LM317 voltage regulator keeps dying.

@zander , LOL, I did not mention that I have also read that blog at least 30 times already, the video is easier for me to see with my old eyes, I watch it on a big screen and pause it when I want. i.e. to remember the capacitor value for example . but did not know about the new feature thank you  , I'll download the PDF file of this project and save it for future reference.

Hi @ddistrbd1,

  The application of diodes to protect the voltage regulator is probably a bit of an insurance policy approach ... many commercial circuits (whole units such as a computer or audio amplifier, not a power supply) manage without them, but they are in a unit where all of the components are determined by the manufacturer of the unit, so there are less 'unknowns'.  With a power supply unit, the components it connects to, etc. are continually changing, so I recommend following the regulator manufacturer's advice as closely as possible.

Yes, obviously if you want three different voltages at the same time, you are going to need three regulators, etc.  It should be fairly straightforward, but you could draw out your proposed circuit and post it on the forum in case somebody spots an issue. Photo of a clear pen and paper sketch should be enough, if you don't want to use CAD.

Remember, the total current demand (at any one time) from all of your regulators must not exceed the current rating of your transformer and rectifier.

A fuse in series with one of the wires from the transformer to the rectifier would provide at least some protection, though fuses are not 'precision' devices .. and even the way they are specified differs markedly on opposite sides of the Atlantic, so treat it like a seat belt in the car ... it might save you in an accident, but better to drive carefully and avoid the accident in the first place.

Best wishes, Dave

Posted by: @davee

↑

Hi @ddistrbd1,

  

Yes, obviously if you want three different voltages at the same time, you are going to need three regulators, etc.  It should be fairly straightforward, but you could draw out your proposed circuit and post it on the forum in case somebody spots an issue. Photo of a clear pen and paper sketch should be enough, if you don't want to use CAD.

Remember, the total current demand (at any one time) from all of your regulators must not exceed the current rating of your transformer and rectifier.

Informative post Dave, I appreciate you / Ron's thoughts and comments , and honestly couldn't have asked for better responses . anyhow I thank you both sincerely.

I am exploring my options and thinking out loud here, I do admit of having (less than) little knowledge of this field which could make me as they say " dangerous" but I'm old enough to know I should ask every step of the way , that's why when I introduced myself I said I would be asking a lot of silly questions.

The two options I'm exploring at this time are:

A) use a 7824 and 7912 regulators instead of an adjustable one ,  route each to a tab on the rotary switch and it's own Pos. & Neg. outlets. 

B) instead of having one positive and one neg. from the filter capacitor, have 3  out, (each with an inline fuse) , then connect one to the ld1085 to make one  adjustable output, each one of the remaining two will be connected to a step down adjustable buck module, and adjust to 9V and 12V. 

I have not done enough research on how far I could push these fixed regulators (7824,7912) and not even sure it is wise to mix buck modules with this linear power supply , my only hope is not to lose your (and everyone else's)interest in this thread.

.

Posted by: @davee

↑

  The application of diodes to protect the voltage regulator is probably a bit of an insurance policy approach

Best wishes, Dave

 

I have a few 1N4007 diodes, but I might not use one since it was not in Bill's original plan and as you said it is not a "must" , to use one, just for peace of mind.

That's all I have at this time , just wanted to reply to your last post before life gets in the way again . I will let you know may even post a hand drawn circuit of what I am planning in my head but please feel free to stop me any where before I get carried away with all the unwise possibilities.

Thank you both again and also Bill for making this site free and available to us .

Regards.

Ken .

@davee @ddistrbd1 Dave, can you explain why the difference in the two diagrams of Bill's, the first being the single variable output and the second being the 3 fixed plus 1 variable with respect to the wiring of the variable pot? In the single output diagram, the pot is wired with one end of the pot as well as the wiper to G with the other end to the voltage source, while in the multi-output the pot has the more simple and obvious one end to G and the wiper to the voltage source with the other end not connected. I know I am old and it's been a long time since I studied this stuff plus not anywhere as deep as Dave but isn't the first case with one end connected to G not needed, we get a variable resistor with just 2 connections, any end plus a wiper. If I am wrong please straighten me out, I will even accept one of your famous sermons 😀 

