I built a regulated tube filament power supply using an LT1084CP regulator. I tested it at 600ma and 6.3vdc. The heatsink temperature rose to 135 degrees Fahrenheit/57C in a few minutes and steadied there. The final load will be 1.2A. The data sheet lists the max temperature of the regulator as 150C. I plan to test it at a higher load but wanted to see if I am ok on the temperature so far.
Is this about right?
Is this about right?
I am just using a thermal pad with no paste. For me it would be too much guesswork trying to do the calculations. I probably just need to get a better heatsink.
First try a "very thin" layer of thermal paste on each side of the thermal pad.
That should help significantly. Torque down the mounting bolt fairly tightly.
How much higher current do you want?
If the input voltage is over 9V, you could drop the excess with a power resistor.
Add a large capacitor to ground after the resistor.
That should help significantly. Torque down the mounting bolt fairly tightly.
How much higher current do you want?
If the input voltage is over 9V, you could drop the excess with a power resistor.
Add a large capacitor to ground after the resistor.
Last edited:
At what ambient??
Under some assumptions, you are cool. Say ambient is 25C. Then you have 32C rise. And double current could be 62C rise, 87C. For class B thumpathump audio we learned to try for 50C rise or fatigue would crack the seals in the field. A steady heat load, 62C rise may be fine and 87C is in spec.
OH, these numbers really apply to the chip not the package or sink. For a small sink there may be little difference. If it is hot on a big sink you want to do some math.
It also would stress the 6.3V winding of the mains transformer much less, because with the pulsed current when rectifying. The current peaks are several times(7-10x) higher than the average, especially with big caps to load. This might drive the transformer into saturation and it will emit a humming sound.
In this case you insist on DC feed, get a separate 7.5V/20W transformer.
The regulator heat sink should have 5°/W, so enlarge it and mount it in free airflow or vertical against the steel frame.
Exactly, transformers do not saturate from load current but from too much magnetization = volt-seconds on the primary. Choke/filter coils on the other hand do, because the load current is the magnetizing current.
Temp rise is proportional to power loss on a first order, but cooling gets more effective for bigger differences because A) hotter air rises faster (more convection) and B) radiative cooling is proportional to the forth power of absolute temperature. You should get less than 64C rise, more like 58-60C, also because the unregulated supply will droop a little more at 1.2A.
You should be safe at twice the current, but I recommend a good insulator pad (there is a 1:5 spread in conductivity between some) and a little (!) grease.
And never forget it is rise above LOCAL ambient. Once in a chassis, with 60C inside air, your sink is at 120C and the chip even hotter!
Temp rise is proportional to power loss on a first order, but cooling gets more effective for bigger differences because A) hotter air rises faster (more convection) and B) radiative cooling is proportional to the forth power of absolute temperature. You should get less than 64C rise, more like 58-60C, also because the unregulated supply will droop a little more at 1.2A.
You should be safe at twice the current, but I recommend a good insulator pad (there is a 1:5 spread in conductivity between some) and a little (!) grease.
And never forget it is rise above LOCAL ambient. Once in a chassis, with 60C inside air, your sink is at 120C and the chip even hotter!
Ambient is 23C. I am going to change a few things based on the input here. Add a little paste and lower the input voltage to just what is needed. Right now it is a little high (12vac into the rectifiers, 15.8 into the LT1084).
I am just playing around with a circuit right now, don't know if I will actually use it yet. I spent 2 years in electronics school in the mid 70s, I passed the associate CET and licensing exams back then but got out. Now I just build/modify and play with circuits for a hobby.
Thanks everyone
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.