Ohh sorry.. its 0-18 to use as a lab VRPS.. And always there wont be a load as it is a lab power supply.

OK. In that case, it will be a 1.25v-18v supply.

And I definitely have to use a heatsink. I know I cant otherwise.

The LM317 would go into thermal shutdown, which is hard on it.
Note that for a given current, the power dissipation in the LM317 will increase as the voltage drop across the LM317 increases.
For example: you have 20VDC on the input, 18V on the output, with a load of 1 Ampere. Power dissipation is (Vin-Vout)*current, or (20-18)*1 = 2*1 = 2Watts (simplified formula).
You then choose to lower the output to 2v, with the same current (1A)
Power dissipation is now (Vin-Vout)*current, or (20-2)*1 = 18*1 = 18 Watts.

Download National Semiconductor's datasheet for the LM117/LM317. There are recommendations and formulas to determine heatsink requirements beginning on page 9.

Ahh... By the way, can i bolt the LM317 to the metal case? will it short the chip?? Assume as if my case is ground.

The TO-220's tab is connected to VOUT as well as pin 2. If you bolted the tab to the metal case, you would short out the regulator. The only good thing that would result from this is that you would reach 0V for an output - but that's all you would get!

In a previous reply, you mention using a couple of diodes to obtain voltage lower than 1.25v. Keep in mind that the Vf (forward voltage) across diodes is not fixed; it depends upon the current flow. See the attached plot of a 1N4002 diode's Vf vs current.

Note that the plot is only up to the first 10% of that particular diode's current rating; as the current increases towards maximum, the Vf will dramatically increase.

 

Remember to pay attention to the requirements for decoupling capacitors to prevent the IC oscillating at high frequency.

 

I liked the whole way you replied buddy..

Download National Semiconductor's datasheet for the LM117/LM317. There are recommendations and formulas to determine heatsink requirements beginning on page 9.

I already have that long back... Dont worry I never compromise on heatsinks.. have learned several times

The TO-220's tab is connected to VOUT as well as pin 2. If you bolted the tab to the metal case, you would short out the regulator. The only good thing that would result from this is that you would reach 0V for an output - but that's all you would get!

Oh God... thanks for that Info, I tested with a multimeter, but gave no signs of ground as clearly it goes through a resistor in circuit.. I needed that Info

In a previous reply, you mention using a couple of diodes to obtain voltage lower than 1.25v. Keep in mind that the Vf (forward voltage) across diodes is not fixed; it depends upon the current flow. See the attached plot of a 1N4002 diode's Vf vs current.

I remember studying that long ago but just stated in case someone needed a power supply to 0v with a few compromises..
Anyway I learned a lot discussing it over here rather than jumping in and burning fingers and chips.. thanks guys..

 

re: tab connected to Vout:
Oh God... thanks for that Info, I tested with a multimeter, but gave no signs of ground as clearly it goes through a resistor in circuit.. I needed that Info

Take a look at page 2 of that same datasheet. It's a little hard to see unless you zoom in on it, but if you zoom in you'll see "<- Vout" to the right of the tab.

Although the LM117/LM317 regulators are pretty easy to use, it's little things like that which will trip you up if you're not careful.

The tab being attached to Vout is quite different from the 78xx series of fixed voltage regulators; with those, the tab is connected to the GND pin. This difference has bitten more than a few people who attempted to replace a 78xx regulator with an LM317.

 

COMPONENTS
S1-SPST 250V 1Amps
F1-250V 1Amps Transformer Power =25x2=50W, So Input=50/250=0.2A
T1-250 To 25V 2A Non-CT Transformer
D1-400V 6A x 4 Bridge
C2-2200uF 63v DC Filter Capacitor (1000uF is OK,2200uF Better for heavy Load Voltage Regulation)
And NOTHING less than 63V as rectified DC voltage reaches over 35v for 25v Transformer
R3-2.5k voltage/current limiter for LED (Optional,Recalculate for other colours)
D4-Red LED
U1-National LM317T 37v 1.5A Regulator TO-220 Plastic Package
D3-1N4002 100v 1A Protection Diode against Output Capacitor C3 (unnecessary if C3<=25uF)
D2-1N4002 100v 1A Protection Diode against Bypass Capacitor C1 due to I/O short-circuit(Not compulsory for Vout below 25v)
R1-2k Linear Variable Resistor For coarse voltage Adjustment
R4-200R Linear Variable Resistor For Fine voltage Adjustment (Vout=1.25*(1+(R1+R4)/R2)
R2-1200R 5% Resistor For current adjust
C1-10uF 63v Ripple Rejection Capacitor for higher output voltages
C3-220uF 63v Capacitor for Loop Stability and higher Output Impedance (10uF is fine for stability)

FEATURES
Output Volatge 1.25-18 Volts DC Rectified at maximum 1500mA (Possibly Reach 3.4A!!)
Current Limit,Thermal Overload, Short circuit and safe area Protection

NOTES
A Distance of less than 6" between Filter capacitor and IC ensures maximum ripple rejection
Tantalum Capacitor Recommended for Capacitor C3 with 1uF
Soldering the IC must never take more than 10 seconds and maximum 300 degrees.
Working Temperature Range -65 to +150 degrees maximum
Load Regulation Available only above 10mA Vout
Ripple Rejection >=80dB
If possible tie R2 directly to the heatsink/IC case to minimize line volatge drop.It improves load regulation
and is more than you can imagine.

CAUTION
USE A FAIRLY LARGE HEATSINK IF YOU DONT WANT THE IC TO ALWAYS BE IN THERMAL SHUTDOWN MODE!! It is Compulsory!!
The heatsink should Never be the case/cabinet

 

Howz it guys?? Post comments on it if you find a problem or a better way..

