I absolutely love your approach to video making. You were guiding us through the process of designing a linear power supply step by step. Keep up the good work mate 😊
Nice video, loved your use of the LM358. I recently used it to make a sound amplifier from 0-5V with a 2.5V offset and it sounded relatively clean (depending on the speaker frequency range as well) its not the best for sound quality but it was alright as long as signal doesn't clip
@@keisorestis6487thats right, an Op amp cant provide much current which is why you need to amplify the current. Transistors will be your goto for amplifying current
Op amps have a higher Gain than transistors, imagine you want to amplify a veeeeery small signal, in the mV scale. Ideally an op amp has infinite gain, but in reality its gain is a bout 100.000 max. its more than enough to transform an mV scale signal to V scale, an op amp is useful for these applications, voltage gain, of course there are other applications but our topic is signal amplification. On the other hand, the gain of a transistor for voltage gain is a lengthy mathmatical process if you want all the details, i can't tell you much about it, other then the fact that they are useful for amplifying the current of our voltage boosted signal from the op amp. the current gain is easier to calculate.
I don't know what it"s like in the Americas, but in Western Europe, Poland and Czechoslovakia, the LM723 integrated ciecuit was commonly used in linear power supplies.
Bravo on all your hard work. Bravo for sharing it with us. I offer the following questions, without intending any criticism... You built a linear supply' which has the significant advantage of being a non-switch-mode low noise source. Then, you used a charge pump within the circuit. A charge pump uses switch-mode operations, as you know. So, did you add back the noise that you were trying to keep out? I am still thinking about that. You used robust components to build the fixed negative supply, to enable true 0-12V operations. Might it be useful to make the fixed negative supply to be variable, as well? Then, you would have a -12V to +12V supply with the ability to trim it to 0V. I have not thought this completely through, but I think that you have already incorporated the main components. Maybe it is an idea for you for a future build? Two small things to help our understanding. Please post a complete schematic at the front of all your videos. This way, we can refer to it, as we follow along in the lecture. Yes, I know that you posted it on your website, but that is not so convenient, when watching the video. As you are reading your text from a script, please remember to speak slowly and use pauses. We need time to understand what you said and to incorporate your present points with your former points. I offer these things in the spirit of Community. I hope you take them that way. You did great.
You are right about those improvements. Perhaps I could use a center-tapped transformer along with the LM337 to get those +/-12 volts. And thanks a lot for your suggestions, they are very helpful as feedback. Thanks for watching. :)
From my testing that I in my incompetence have done, (using a scope) I have found no noise created by the charge pump on the LM 337's output, that or it is just outside the sensitivity range of my scope. The regulator deals internally with this and cancels it out, although, if you are unsure about this you can always add a set of ceramics to kill any ac that might want to make it through the regulator, another thing that I have found that works well to kill any oscillations on the output is to add in a small choke I have done this myself and find no noise. All around a well designed circuit, I am just adapting it for a bigger scale unit.
Isn’t the lm337 identical to the lm317 except that it is a negative voltage regulator rather than a positive one? Therefore I wouldn’t expect any noise from that that you wouldn’t already be getting from the lm317. The specs for RMS noise are the same for both ICs.
@@robertthompson5908 Both the LM317 and LM337 are not magical, there are different types of noise that they will be good and bad at removing. Slow rectified mains ripple they should be quite good with. Higher frequency maybe not so much. But it all depends on the level. With enough capacitance the LM337 might be able to diminish the charge pump to negligible levels.
I feel like you overcomplicated the current limiter - in my power supply I used an LM324N and a MOSFET for the current regulation, and it limits well without the shunt voltage amplifier with 0.1 Ohm shunt, but then again I also didn't use an LM317 for the voltage regulation, but an LD1084V without the minimum voltage compensation, as I don't work with circuits that reqire such low voltages. Anyway, great video, keep up the good work!
Love this power suplly video, one of the best and I have watched many! Is it possible to connect two of these supplies in series, to make +12v 0v -12v? As many audio circuits use this
You could use a ~45 degree normally open bimetal switch to turn on the fans. I didn't notice you mention any earth grounding plan during the construction of the video, but I hope you wired earth ground to the case and the transformer.
All of the grounds on the schematic are connected. Mains power is isolated from the circuit through the input transformer. Don't connect mains power to anything other than the transformer.
On the primary side of the coil, I connected the two wires that come from AC mains voltage. On the secondary, the schematic will show you the connections. Every time that you see a ground symbol, it is electrically connected with 0 resistance to all other ground symbols. If you want to know about the third AC "ground" pin, then you can connect it to the metal shell of your case. It is not required to run the circuit and is only in place for safety reasons. This third pin is not included in the schematic and is not connected to the ground symbols.
Hey! Nice video as always. Btw, why not put the lm317's voltage divider after the current sense resistor? That basically removes the effect of series resistance from the circuit and improve the load regulation. It will obviously sense extra current of the voltage divider but thats small. Thus current regulation only slightly worse but voltage regulation much better specially if designed for high currents.
