Variable PSU problem......

[TW462]

I hope someone may be able to offer some insight into this, i have been playing with a circuit derived from LM317 datasheets for a variable power supply.
The constructed circuit has a digital volt meter which is powered from a separate source; with no load the supply can be swept through a range of 0v to 35.9v. However, when a load is attached the voltage drops down to virtually nothing, a small 12v bulb will only glow dimly even on maximum output (reading of 4v on the meter), an LED lights up but doesn't burn out when you increase voltage to maximum (reading of 3v on meter).
I'm confused by this, i originally drew the circuit in multisim for reference before i started construction, running a simulation yields the same results, the virtual circuit behaves exactly the same way as the real one with a pitiful output.
I have tested the components and found no faults, replaced the original lm317t with a lm317k, also replaced the linear pot. I still can't see the problem. I only wanted a small variable PSU for testing other circuits up to 24v!
I am attaching the drawing for what it's worth, it is quite a basic and common design
Thanks in advance.......... TW

 

[(*steve*)]

12V 25W!!! That's a 2A load. The LM317 is rated for 1.5A MAX.

 

[Leighcusack]

You can get the reg in a TO3 package. It can handle up tp 3 Amps. Put a heat sink on it though. As the temperature increases the power output decreases.

I have one on a heat sink designed for TO3's and have an old CPU fan cable tied to the heat sink. Barely gets warn to the touch.

the TO3 packages are expensive though.

Good luck

 

[Leighcusack]

Don't forget that it's the power the reg is dissipating that is the concern.
You,ll still struggle getting that current at 12 volts off a 30 volt ac supply.

 

[(*steve*)]

Some of the fixed regulators are available in 5A versions in TO3 packages. My brief look at the specs for the LM317 suggests it's not (however I'm willing to be proven wrong).

Oddly enough the datasheet I looked at had a lower junction to case thermal resistance for TO-220 than TO-3.

I would also be checking to make sure that my transformer was rated appropriately. The rated secondary current needs to be significantly higher than the DC current required. See the last page of this for a quick summary.

 

[Leighcusack]

Try lm350.
That's the reg I'm using.
I also had the lm317 next to it.
Sorry bout the confusion.

 

[TW462]

in response to the replies:

(*steve*): 12V 25W!!! That's a 2A load. The LM317 is rated for 1.5A MAX.

You'll have to excuse me on that one, that was simply my choice of load for the simulation; a 10w bulb was blowing during the simulation even though the voltage didn't read above 4 volts, i thought it better to use a 25w bulb to demonstrate the lack of voltage when the output was at maximum than a blown bulb that would have opened the circuit. For my purposes 1 amp is plenty of current which the lm317 should provide.

Leighcusack: You can get the reg in a TO3 package. It can handle up tp 3 Amps. Put a heat sink on it though. As the temperature increases the power output decreases.

As i mentioned in the original post, i swapped the lm317t for a lm317k; the difference being the lm317t is a to220 and the lm317k is in fact a t03. This was my first consideration as the lm317k does indeed handle higher current and i had one to try. It is mounted on a large heatsink with a cooling fan powered from the same separate power source that supplies the voltmeter.

A little more information may help here; there is an auxiliary circuit running in the same box. The auxiliary circuit is powered by its own 12v transformer; it provides power to the fan for the lm317k heatsink, it provides power to the voltmeter, and it also provides power to a buzzer/continuity checking circuit for high voltage diodes that need more power to test than the typical handheld multimeter can provide (some need more than 10volts). As i have said though, these are all powered from a separate source and should not interfere, just wanted to make clear that i'm not overloading the variable psu circuit with anything else.

I had a look at the end of the page you reffered to steve, had trouble following it as the language lowered further as the page went on. The 30v transformer i have for the circuit is a 50va model from RS, i thought that would have been plenty for this circuit, i hadn't considered swapping the transformer to see how the circuit behaves because the simulation was behaving the same way with a virtual 30v supply.

Thanks for the feedback and help so far, as unlikely as it sounds considering how basic the circuit is, i believe the fault must lie in the design for it to fail both physically and virtually in the same way.

 

[(*steve*)]

I had a look at the end of the page you reffered to steve, had trouble following it as the language lowered further as the page went on.

I didn't really notice the swearing

The 30v transformer i have for the circuit is a 50va model from RS, i thought that would have been plenty for this circuit, i hadn't considered swapping the transformer to see how the circuit behaves because the simulation was behaving the same way with a virtual 30v supply.

the practical upshot of that note is that you need a transformer rated for between 1.5 and 2 times the DC current you're after. I've always calculated it using 1.4x, but either way, 2A from your DC source is way too much.

Measure the input voltage to the regulator to make sure it's not drooping too much. I suspect that the regulator is simply falling out of regulation due to overcurrent and/or thermal shutdown.

 

[NickS]

would you mind zipping up that simulation and attaching it here so I can play with it?

 

[TW462]

Measure the input voltage to the regulator to make sure it's not drooping too much. I suspect that the regulator is simply falling out of regulation due to overcurrent and/or thermal shutdown.

The datasheet lists max voltage at 37 to 40, i measured the input at 48v
I split the 30v transformer to 2x15v secondaries, one for auxiliary and one for the regulator; this gave nearly 24v input to the regulator and the circuit works now, the lower voltage range still meets my requirements so i'm happy with it, current is high enough to blow bulbs and the voltage doesn't drop too much under load. So thanks steve, this has certainly taught me an embarrassing lesson in checking the obvious, no matter how confident i am that the root of the problem must be much more complicated than it needs to be! I'm just glad i didn't destroy the regulator running it that high.

would you mind zipping up that simulation and attaching it here so I can play with it?

View attachment vpsu2.zip
You're welcome to play around with it, I deleted the LM317k from the circuit and replaced it with a new one, must have been from a bad batch as it started behaving after that , this is not the first time i've had simulation problems that have disappeared after redrawing the circuit and starting afresh, but i still think multisim is a great little program even if you only use it to plan or print circuits

Thanks again everyone for your help