RthetaJC - top, bottom??

[chopnhack]

I noticed that the datasheet of an LDO give the RthetaJC of the top at 16.1 deg C/W and then gives the bottom in two parts, one for the control section and one for the output section being 1.5 and 4 deg C/W respectively.

When calculating for power dissipation and determining heatsinks, are we to use the top since that is closest to the metal tab that would conduct the most heat? Or do we truly have to go through the full equation in the new standard below?
- SPRA953B.

 

[Harald Kapp]

Could you link to the datasheet of the LDO, please?
Rtheta surely is different with respect to different parts of the case due to varying distance and material composition from the source (chip) to the point on the case. I can't remember, however, having seen different Rthetas for different parts of one chip.

Unless you are on a very tight thermal, spatial and/or financial budget I don't think it is worth going the tedious way described in the document you linked to. You'll have to consider how an where the heat sink is mounted, however.

 

[chopnhack]

Thank you Harald, I was trying to calculate out the numbers to see if forced air would be required. I intend to use a solid aluminum chassis, probably about 1/16" thick and attach the LDO directly to it.
Here is a link for the pdf of the LM1086-5. The section is on page 5, 6.4.

 

[Harald Kapp]

I'm afraid you'll have to follow the instructions on page 19, chapter 10.3 of the datasheet.
Essentially you calculate the required heat sink for the control section (which has lower power consunmption) and the output section (which typically has higher power consumption). You then select the lower value of both results and use the matching heat sink.

This is the first time I came across such data, thanks for spotting it.

 

[chopnhack]

Thanks Harald, essentially taking into account the worst case scenario and heatsinking for that.

 

[Harald Kapp]

I got a response from a TI application engineer: The LM1086 is a monolithic chip, no separate units for control and power. Therefore the different data for Rth(control) and Rth(power) do not make sense. TI is going to correct this in a future update of the datasheet.

 

[chopnhack]

Wow, that is awesome Harald! I think I should report their misdrawn bridge diode schematic! In, I believe their 371 datasheet someone on the forum spotted it drawn as a ring modulator

Any prizes for helping? Maybe they can send me a few samples