That's an interesting improvement. I'm not 100% sure how it would perform, but it definitely could work. The next time I make a linear supply I'll keep your suggestion in mind! :)
@@SineLab I realized that would not work directly. For my project, I will sense the current at LM317's Input. By the way, can you please give me all the references of that intelligent current limiting mechanism? I need some guide to decide component values.
@@MohammedAyazQuadri Sure, R2 is the current shunt resistor, and you should pick a small value to minimize the voltage drop. U4B and the surrounding 4 resistors form the differential amplifier. It takes the voltage drop across the shunt and then multiplies it by 10 (100k / 10k = 10). I would adjust R3 and R4 as necessary to get your desired gain. The gain of 10 and the 100m shunt resistor together will give us an output voltage of 1:1. That means that 1A of current gives us 1V out on the differential amplifier. U4A is simply a comparator. The inverting input is connected to RV1. This is a potentiometer that acts as a voltage divider. The max voltage of the divider is 1.2 volts, generated by the zener diode. If the current sensing circuit is of a higher voltage than the limit adjustment, then it turns on the comparator. This comparator turns on an NPN, which will pull the adjust pin down to -1.25 volts, therefore giving us 0V on the output in the event of an over current. I hope that this helps :)
The design is fine and all but IMO too many ICs, especially for a beginner, i built mine using only LM358 and damn is it unstable but you basically cannot kill it xD
@@SineLab It's not very complicated, first opamp in lm358 regulates the voltage using a darlington pair (simple linear reg) and then a single npm transistor pulls the current from the darlington pair when the current through the shunt goes above the threshold Biggest mistake i think was that i used zeners as references but they are not stable at all, with some caps and stuff this could probably be kinda stable but you'd need a oscilliscope to check and i do not have one
I absolutely love your approach to video making. You were guiding us through the process of designing a linear power supply step by step. Keep up the good work mate 😊
Thanks! :)
Nice video, loved your use of the LM358. I recently used it to make a sound amplifier from 0-5V with a 2.5V offset and it sounded relatively clean (depending on the speaker frequency range as well) its not the best for sound quality but it was alright as long as signal doesn't clip
It sure is a handy little ic. :)
Is it possible to drive speakers with the 358? I thought the max out current was too low for that application
@@keisorestis6487thats right, an Op amp cant provide much current which is why you need to amplify the current. Transistors will be your goto for amplifying current
@@Inti72600 Since you can amplify your signal with a transistor what's the upside of going through an Op amp first?
Op amps have a higher Gain than transistors, imagine you want to amplify a veeeeery small signal, in the mV scale. Ideally an op amp has infinite gain, but in reality its gain is a bout 100.000 max. its more than enough to transform an mV scale signal to V scale, an op amp is useful for these applications, voltage gain, of course there are other applications but our topic is signal amplification. On the other hand, the gain of a transistor for voltage gain is a lengthy mathmatical process if you want all the details, i can't tell you much about it, other then the fact that they are useful for amplifying the current of our voltage boosted signal from the op amp. the current gain is easier to calculate.
I don't know what it"s like in the Americas, but in Western Europe, Poland and Czechoslovakia, the LM723 integrated ciecuit was commonly used in linear power supplies.
Bravo on all your hard work. Bravo for sharing it with us. I offer the following questions, without intending any criticism... You built a linear supply' which has the significant advantage of being a non-switch-mode low noise source. Then, you used a charge pump within the circuit. A charge pump uses switch-mode operations, as you know. So, did you add back the noise that you were trying to keep out? I am still thinking about that. You used robust components to build the fixed negative supply, to enable true 0-12V operations. Might it be useful to make the fixed negative supply to be variable, as well? Then, you would have a -12V to +12V supply with the ability to trim it to 0V. I have not thought this completely through, but I think that you have already incorporated the main components. Maybe it is an idea for you for a future build? Two small things to help our understanding. Please post a complete schematic at the front of all your videos. This way, we can refer to it, as we follow along in the lecture. Yes, I know that you posted it on your website, but that is not so convenient, when watching the video. As you are reading your text from a script, please remember to speak slowly and use pauses. We need time to understand what you said and to incorporate your present points with your former points. I offer these things in the spirit of Community. I hope you take them that way. You did great.
You are right about those improvements. Perhaps I could use a center-tapped transformer along with the LM337 to get those +/-12 volts. And thanks a lot for your suggestions, they are very helpful as feedback. Thanks for watching. :)
From my testing that I in my incompetence have done, (using a scope) I have found no noise created by the charge pump on the LM 337's output, that or it is just outside the sensitivity range of my scope. The regulator deals internally with this and cancels it out, although, if you are unsure about this you can always add a set of ceramics to kill any ac that might want to make it through the regulator, another thing that I have found that works well to kill any oscillations on the output is to add in a small choke I have done this myself and find no noise. All around a well designed circuit, I am just adapting it for a bigger scale unit.
