LM1084 overheating at just 250mA

[jayanthyk192]

Haha, true. My fingers say really hot.:grin:

 

[d123]

Hi there,

I forget to "check digital multimeter functioning correctly (one exemple: lo bat can give false readings) with a known voltage source before measuring anything, then checking again on known voltage source after" - I bet few people actually do, but it's a thought to avoid erroneous and misleading readings.

If you can, and what you're using isn't around the following value, you could try to get hold of a more or less 5ºC/W TO-220 heat-sink. TO-220 is the package shape of the voltage regulator in your first post.

Flat shaped metal, that sounds like fake coins jingling will be high ºC/W, solid metal and/or with big "antlers" or several thick metal grooves is low ºC/W - the latter are the ones of interest for heat-sinking.

From experience, "small" heat-sink, if without (dissipation) fins may be a high ºC/W (as stated a few posts ago like a 23ºC/W - a bad purchase on my part several months ago), and to exaggerate a little, when not used with other cooling methods like a fan and/or for too high a voltage regulator power dissipation - like Treez mentioned - is almost a worse cure than no heat-sink at all, it just adds more heat to say it that way.

I don't know the ground pin current of the 1084, but from re-reading about different designs for regulators it can range from a few milliamps to adding up to 40mA at full load (which you are far from anyway at 250mA), so doesn't hurt to add a few mA to any calculations for the actual current passing through the regulator.

Everything gets hot (hopefully not when it shouldn't!), a regulator could be very, very roughly described/considered as one big resistor so when you multiply voltage by current you get power dissipation which is heat. Specifically 12V in - 5 out = 7V "burnt up" by the regulator; 7V * 0.25A = 1.75 Watts.

The series resistor is supposed to be on the input side of the regulator, not the output, and it's a solution I don't see any point in as all it does is dissipate the "extra" voltage through the resistor, which also has to handle the current, so you can end up with an "expensive" high wattage resistor doing a part of the regulators job, but it does reduce the regulator dissipation.

A 10 ohm, preferably 2 Watt (1 Watt will do) resistor in front of the regulator input should drop 2.5V, bla bla bla calculation again, that would reduce the regulator dissipation to about 1.125 Watts - a whole saving of over 0.6 Watts thrpough the poor overheated regulator.

You could even stretch to a 15 ohm resistor in front of regulator input and still scrape by using a 1 Watt type if the maximum current to the Arduino is only ever going to be 250mA - and now the regulator would only be dissipating 0.8125 Watts - you might even be able to hold it for a long time (don't do that if connected, please - fingers are very good but also probably not a good idea to go touching electrically "hot" regulator tabs when connected to a power supply).

Good luck on Monday!

 

[Audioguru]

Meanwhile are there any ICs that don't burn off like this one?

There are 5V switching regulators available that stay cool. **broken link removed**

 

[jayanthyk192]

Hi there,

I forget to "check digital multimeter functioning correctly (one exemple: lo bat can give false readings) with a known voltage source before measuring anything, then checking again on known voltage source after" - I bet few people actually do, but it's a thought to avoid erroneous and misleading readings.

If you can, and what you're using isn't around the following value, you could try to get hold of a more or less 5ºC/W TO-220 heat-sink. TO-220 is the package shape of the voltage regulator in your first post.

Flat shaped metal, that sounds like fake coins jingling will be high ºC/W, solid metal and/or with big "antlers" or several thick metal grooves is low ºC/W - the latter are the ones of interest for heat-sinking.

From experience, "small" heat-sink, if without (dissipation) fins may be a high ºC/W (as stated a few posts ago like a 23ºC/W - a bad purchase on my part several months ago), and to exaggerate a little, when not used with other cooling methods like a fan and/or for too high a voltage regulator power dissipation - like Treez mentioned - is almost a worse cure than no heat-sink at all, it just adds more heat to say it that way.

I don't know the ground pin current of the 1084, but from re-reading about different designs for regulators it can range from a few milliamps to adding up to 40mA at full load (which you are far from anyway at 250mA), so doesn't hurt to add a few mA to any calculations for the actual current passing through the regulator.

Everything gets hot (hopefully not when it shouldn't!), a regulator could be very, very roughly described/considered as one big resistor so when you multiply voltage by current you get power dissipation which is heat. Specifically 12V in - 5 out = 7V "burnt up" by the regulator; 7V * 0.25A = 1.75 Watts.

The series resistor is supposed to be on the input side of the regulator, not the output, and it's a solution I don't see any point in as all it does is dissipate the "extra" voltage through the resistor, which also has to handle the current, so you can end up with an "expensive" high wattage resistor doing a part of the regulators job, but it does reduce the regulator dissipation.

A 10 ohm, preferably 2 Watt (1 Watt will do) resistor in front of the regulator input should drop 2.5V, bla bla bla calculation again, that would reduce the regulator dissipation to about 1.125 Watts - a whole saving of over 0.6 Watts thrpough the poor overheated regulator.

You could even stretch to a 15 ohm resistor in front of regulator input and still scrape by using a 1 Watt type if the maximum current to the Arduino is only ever going to be 250mA - and now the regulator would only be dissipating 0.8125 Watts - you might even be able to hold it for a long time (don't do that if connected, please - fingers are very good but also probably not a good idea to go touching electrically "hot" regulator tabs when connected to a power supply).

Good luck on Monday!

Wow, that is one really big reply and a helpful one. Thank you. I'll change the batteries on my Multimeter and check. Any way audioguru's suggestion seems to be better off for now. It hardly heats up since its PWM switched regulator. Anyway, thank you for the advice.

There are 5V switching regulators available that stay cool. **broken link removed**

That's exactly what I was looking for. I never remembered that. PWM is efficient! But do they work well with the board computers? The saw tooth wave can mess up the input voltage since the regulators and filters are smaller and more sensitive. So will it work?

 

[Audioguru]

If the 5V switching regulator module is designed properly then it will work fine. But ebay sells some poor quality stuff that is designed "over there".

 

[jayanthyk192]

If the 5V switching regulator module is designed properly then it will work fine. But ebay sells some poor quality stuff that is designed "over there".

I took my chances and already bought one. If it doesn't work, I'll redesign the circuit.