve the layout as soon as possible.
But due to lack of space (I mentioned this before), this will very probably require switching to microvia process.

There is one last question I worry about:

Does a junction temperature of e.g. 150°C derate the analog performance (slew rate, noise, gain bandwidth, ...) of the OPA1652?

Regards, Niels