Thanks to Bill, Sgtwookie and all others who helped me..

 

As a matter of good design practice, ask yourself what would happen of the wiper lifted off one of the potentiometer tracks, even momentarily?

Can you think of a better arrrangement?

 

OK, here is a potential overcurrent indicator. I have never built it, but it is very similar to the sense circuits used in current foldback schemes. Rs (Sense Resistor) is under 0.1Ω, you'll need to parallel and test select to get the value you want.

 

As a matter of good design practice, ask yourself what would happen of the wiper lifted off one of the potentiometer tracks, even momentarily?

Can you think of a better arrrangement?

Ohh Thanks to you ... I thought abt that problem a long ago but thought that the chip would "see" it as an open circuit.. When you stated it, I tested the circuit which I hadn't even built till then... To my disgrace, it was as thought.. Full voltage with one pot leg disconnected... This cannot be allowed.. Im thinking over it.. Do you have any idea Studiot? A Diode or resisor parallel to it?? Paralleling is a bigggg trouble....

And Bill, thanks for the great Idea.. Looks like an Ammeter arrangement.. current sensing method.. What values would you reccommend for the resistors and transisitors??? And Bill (Do you mind me calling like that?) any idea on that problem above??

 

Why would you need that last diode in the circuit?? To protect from back EMF when running coiled devices????

 

Opps, I was trying to retrace yours. I also made a mistake that EBLC1388 was kind enough to point out, the resistor Rs would be 1Ω or slightly less for 1A sense.

 

This is why pots are often wired with the wiper shorting a section of track. There is always some resistance in the path.
If a full zero is unacceptable, add a small resistor (small relative to the pot) in series.

 

Why would you need that last diode in the circuit?? To protect from back EMF when running coiled devices????

D2 protects the regulator in the event of a short on the output by discharging C1.
D3 protects the regulator in the event of a short on the input by discharging C3.

 

R2 should be 120 Ohms.
As you have things shown currently, the regulator will not achieve guaranteed regulation due to less-than-specified (>=10mA) current flow via R2/R1/R4.

You have the fuse in one mains line, and the switch in the other line.
In the States, the narrower of the two flat prongs is the "hot" lead, the wider is the "neutral" lead.

The fuse must be on the "hot" side, and be the only component connected to "hot".
The switch should be connected to the other side of the fuse.

You do not show the plug's ground pin connected to anything. It should be connected to ground. If there is no ground pin on electrical plugs in your country, then you need to find out which terminal on your plug is neutral and which is hot.

If your plug can be inserted into an outlet either way, then you have a potentially dangerous situation.

 

Bill, Can I please get an easy transistor??

This is why pots are often wired with the wiper shorting a section of track. There is always some resistance in the path.
If a full zero is unacceptable, add a small resistor (small relative to the pot) in series.

I have seen that practice in audio amps but thought it would be used to linearize the resistance if the pot value is not accurate.. Will tying wiper with the other unused terminal solve the problem?? Still it will reach peak as the max pot value will be the resistance in such a condition.. Even if a small series is connected.

Quote:
Originally Posted by jj_alukkas
Why would you need that last diode in the circuit?? To protect from back EMF when running coiled devices????

D2 protects the regulator in the event of a short on the output by discharging C1.
D3 protects the regulator in the event of a short on the input by discharging C3.

I didnt mean the diode in my diagram, Its in Bill's... I didn't understand the last ones use.

The R2 value I decieded according to the datasheet specification.. They say 240 is used mostly.. And it will attain the best regulation.. will 120 be better?

I India we have 3 pins.. 2 similar round for neutral and phase and we normally wire as you specified.. I'll change it.. just made it to accomodate the space..
But normal consumer devices have no earth.. Its two pin thats standard here..

 

I have seen that practice in audio amps but thought it would be used to linearize the resistance if the pot value is not accurate.. Will tying wiper with the other unused terminal solve the problem?? Still it will reach peak as the max pot value will be the resistance in such a condition.. Even if a small series is connected.

So now you know the real reason why. Shorting the track can't make anything more (orless) linear than it already is.

It is up to you to size the pot and possible series resistor accordingly.

 

I Found a solution for it.... Wire a 6k dual pot in parallel to get effective 3k.. if one doesn't touch, the other will help.. Hows that??

 

Edited::: 240 --> 120 ...
Not Redrawn pots, shorted wiper and ends
Added switch and fuse on same line.. mistaken by plug, will redraw tomorrow..

Any corrections??

 

I didnt mean the diode in my diagram, Its in Bill's... I didn't understand the last ones use.

He ham-fingered the schematic. It's supposed to be connected to drain the cap on the ADJ pin in case the output is shorted.

See the LM117 Application Note (AN181) at National Semiconductor's site:
http://www.national.com/an/AN/AN-181.pdf

The R2 value I decieded according to the datasheet specification.. They say 240 is used mostly.. And it will attain the best regulation.. will 120 be better?

The LM117 and LM317L (100mA version; different datasheet) only require 5mA load for guaranteed regulation. The LM317T (TO-220) requires 10mA for guaranteed regulation.

I India we have 3 pins.. 2 similar round for neutral and phase and we normally wire as you specified.. I'll change it.. just made it to accomodate the space..

It is important that your schematic match how it is to be wired, and the other way around, too. Please remember that your safety is the most important thing.

But normal consumer devices have no earth.. Its two pin thats standard here..

That would be for table lamps and similar things. For this application, you really need the chassis connected to earth ground for safety.

 

How abt my idea of using a dual pot??