Isn’t the lm337 identical to the lm317 except that it is a negative voltage regulator rather than a positive one? Therefore I wouldn’t expect any noise from that that you wouldn’t already be getting from the lm317. The specs for RMS noise are the same for both ICs.
@@robertthompson5908 Both the LM317 and LM337 are not magical, there are different types of noise that they will be good and bad at removing. Slow rectified mains ripple they should be quite good with. Higher frequency maybe not so much. But it all depends on the level. With enough capacitance the LM337 might be able to diminish the charge pump to negligible levels.
I feel like you overcomplicated the current limiter - in my power supply I used an LM324N and a MOSFET for the current regulation, and it limits well without the shunt voltage amplifier with 0.1 Ohm shunt, but then again I also didn't use an LM317 for the voltage regulation, but an LD1084V without the minimum voltage compensation, as I don't work with circuits that reqire such low voltages. Anyway, great video, keep up the good work!
Love this power suplly video, one of the best and I have watched many! Is it possible to connect two of these supplies in series, to make +12v 0v -12v? As many audio circuits use this
I haven't tried it myself, but it should work since the transformer inside isolates it from mains power.
Awesome!
Thanks!
You could use a ~45 degree normally open bimetal switch to turn on the fans. I didn't notice you mention any earth grounding plan during the construction of the video, but I hope you wired earth ground to the case and the transformer.
An excellent video!
Glad you liked it!
Nice video! And did you add the soft start section in your final schematic?
Yes I did. It's made up of Q3, R7, and C1.
Is the ground on the AC side and the DC side same or are they isolated?
All of the grounds on the schematic are connected. Mains power is isolated from the circuit through the input transformer. Don't connect mains power to anything other than the transformer.
So where is the ground from AC connected to. Do you have any idea on resistance reading between AC ground and DC ground?
On the primary side of the coil, I connected the two wires that come from AC mains voltage. On the secondary, the schematic will show you the connections. Every time that you see a ground symbol, it is electrically connected with 0 resistance to all other ground symbols.
If you want to know about the third AC "ground" pin, then you can connect it to the metal shell of your case. It is not required to run the circuit and is only in place for safety reasons. This third pin is not included in the schematic and is not connected to the ground symbols.
Hey! Nice video as always. Btw, why not put the lm317's voltage divider after the current sense resistor? That basically removes the effect of series resistance from the circuit and improve the load regulation. It will obviously sense extra current of the voltage divider but thats small. Thus current regulation only slightly worse but voltage regulation much better specially if designed for high currents.
That's an interesting improvement. I'm not 100% sure how it would perform, but it definitely could work. The next time I make a linear supply I'll keep your suggestion in mind! :)
@@SineLab I realized that would not work directly. For my project, I will sense the current at LM317's Input.
By the way, can you please give me all the references of that intelligent current limiting mechanism? I need some guide to decide component values.
@@MohammedAyazQuadri Sure, R2 is the current shunt resistor, and you should pick a small value to minimize the voltage drop.
U4B and the surrounding 4 resistors form the differential amplifier. It takes the voltage drop across the shunt and then multiplies it by 10 (100k / 10k = 10). I would adjust R3 and R4 as necessary to get your desired gain.
The gain of 10 and the 100m shunt resistor together will give us an output voltage of 1:1. That means that 1A of current gives us 1V out on the differential amplifier.
U4A is simply a comparator. The inverting input is connected to RV1. This is a potentiometer that acts as a voltage divider. The max voltage of the divider is 1.2 volts, generated by the zener diode.
If the current sensing circuit is of a higher voltage than the limit adjustment, then it turns on the comparator. This comparator turns on an NPN, which will pull the adjust pin down to -1.25 volts, therefore giving us 0V on the output in the event of an over current.
I hope that this helps :)
Re heat sinks. Drill ànd tap for 220...
The design is fine and all but IMO too many ICs, especially for a beginner, i built mine using only LM358 and damn is it unstable but you basically cannot kill it xD
It's probably not the most beginner friendly unfortunately. I'd like to see that LM358 design too haha
@@SineLab It's not very complicated, first opamp in lm358 regulates the voltage using a darlington pair (simple linear reg) and then a single npm transistor pulls the current from the darlington pair when the current through the shunt goes above the threshold
Biggest mistake i think was that i used zeners as references but they are not stable at all, with some caps and stuff this could probably be kinda stable but you'd need a oscilliscope to check and i do not have one
@@alexstone691 Sounds cool :). Maybe you could make it stable if you ever end up getting an oscilloscope
@@SineLab maybe it would be a good video to do, i certainly would watch it :)
@@alexstone691 It's a good video idea!
Nice job but speak way to fast for me to keep up lol. You give my brain no time to even think about what you’re doing
@@sirmongoose I wish I could
I didn't have that problem, but you can always drop to .75x speed playback! I love that feature
@@evanbarnes9984 as an amateur I try to process everything to understand completely. Yes I have done that and it helps a lot. Thanks