@ddistrbd1 @davee Ok, Ken you have totally lost me. I don't understand why you would use a buck converter when an LDO VR is what the project is about. As I recall, you want to eliminate the rotary switch and instead have 4 sets of output terminals. 4 outputs mean 4 LDO VR's. Y.ou may want to use the rotary switch to dial the meter into one of the 4 outputs unless you plan on using 4 meters

Posted by: @zander

↑

@ddistrbd1 @davee Ok, Ken you have totally lost me. I don't understand why you would use a buck converter when an LDO VR is what the project is about. As I recall, you want to eliminate the rotary switch and instead have 4 sets of output terminals. 4 outputs mean 4 LDO VR's. Y.ou may want to use the rotary switch to dial the meter into one of the 4 outputs unless you plan on using 4 meters

I regret causing the mixup, it's due to my lack of using the correct terminology.

I can see where I caused the confusion, the rotary potentiometer switch remains part of the design since it allows me to have an adjustable output, the selector switch also stays on to choose both (A) adjustable output and (B) to select a fixed output which is going to be connected to a USB outlet.

I will from now on call the rotary selector switch as "selector switch", in my first post I did say I would like to have each click of the selector switch to be connected to it's own output, not like the way Bill's design is which is: all the 3  trim pots as well as the adjustable rotary potentiometer go to a single pos/neg output.

The reason for exploring the buck module idea is, as Dave mentioned,  in order to achieve 3 separate outputs , I'll need 3 VRs. so my question was/is, why not use  buck converters  instead of the other two voltage regulators? so the end result will be one LD1085 and two buck converters (or maybe just one ) ,it was just a question whether mixing a linear power supply with buck modules was a possibility.

So  if you could ,picture this diagram in your mind: a 24VAC 3Amp transformer, then abridge rectifier, then a filter capacitor, now my question starts from this point:

 Can I have at least two positive and two negative wires to come out of the filter capacitor, one (POS. & NEG.) set goes to LD1085 VR, the other set goes to a buck converter which will connect to a USB outlet, I need to find a way to connect a selector switch (some where in the diagram, maybe just passed the filter capacitor) to  switch to either (A) the adjustable output ,and second position of the selector switch would connect to a (type C) USB outlet.

I also  mentioned in one of my posts that a store bought USB charger can be used instead of  a usb outlet , the advantage is it has it's own power supply but still needs to be connected to a 12V to 24V source.

This is really getting exciting to get all you knowledgeable members read /comment  /get involved, thank you for all your help.

@ddistrbd1 Post a diagram

Hi Ken @ddistrbd1,

  Sorry, but I agree with Ron, (@zander), I think you need to draw out a diagram as the description is getting too confusing - not your fault - just a case of a picture being worth a 1000 words.

And sorry, I may not be answering all of your questions for the same reason, but I'll make a start with some comments.

Of course, you may wish to build a power supply which extends upon the original article.

It is reasonable to use more than 1 voltage regulator if you want more than 1 voltage at a time.

Similarly, it is perfectly reasonable to use a mix of variable and fixed voltage regulators, and maybe one or more buck regulators, depending on what you want your power supply to do.

Each output would have its own positive output terminal, and may conveniently have its own negative terminal as well, although all of the negative terminals will normally need to be connected together inside the power supply unit. This means there can only be one current meter if it is measuring the negative (return) line, as per Bill's design.

Each output could have its voltmeter. 

If each output option has its own regulator and output pin, I don't see the need for a rotary switch, but perhaps a diagram will explain.

A potentiometer can indeed be used as a variable resistor by connecting between the slider and one of the track end connections. However, it is often considered good practice to connect the slider to the other 'free' end of the track, as Bill shows in one of his diagrams:

The reason is the wiper to track connection is the 'most vulnerable' from a fault viewpoint, especially whilst it is being operated .. By adding the track connection, if the wiper momentarily loses contact, the resistance becomes equal to the maximum resistance, and hence in 'normal' operating range, not an open circuit. Whilst the wiper is making contact with the track, the extra connection makes no difference.

Whilst I commented the diodes are a bit of insurance policy, when making a power supply that may be connected to a range of loads etc., that is just the sort of 'high risk' usage that they are most likely to save a failure, so I recommend you fit all of those recommended in the data sheets. Of course, many people will say they have not bothered in decades and never seen a problem... take your pick!

