I’m working on a system design and thermals are a concern with the LMS7002M’s commercial temperature range. First question: is the -40°C - 85°C temperature range ambient or junction?
To ensure we stay under TJmax, we need the package thermal resistance. Is there any θJA, θJB, and θJC data available? I apologize in advance if I missed this in the documentation, but would appreciate if you could point me to it in that case.
Our design will be conductively cooled through the top and bottom of the package with no airflow, so θJB and θJC matter the most to us, but we will take any information available.
Assuming worse case thermally, heat is only dissipated from the heat spreader at the bottom of the LMS7002M package, and then passes only through the vias underneath the heat spreader to the rear of the PCB, and then to an exposed area of copper nikel area on the bottom of the PCB under the chip, as done on the LimeSDR Mini PCB https://myriadrf.org/projects/limesdr-mini-2-0/_images/LimeSDR-Mini_v2.2_3D_bot.png
The LMS7002M typically dissipates about 1W of heat in operation - see datasheet for example scenarios.
The package manufacturer’s guide number is 21’C/W for active junction to the bottom of the package.
Through plated via (TPV) 1mm diameter drill hole in 1.6mm PCB, with 20u of copper plating, gives about 60’C/W/TPV. So 20 or so vias is sufficient.
The exposed area, approx 5mm x 5mm, is usually sufficient to convect/radiate away the heat.
As I say, this is the worse case estimate, real life should be a little bit kinder.
The original LMS7002M evaluation board was run by one of my colleague in a thermal chamber at 110’C without problems. (Test was limited by wire insulation melting above that).
Thru vias aren’t an option with the density of our design, but we will run a full thermal sim with the 21°C/W θJB and our complete layout including other dissipators. We’re using analog IQ with the TSP and ADC/DAC powered down, so our PDISS within the LMS7002M should be manageable.
Based on your colleague’s test, the 85°C max temp is neither ambient nor junction? Was this just some kind of accelerated life test? Typically I would expect CMOS to be 125°C-150°C TJmax. Your 110°C ambient with 21°C/W with 1-2W PDISS would align with a 150°C TJmax. Is that accurate?
As you appreciate, our internal design rules and testing procedures included margins, so we can guarantee performance of delivered parts to the usual commercial temperature ranges, -40 to +85. For commercial sold parts, we make no guarantees outside of that range. If you require parts for high reliability, I suggest you discuss your requirements with the sales team.
