This thread has been locked.

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.

CSD19537Q3: Junction to case 1.5°C/W parameter question

Part Number: CSD19537Q3

Hello. I have a question regarding the parameter of thermal resistance of Junction to case, specified at 1.5°C/W.

The notes state:

RθJC is determined with the device mounted on a 1-in2 (6.45-cm2 ), 2-oz (0.071-mm) thick Cu pad on a 1.5-in × 1.5-in (3.81-cm × 3.81- cm), 0.06-in (1.52-mm) thick FR4 PCB. RθJC is specified by design, whereas RθJA is determined by the user’s board design. (2) Device mounted on FR4 material with 1-in2 (6.45-cm2 ), 2-oz (0.071-mm) thick Cu

I would like to confirm my interpretation of the first note. I interpret this note as follows:

1) TI has derived (measured) the junction to case parameter of 1.5°C/W by using the test conditions cited in note 1.

2) "RθJC is specified by design" means specified for the MOSFET design as opposed to the board design: so I conclude that the junction to case parameter of 1.5°C/W is constant for all board designs, while RθJA varies with board design.

Can you confirm my understanding?

IN our applications. we are using a FLIR camera to measure the case temperature of the MOSFETs under load, and then we would like to use RθJC to calculate the junction temperature, based on MOSFET power dissipation.

Thanks, Charie.

  • Hi Charles,

    Thanks for the question. You are correct that in your understanding. We can only control RthetaJC while RthetaJA depends on the board layout and stackup. Please note, RthetaJC specified in the datasheet is measured on the thermal pad on the bottom of the package. The primary path for removing heat from this package is thru the thermal pad and into the PCB. We don't measure or specify RthetaJC to the top of the package. I'd estimated it is in the range of 20C/W to 25C/W based on my experience.

    You can find more information in this blog:

  • Thanks John! In your experience with case temperature measurements, were they based on a thermocouple type measurement or FLIR? We are using a FLIR imaging camera to measure case temperature, Would a FLIR measurement yield a measurement closer to the junction vs a thermocouple on the surface? The FLIR shows a hotter temperature than the immediate surrounding copper of the FET. Thanks in advance! Charlie

  • Hi Charlie,

    For our FETs, TI does not use the FLIR camera to make temperature measurements to calculate RthetaJC specified in the datasheet. For our SMT packages, we use a special PCB design with a hole that allows a thermocouple to make contact with the drain pad (large thermal pad). In your case, you can try to attach a thermocouple to the PCB right next to the package. I know a lot of customers rely on thermal imaging to characterize power devices in their system. In my experience, if you're using a FLIR camera, the top of the case is usually within 5 - 10degC of the junction temperature. This is because a larger fraction of the heat is removed thru the thermal pad and into the PCB. For example, if you're dissipating 1W, you can estimate 70% is dissipated into the board and 30% thru the top of the package. Say you measure the top of the case to be 50degC with the FLIR. Then you can estimate the junction temperature rise above case is:

    1W x 30% x 25degC/W = 7.5degC, Tj = 50 + 7.5 = 57.5degC

    Obviously, these are estimates based on assumptions. I've seen estimates between 70% to 80% of the power dissipated goes into the PCB with a decent thermal design using thermal vias to multiple internal copper layers. This spreads the heat into the PCB. Having a large copper shape connected to the thermal vias on the backside of the PCB will also help dissipate the heat. I hope this helps. I'm no thermal expert and you may have to resort to thermal modeling such as Flotherm or IcePAK to do a more detailed analysis.

  • Thanks again. In one of our applications, we are measuring 60°C case temperature at a worst case condition, I am estimating a junction temperature of about 86.5°C, with an estimated total mosfet power loss of 0.63W. We do not have a lot of copper on the PCA, and we are using 1 oz. Perhaps 2oz can buy more margin if needed. My calculated Rj-a is about 86.5 C/W, which is in between the datasheet figures, so it all looks about right. In our application we only expect it to operate like this for a few minutes at startup, then it will be well below 40C.

    In most other applications (Step motor) we are seeing a worst case temperature of around 47C when at full current, and here again we do not dwell under these conditions.

    I can probably use the FLIR to measure copper temperature very close to the package.