Newbie power supply design

[Mandelbrot]

After looking for a project to work on between spring and summer classes, I was told that it would be a good idea to try and design a linear power supply. I did quite a few hours of research over the past few days and came up with a design largely but not entirely based from some I found during research. I'm not so confident in my design so far, so I was wondering if anyone could find anything wrong with it. I am trying to achieve 1.25V-30V and up to 1A. I have many other questions as well, but I guess I should get the basic design down first.

I hope I attached the schematic correctly...

Thanks,

Ryan

[mariush]

Please read the following thread, where me and others have made some observations that also apply to your design :  https://www.eevblog.com/forum/beginners/lm317-power-supply-1uestions/

I don't want to write another long post with the same things.

in addition, i don't see how you're going to get a 1k pot capable of handling 1A of current. But you have bigger problems than that.

[Mandelbrot]

I thought I read all the recent threads on this subject, somehow I missed that one.

So the main problem with my circuit is that there is too much power being dissipated in the ICs? I forgot about rms, so I can, and should, choose a smaller transformer, like a 24V output with two secondary windings to get 12V rms for smaller voltages. About the potentiometer, I was able to find quite a few on DigiKey with power ratings larger than or equal to 1.5W unless that doesn't mean what I thought it meant. I was also going to mount the LM317s on a big heat sink and keep them away from resistors and capacitors. If I use a switch to choose between one or both secondary windings, possibly automatic with a relay at a set voltage, with that solve the power problems, or is there something else I'm missing?

Thanks very much! I hope to learn a bunch from this project.

[Mandelbrot]

I fixed the transformer so that the power supply can be switched between high and low voltage modes to reduce power consumption. I also changed the resistors on the voltage selection side, but I may change it back because I am even more doubtful of those values than the ones in the previous design.

I figure the worst case scenario for power dissipation in the voltage selection IC is (24V * sqrt(2) - 1.4V - additional drops - 12V * sqrt(2)) * 1A = 15.57W - drops across other devices

It's still not great, but I'm going to use a decent heatsink.

I'm still really new to this, is there anything else that I did wrong?

Thanks for the help, I hope I will be able to pass on the knowledge later.

[Kremmen]

The regulator circuit is somehow fundamentally wrong. You just cannot have the 1k pot R1 in the main circuit. It makes no sense in series with a 1.25 ohm resistor. I am too laz... ee busy to check whether the topology with the 2 regulators can work but i suspect not. Is this supposed to be a coarse/fine adjustment setu? In that case i am not sure it will work as intended..
A couple of practical observations:
-suggest you move the fuse to be the first component in from the line input. That way the main switch is also protected by the fuse against a short to the ground.
-a dual secondary transformer could be a good idea. Then you could switch the secondaries either in series to get more voltage or in parallel to get more current.

[Mandelbrot]

Thanks for the feedback. The first regulator is supposed to set the current limit. The pot is for changing the current limit while the 1.25 ohm resistor is for limiting the max current to 1A when the pot is all the way down. I guess I should also consider that the pot isn't quite 0 ohms at the low end. I may have messed this part up; I was using Analog Power Supply Design as a reference here.

Good call on switching the fuse to be before the switch, I hadn't thought about that.

[mariush]

The Texas Instruments datasheet : http://www.ti.com/lit/ds/symlink/lm117.pdf

is full of examples of regulators with current limiting done right. Though quite a few use external power transistors to do the heavy work.

-  see pages 17-25.

[Kremmen]

Thanks for the feedback. The first regulator is supposed to set the current limit. The pot is for changing the current limit while the 1.25 ohm resistor is for limiting the max current to 1A when the pot is all the way down. I guess I should also consider that the pot isn't quite 0 ohms at the low end. I may have messed this part up; I was using Analog Power Supply Design as a reference here.
[...]

OK i found the circuit in the design appnote. It could kinda work but still, putting a 1k (250 ohm in the appnote) pot in the main circuit is IMO just a strange design. Suggest you consider implementing the current limit in a way that is more "mainstgream". I don't have much use for 317s so can't offhand say if there is a neat way to limit current with that circuit but the "canonical" way is to use a difference amp to rob drive signal from the main pass element based on max current reference and measured actual current. The principle used in Fig 46 of the datasheet mariush linked above is an example of the generic way of implementing current limit. Works with all kinds of basic circuits and is well worth understanding.

[InterestedTom]

You're current limiting will work fine, so long as the chosen potentiometer can take the (V^2)/R.  A current limiting arrangement like in figure 46 of the ti datasheet would save you the extra cost of a high power potentiometer.  Check out dave's power supply videos https://www.youtube.com/playlist?list=PLBF35875F73B5C9B5
and https://www.youtube.com/playlist?list=PLE37A0E6F75C37452 for further info