Best wishes, Dave

@davee FYI  @ddistrbd1 Thanks, Dave, I totally missed the trick with the meter in the common return leg. I just ordered 4 of them, but I know there is another hiding here, so I will use all 5. V and A for each output, and a 5th meter measuring total A as well in case I am using 2 or more outputs. I am adding an ON/OFF switch and LED to each output as well just in case that proves needed. I have built or have parts for several other supplies, but I think this will become my main supply. Now I will need to check the transformer to see if it will supply enough current.

I suspected that was the case with the pots, but was not 100% positive as my education was over 60 years ago and was more practical than theory. Thanks for the expert explanation, I get it.

NOTE/Query If the meters are for each output, do I place them in the positive lead and therefore the red wire goes to the positive output jack?

Hi Ron @zander,

 PS Ron, I am just wondering if I have just written an explanation of what you also realised ... if so apologies .... hopefully somebody will find it helpful. Take care my friend, Dave

NOTE/Query If the meters are for each output, do I place them in the positive lead and therefore the red wire goes to the positive output jack?

This is where having multiple outputs gets tricky. Voltage measurement is fine, (though draw out a sketch first as it can be a little confusing), as it only needs the 0V-Ground line and the positive voltage out. (Plus power for the meter itself, which can be taken from the unregulated 24V line, providing the meter module is chosen appropriately.)

The troublesome bit is current measurement, as the cheap modules Bill uses measures the return current in the 0V line. With just 1 output, no big problem ... but more than 1 is not so easy.

I hope Bill won't mind if I put his diagrams from https://dronebotworkshop.com/linear-dc-power/

side by side to show the whole story with just one output.

The NEG output terminal of the power supply, connects through a low value resistor inside the meter, through to the "Negative Out" from the rectifier/smoothing capacitor.

Say the value of this sense resistor is 0.1 Ohms, so that when 1A flows through it, it will have a 0.1V drop, making the potential at the output "NEG" terminal, 0.1V with respect to the "Negative Out" point.

Now for a load that is powered by the POS and NEG terminals output, that does not connect to any other electrical circuits, this is 0.1V drop is usually of no significance.

But now consider you 'upgrade' this circuit to still have just one transformer+rectifier+smoothing capacitor, but two voltage regulator + voltage & current meters.

Lets imagine one voltage regulator output is set to 5V and the other set to 12V, (to keep the description simple).

Now imagine our 5V output is supplying 1A to its load, and the 12V output is supplying 0.1A to its load.

Thus the NEG output of the 5V output will be 0.1V positive of "Negative Out" point but the NEG output of the 12V output will only be 0.01V positive of the same "Negative Out" point, because the current flowing through the 12V output current sense resistor is only 0.1A.

If the two loads are electrically isolated from each other, this is still fine, but if they were connected in some way, their private "0V" lines would be at different potentials ... not a good situation in general, that is likely to lead to all sorts of issues!

----------------

This is why it is more convenient to measure the current flowing in the 'live' (often called 'high') side, not on the current return (or 'low') 'ground' side, when more than one voltage supply is involved.

The downside is that it is more difficult, and often more expensive, to measure the 'high' side current, and hence not supported by the "very competitively priced" modules described in Bill's design.

---------------

I have only described the two output case, but for 3, 4, 5 or more outputs, the problem is just replicated for each extra output.

---------

I hope that is clear, if maybe not what you were hoping for.

Best wishes,

Dave

@davee I changed my mind, I will put 4 meters for the separate voltages in its own negative return, and a 5th that measures all the returns.

@davee I changed my mind, I will put 4 meters for the separate voltages in its own negative return, and a 5th that measures all the returns.

Hi Ron @zander,

   Of course you know what you want to use it for.

Just beware that if you have resistors in the return lines (from the current meters), you could get problems if you are supplying (say) two PCBs from different outputs, if you then try and make a data link with something like I2C between the two boards, as you will not be able have a common earth wire between them.

That is not to say there are not sneaky ways that might still work, but you are creating a minefield of possible rabbit holes to fall into.

Good luck and have fun. Dave

@davee There are no resistors in the return lines, the circuit is negative post to meter in, meter out to the common negative line back to the rectifier.

Hi Ron @zander,

re: There are no resistors in the return lines, the circuit is negative post to meter in, meter out to the common negative line back to the rectifier.

  Sorry, I may have misunderstood which meters you are using ... if you only have voltmeters, then I agree with you.

  However, if you use the combined voltmeter and current meter that Bill used (and I showed in the picture about 5 messages up), then the current meter includes a resistor in the return line.

Best wishes, Dave