PCF8575: thermal consideration

Hi Michael,

This is a good question. There's a lot of information surrounding IC temperature and thermal parameters so it can be tricky to find what values you need based on what is defined in the datasheet. We can look at a few scenarios to see where these parameters apply and when they may be considered an issue.

The first this to consider is that an active device will also be producing some amount of heat. This means that in a steady operating state, the device will be warmer than the ambient temperature and thus be cooled ambiantly. The rate at which the device will transfer heat to its ambient surroundings is given by R-theta JC which describes the thermal resistance between the device's junction to the ambient environment. This unit is described in C/W meaning that the device will be so many degrees warmer than the ambient temperature for every watt it consumes. There are other mediums a device will dissipate heat through such as through thermal pads and pins the a PCB, but let's just focus on ambient cooling for our examples. 

Two very common specifications that appear on device datasheets are Operating free-air temperature (TA) and Junction to ambient thermal resistance (R-theta JC). An important metric that may not appear on the datasheet is the maximum junction temperature (TJ) for the device. This is the Maximum temperature that the device itself (such as measuring with a thermocouple) will permit while continuing to operate normally. Given the two specifications we normally have above, we can estimate a theoretical Max TJ for a device by calculating the maximum specified power consumption. - Note that this method excludes cooling from other sources so actual TJ will vary based on factors such as PCB design and other thermal resistance parameters. 

Now to get to your question. PCF8575 has a Maximum operating free-air temperature of 85C. Knowing that an operating device will always be warmer than ambient, we know that TJ for this device is greater than 85C (85C < R-theta JC * Power dissipated + 85C). Using this logic, we can conclude that the device can operate within normal ranges with a junction temperature of 80C. 

In scenarios where the device temperature exceeds the Maximum ambient free-air temperature, some calculations would need to be done to know for sure. However, if we are to assume that the device is operating below the maximum power specifications and the ambient temperature is not near the specified maximum, it is likely that the junction temperature will remain below its actual maximum. 

If you wish to do these calculations yourself, I would encourage you to consult the application report that is linked in the thermal characteristics section of the device datasheet:
Semiconductor and IC Package Thermal Metrics

Let me know if you have any other questions regarding thermal metrics or if you have concerns about your specific implementation of PCF8575.
Regards,
Eric

[Edit]: Formatting

Hello Eric,

thanks for the information and your help.

I know that this is a very difficult topic, so I have further questions.

In your answer you wrote that:

Now to get to your question. PCF8575 has a Maximum operating free-air temperature of 85C. Knowing that an operating device will always be warmer than ambient, we know that TJ for this device is greater than 85C (85C < R-theta JC * Power dissipated + 85C). Using this logic, we can conclude that the device can operate within normal ranges with a junction temperature of 80C. 

• so if I have a case temperature of 80°C the PCF8575 will operate normally because this temperature is lower than the max ambient temperature of 85°C…right?

For my other working scenario the case temperature of the Chip can rises up to 90°C.

How can proof if the part in this scenario will also work normally!?

In the application report there are two possibilities:

Equation 1

Tj = Tc + (R0jc * Power) 

-> to calculate my junction temperature with equation 1 I need the thermal resistance R_0JC, but this resistance is not in the datasheet.

-> also I need the max. junction temperature T_J of the chip to verify may calculation

Equation 2

Tj = Ta + (R0ja * Power) 

-> with this equation I can calculate the TJ, but I have measured the case temperature…can I use my measured case temperature for T_A? I think not, that`s why I first locked to equation 1!

-> here is also the TJ missing to verify may calculation

So I think I have to use Equation 1 because I measured the case temperature, but in the datasheet there is missing the R_0JC and the max. T_J to verify the calculation, so I can`t use this equation!

But I also can`t use equation 2, because there is also missing the max. T_J and additional I have no reference to the actually measured temperature in my system, because I measured  the case temperature, not the ambient temperature!?

