Ambient vs Junction Temperature For ICs?

[dknguyen]

Hello. For ICs and the like, they have something called operating temperature (which I assume is also the same as ambient operating temperature), but there is also another spec called the junction temperature which is usually 25C to 50C higher than the operating temperature. What is this? The temperature of the leads? of the silicon die?

Thanks.

 

[Hero999]

The temperature of the die, it's lower than the ambient because powering the IC produces heat which warms the die. The operating temperature depends on the, maximum junction temperature, thermal resistance of the die to ambient and the maximum power dissipation, the latter is also linked to the maximum junction temperature.

 

[philba]

ambient is the temperature of the location where the case sits but it usually isn't spec'd.

Operating temperature is usually for the junction. specified as a range.

Now, you need to take the thermal resistance into account to make these number mean anything. Thermal resistance junction to ambient is for when there is no heatsink. multiply that times the wattage dissipated and add to ambient temp to get max temperature. If it exceeds the operating range, get a heat sink. I usually try for something much much less than the top end of the operating range.

ambient temp does come into play here. especially if the device is in an enclosed space (like a project case...). It will raise the ambient temp and possibly cause a runaway condition if you aren't conservative in your ratings..

 

[RadioRon]

Here's a bit of background:
https://en.wikipedia.org/wiki/Thermal_resistance_in_electronics

 

[Papabravo]

In thermodynamic terms the ambient temperature is the temperature of the surroundings. It will typically be lower than the case temperature, and the junction temperature. This is the important point; the ambient temperature is not significantly affected by the chip getting hot. That is the surroundings have the ability to absorb large quantities of heat without the raising their temperature.

Naturally the higher the ambient temperature, the less heat the surroundings can absorb before the maximum junction temperature is reached.

 

[philba]

papabravo, consider the situation where the chip in question is in a small enclosed space. it will raise the ambient temp. that's why I said you need to be fairly conservative in your thermal design.

Ron, that wikipedia entry has what looks like a significant error. I think they omitted the heat sink's thermal resistance. they never say heatsink but imply it with the heat transfer pad. 28 watts is probably the maximum dissipation with a perfect, infinite heatsink (thermal resistance of 0).

the same calculation for a to220 in free air would be something like 45/65 or about 2/3 Watt. however, I would never plan for a to220 to run at 125C. Also, ambient of 70C is way high. I would think more like 30-35C. Not sure who did that wikipedia entry but it's kind of misleading - 28 watts through a TO220, I don't think so.

 

[Papabravo]

I always consider the small enclosed space. In them, higher ambient temperatures give you less fredom and ability to disipate large amounts of heat. In this case you are most interested in keeping the temperature rise small so you can stay under the maximum junction temperature.

 

[instruite]

Ron, that wikipedia entry has what looks like a significant error. I think they omitted the heat sink's thermal resistance. they never say heatsink but imply it with the heat transfer pad. 28 watts is probably the maximum dissipation with a perfect, infinite heatsink (thermal resistance of 0).

The guy who has written the wiki has consider the heat resistance of the heat sink by considering deltaTHS of 10C. and this is the normal way of calculations because from heat sink to ambient the heat dissipation is due to radiation which is normally determined by experimentation rather than calculation. Typically this difference temperature is consider 10C and maximu case it is considered 20C

the same calculation for a to220 in free air would be something like 45/65 or about 2/3 Watt. however, I would never plan for a to220 to run at 125C. Also, ambient of 70C is way high. I would think more like 30-35C. Not sure who did that wikipedia entry but it's kind of misleading - 28 watts through a TO220, I don't think so.

The ambient temperature in case of industrial/commersial equipment is till 70C in case of automotive application it is 85C normally (and in some cases even 125C)

Basically both the specifications about ambient temperature and maximum junction temperature are important in case of designing for commercial applications
Ambient temperature determines the operatibg temperature of your device/unit if this is increased beyond spec the package of the IC/device might be damaged
Maximum Junction temperature is used to determine how much maximum power should be dissipated by the IC/device such that it will not exceed this spec which may result in burning/damage of the IC/device

 

[philba]

that still makes no sense to me since is assumes that the heatsink will always be less than 10C above the ambient. I still think it makes the assumption that the heatsink has a thermal resistance of zero. 28 watts into a heat sink is going to warm it up pretty fast. Anyone who reads that entry and comes away thinking he can push 28 watts through a to220 will soon be letting the smoke out.

 

[instruite]

that still makes no sense to me since is assumes that the heatsink will always be less than 10C above the ambient. I still think it makes the assumption that the heatsink has a thermal resistance of zero. 28 watts into a heat sink is going to warm it up pretty fast. Anyone who reads that entry and comes away thinking he can push 28 watts through a to220 will soon be letting the smoke out.

It doesnot assume heatsink temperature to be 10C
it assumes difference between ambient temperature and heatsink temperature to be 10C so if ambient is 70C then heatsink will be at 80C

So it doesnot assume heatsink thermal resistance to be 0 (if you calculate from power and assumed temperature difference) the thermal resistance of the heat sink will be 10C/28W = 0.36 C/W or K/W

In my project I have operated the device at 105C with all my ICs runing at max power (different packages not To220 package) and the temperature at heat sink was way less than 20C (was around 14C approximately)