Power supply (components) "heat" questions/issues

GroupDIY Audio Forum

Help Support GroupDIY Audio Forum:

This site may earn a commission from merchant affiliate links, including eBay, Amazon, and others.

jwhmca

Well-known member
Joined
Jan 28, 2009
Messages
931
Location
USA
Hi Guys,

Rectifiers - caps - regulators- getting hot...

I have built an external power supply for a multi channel micpre. Everything is working fine, sounds fine etc... I just have a few questions concerning the heating up of some of the parts in the power supply.

4 voltages +/-15vdc - 12vdc - 48vdc.

Each supply works approximately the same way-
1. Transformer feeds -
2. RB153 1.5A 150c max rectifier.
3. CRC 1000uf-10R-1000uf.
4. LM317/ 337 (well heatsinked/compound)

The main issue (might be non-issue)

The 12vdc supply under 500mA load heats up the RB153 rectifier to my guess of "HOT!" The kinda hot where your finger hurts after 3-5 seconds and you feel the NEED to remove it. At lower loads it doesn't heat as much... duh.

I didn't bother to measure the exact temperature, because I'm sure it's not the 150c it says it can run, and I know it's not the 1.5A it says it can do.

I have built many power supplies using the little 1.5A rectifiers, never pulled more than 50mA, so it doesn't surprise me it's getting hot, just first timer questions... Thoughts?

The CRC capacitors are running warm to WARM! as well,  Panasonic FC, with 2A ripple. The upstream (of the 10R ) cap seems to run twice as warm. It seems like the should be splitting the heat(?) maybe take the 10R out and run them in true parallel?

The +/-15vdc supplies are running about 50-60c under a load of 250mA per side

So, really my main question on all this is how to measure/concern yourself with heat.



 
While not very scientific the old spit test (if spittle boils away) the package is more than 100'C. Don't spit on you circuits, just touch a wet finger to the device and see what happens.

100'c is too hot for external heat sinks, but allowable inside the package, while cooler is always better.

JR
 
Hot capacitors is worrisome.

Are you SURE they are not just picking up heat from adjacent hot parts? (Like that 2.5 watts of heat in the resistor? Mount it on long leads so you can tell where the heat is really coming from.)

Proper voltage? (If a lot high, the caps will soon BURST excitingly.)

Yes, they are rated for ripple induced heat. But in audio, we almost never want high ripple, and don't work very near the ripple current rating.

> The upstream (of the 10R ) cap seems to run twice as warm.

That would make sense (if it IS ripple heat). The 2nd cap works with lower ripple than the 1st. 10r between two 1000uF is not much ripple reduction, so maybe one is twice the other. 1,000uFd is not over-generous for a 500mA load, but it should be "OK".
 
This is the 12v supply for things like LEDs and Relays... no audio.

PRR said:
Are you SURE they are not just picking up heat from adjacent hot parts? (Like that 2.5 watts of heat in the resistor? Mount it on long leads so you can tell where the heat is really coming from.

Maybe, I'll check when I put it back together... I only rated the 10R between the 1000uf's at 1/2watt... how are you getting 2.5w?

PRR said:
Proper voltage? (If a lot high, the caps will soon BURST excitingly.)

50v caps,  I think we are good here...

PRR said:
That would make sense (if it IS ripple heat). The 2nd cap works with lower ripple than the 1st. 10r between two 1000uF is not much ripple reduction, so maybe one is twice the other. 1,000uFd is not over-generous for a 500mA load, but it should be "OK".

Remove the 10R? Caps go to 2000uf better all around? It's just relay and LED... Then truly split the heat/ripple? Increase the 10R?
 
> I only rated the 10R between the 1000uf's at 1/2watt... how are you getting 2.5w?

"The 12vdc supply under 500mA load"

500mA is 0.5 Amps.

0.5 Amps through 10 Ohms is 0.5A*10r= 5 Volts.

5 Volts times 0.5 Amps is 2.5 Watts.

My abacus gets sticky in the rain, check my math.

For this specific question, there is a shorter formula I^2*R. I like to do it the long way, because the intermediate results are useful check-points. Fer example, are you getting 5V across the 10r resistor? If not, then some basic "facts" are bogus. Is the resistor really 10r? Is the current really 500mA?
 
Block-ish idea...

Might be better to remove the 10R so as to not drop the voltage going into the regulator?
 

Attachments

  • Untitled.jpg
    Untitled.jpg
    110.2 KB · Views: 11
jwhmca said:
I didn't bother to measure the exact temperature, because I'm sure it's not the 150c it says it can run, and I know it's not the 1.5A it says it can do.

The max temp spec is for the junction of the power device, not for the heatsink or the room temp. And the recommended max junction temp is not 150ºC but 125ºC, So having 100ºC on the back of the case and 3ºC/W to from the junction to that point allows 8W, or 16V drop at 0.5A, then the heatsink will be even lower temperature. I don't know what are your transformers, maybe you have too much voltage going to your regulators, having bigger caps and lower voltage, so lower ripple and you need lower drop on the regulator make it heat up less. You need just 3 volts from the lowest part of the ripple voltage.

You usually need 5 times the current rating for your diodes than the current on the load, in order to avoid excess heat and manage comfortably the peak currents on the smoothing caps. 3 times as is your case will work but they will get hot. You should check the resistor behavior but I think may be better to take it out. This probably makes your diodes even hotter so consider going for some bigger ones. Measure the DC voltages at each point of the PS, so, at each cap and at the regulator. Also the AC voltage at each cap will give more information.

JS
 
joaquins said:
jwhmca said:
I didn't bother to measure the exact temperature, because I'm sure it's not the 150c it says it can run, and I know it's not the 1.5A it says it can do.

The max temp spec is for the junction of the power device, not for the heatsink or the room temp. And the recommended max junction temp is not 150ºC but 125ºC,
I recall reading 150'C as the target max. My understanding that was related to the (plastic) package integrity, as silicon can survive up to 200'C or so.
So having 100ºC on the back of the case and 3ºC/W to from the junction to that point allows 8W, or 16V drop at 0.5A, then the heatsink will be even lower temperature. I don't know what are your transformers, maybe you have too much voltage going to your regulators, having bigger caps and lower voltage, so lower ripple and you need lower drop on the regulator make it heat up less. You need just 3 volts from the lowest part of the ripple voltage.

You usually need 5 times the current rating for your diodes than the current on the load, in order to avoid excess heat and manage comfortably the peak currents on the smoothing caps. 3 times as is your case will work but they will get hot. You should check the resistor behavior but I think may be better to take it out. This probably makes your diodes even hotter so consider going for some bigger ones. Measure the DC voltages at each point of the PS, so, at each cap and at the regulator. Also the AC voltage at each cap will give more information.

JS
Diode data sheets list peak and average current. Diode heating is pretty much about average current.

For one off DIY build it and use the wet finger test.

JR
 
My LM337/317 heatsinks are getting up to about 65c

If I slightly blow on them they drop down nicely, wondering if I should be okay with 65c or get some air moving in there?
 

Latest posts

Back
Top