How to calculate the size of the heat sink

Status
Not open for further replies.
Jump to Latest
I

iinself

Member
Joined 2012
#1
I am planning on reducing 58V to 42V dc using LM317. The 58V is coming from an old receiver. How to calculate the heatsink size for the LM317 - I am estimating, the difference in voltage across LM317 * reguired current = 16V * 4A (2 * LM1875 chips) = 54W of power dissipation at max load.
 
#5
No.

deltaT s-a = Rth s-a * Power dissipation * Derating Factor
deltaT c-s = Rth c-s * Power dissipation
deltaT j-c = Rth j-c * Power dissipation

Ts = Ta + deltaT s-a
Tc = Ts + deltaT c-s
Tj = Tc +deltaT j-c

Tj < Tjmax stated on the datasheet.
For long term reliability Tj < 100 to 120°C

iinself said:
..............using LM317. ...............* reguired current = 16V * 4A .............
Click to expand...
the maximum current of the 317 is around 1.5 to 2A.

But it will run too hot at this current continuously.
You would need to set a continuous maximum current at a much lower level.

BTW,
I do not recommend regulators for the output stages of Power Amplifiers, unless you know HOW to DESIGN the combination of PowerAmp and Regulator.
 
Last edited:
#6
Agreed with infinia and AndrewT.

The LM317 is a floating regulator, but the voltage differential between input and output must be 40 V or less. At startup Vin = 58 V and Vout is still 0 V, this is too much.
You will need additional circuitry to limit the voltage differential during startup.
But it is possible. I know of a design using a few transistors, resistors, diodes and of course a LM317 to regulate 125 V down to 110 V.

Also, 4 A is too much for a single LM317. But here too, additional components may be used to make higher currents possible.
 
#8
classic 'I gotta used transformer' looking for an amp thread.
buy a 'DX amp' kit or find a smaller transformer.
IF the secondary was with-in range of 10-20% a so called 'bucking type transformer' can be used, but that is working on primary voltages not recommend for noobs.
 
#9
jitter said:
Agreed with infinia and AndrewT.

The LM317 is a floating regulator, but the voltage differential between input and output must be 40 V or less. At startup Vin = 58 V and Vout is still 0 V, this is too much.
You will need additional circuitry to limit the voltage differential during startup.
But it is possible. I know of a design using a few transistors, resistors, diodes and of course a LM317 to regulate 125 V down to 110 V.

Also, 4 A is too much for a single LM317. But here too, additional components may be used to make higher currents possible.
Click to expand...

I was not going to use the LM317 by itself, I was planning to source the required current via a transistor, MJE3055T (10A max collector current) and diode 1N5407 (3A max) as in the attached figure. Will that work?

Thanks for bring the startup issue, I did not think about that.
 

Attachments

  • LM317_high_current_circuit.jpg
    LM317_high_current_circuit.jpg
    19.5 KB · Views: 360
#15
I too would like to know the function of that diode...

I'm no designer, but from googling results it seems to me that there are different ways of going about this.
Personally, I like the circuit on page 16 of TI's datasheet on the LM317. Sure, the transistors are golden oldies, but no doubt modern equivalents are available.

I'm going to try to explain it mainly to see if I understand it myself, so please correct me if I get it wrong.
When very little current is flowing into the LM317, the two transistors won't conduct any current. As the current drawn by the load increases, the voltage across BE (caused by voltage across 22 R resistor) on the 2N2905 turns it on and as a result the same thing happens to the TIP73. The higher the current, the more is taken on by the TIP73. The LM317 still controls the output voltage, so the TIP73 cannot cause the output voltage to rise to the input voltage (unless there were a short).

I'm not sure about if and how the current is limited in this circuit. If the transformer is beefy enough, perhaps the TIP73 would be blown and then the lonely LM317 will limit the current internally. Better use a fuse on the seconday windings if there wasn't already.

But as infinia repeated, you do need to adress the startup voltage first.
Perhaps a beefy zener in series with a 10 k resistor across the input and output of the circuit? Any comments on this?
 
#16
the post14 pic is closer to right.
I think the ONLY reason for the diode is to lower the minimum voltage of the adjustable regulator from 1.25Vdc to 0.5Vdc

The regulator tries to maintain voltage @ pin2.
adding the diode (Vf) and the transistors (Vbe) ruins the regulation.
The output is Vreg - (Vf+Vbe), both Vf and Vbe vary with current.

Use the circuits given by the manufacturers, not this rubbish you are trawling off the web.

How do you stop the three 3055 blowing up?
 
#17
jitter said:
I too would like to know the function of that diode...

I'm no designer, but from googling results it seems to me that there are different ways of going about this.
Personally, I like the circuit on page 16 of TI's datasheet on the LM317. Sure, the transistors are golden oldies, but no doubt modern equivalents are available.

I'm going to try to explain it mainly to see if I understand it myself, so please correct me if I get it wrong.
When very little current is flowing into the LM317, the two transistors won't conduct any current. As the current drawn by the load increases, the voltage across BE (caused by voltage across 22 R resistor) on the 2N2905 turns it on and as a result the same thing happens to the TIP73. The higher the current, the more is taken on by the TIP73. The LM317 still controls the output voltage, so the TIP73 cannot cause the output voltage to rise to the input voltage (unless there were a short).

I'm not sure about if and how the current is limited in this circuit. If the transformer is beefy enough, perhaps the TIP73 would be blown and then the lonely LM317 will limit the current internally. Better use a fuse on the seconday windings if there wasn't already.

But as infinia repeated, you do need to adress the startup voltage first.
Perhaps a beefy zener in series with a 10 k resistor across the input and output of the circuit? Any comments on this?
Click to expand...

Currently I am using this circuit to output about 30V each to clean up the output from a dc dc boost converter which is driven buy a couple of 12V/5A walwarts, to drive 2 LM3886s. I use only one transistor (MJE3055 / 10A) and a diode (3A). The output feeds into a LC filter with cutoff of 49Hz. I don't have any startup problems with this voltage. When I get time I will try the same circuit with 58V input and 48V output for about 10A, and we will see if something blows up 🙂
 
#20
any path will take you there, uh where are we going again, he said it didn't matter.
keep taking a left turn will also take you nearer to your starting point

anyway enjoy your experiments Im sure you will learn a bunch
 
Last edited:
Status
Not open for further replies.
Top Bottom