Another point is how to calculate the necessary Power in both equations?

So I come back to my question…how can I calculate if the chip can operate normally with a case temperature of 90°C!?

Thanks and best regards

Michael

[Edit]: removed duplicate text for readability - Eric

Hi Michael,

Tj Max is generally a measured value when it is included on a device datasheet. As this is not a simple measurement to perform, it is not included for all devices. What the above equations are used for is to approximate the Tj Max value in cases where it is not provided by using other specified values on the device datasheet. In the case of PCF8575, values for Ta, R-theta-JA, and max power parameters are included on the datasheet. These values can be used in equation 2 to find our theoretical calculated Tj Max.

For this example: Ta(max) = 85C, R-theta_JA(PW package) = 88 C/W, max power = (100mA) * 5.5V = 550mW
Tj(max) = 85C + 88 C/W * 550mW = 133C.

Using these estimations, the device can be expected to maintain normal functionality as long as the junction temperature remains below this value (with some generous tolerances). 

Let me know if this makes sense.
Regards,
Eric

Hi Eric,

your calculated example makes sense, so now i know that the absolute max. junction temperatureof the PCF8575 is 133°C.

But i can`t meassure this junction temperature. I only can meassure the temperature directly on the surface of the PCF8575 and in my understanding this is called the case temperature.

And this is exatly my problem...Now i know the max Junction temperature but i can`t proof if the case temperature (90°C) of the chip is ok because i have not the R-theta_JC!?

I hope you understand my problem.

Thanks and best regards
Michael

Hi Michael,

I understand. Getting specific values and measurements for thermal parameters can be very tricky because of the sheer number of variables and variance in real-life thermodynamics. That's why much of these calculations are estimated. Unless a system plans to push the thermal limits of the device, it is often not worth the complicated testing to classify strict thermal parameters. 

I can say that the Junction-to-Case thermal resistance will be a much smaller value than Junction-to-Ambient. Typically the temperature of the device junction is within a few degrees of the case. For example, a similar device has a characterized R-theta-JC of 6C/W. Using the values we used above, this would translate to around a 3C difference between Junction and Case temperatures. 

Generally as long as the case temperature is not close to the calculated max Junction temperature, it can be assumed that this limit is not exceeded. In the case where the device temperature is 90C, this is still well below our calculated limit. Let me know if you are planning to operate near the maximum limits of these thermal parameters and what device and package will be implemented. 

I hope this information helps.
Regards,
Eric

Hi Eric,

now i think i have understand.

Let me sum up:

With the equation Tj = Ta + (R0ta * Power) i can calcuate my max. Junction temperature. With Ta of 85°C, R0ta of 61°C/w (DBQ-package) and the max. Power of 550mW i get a max. Tj of 118,55°C.

So my max. meassured Case Temperature of 90°C is well below this max. junction temperature, so the PCF8575 will work normally, right?

But i can now proof the meassured temperature of 90°C with the equation Tj = Tc + (R0tc * Power). Here i use R0tc = 6°C/W, Tj = 118,55°C and Tc = 90°C.
The calculation shows that i need a power of 4,167 W to generate so much heat to reach the max. junction temperature. And that is quit a big current of 758 mA which i never use in my system!
-> So we have proven that the PCF8575 will work normally with an case temperature of 90°C, right?

So now i hope my sum up is correct.

Thanks for your patience.

Best regards
Michael

Hi Michael,

Your use of the Junction-to-Ambient thermal parameters looks good. The second equation also looks good, but the large power numbers may be a bit misleading. Generally you would apply this equation near the maximum ambient temperature so you wouldn't expect such large power consumption to be needed to exceed the max junction temperature. Because it doesn't seem like you will be operating under these conditions (case temp is only 90C), you should be fine here.

"So my max. meassured Case Temperature of 90°C is well below this max. junction temperature, so the PCF8575 will work normally, right?"
Yes

Regards,
Eric