Hi guys, I’m trying to build a power supply based on lm317. The skematic is posted below


I made this circuit recently but the output voltage is not the value I expected. It is suppose to be controlled by the 5kPot and 2.2k parrallel network. However when I turn the pot the voltage remains the same at the output, approximately 16v. (The same occurs unload or loaded).

Below is a photo of the solder trace, All connections were made according to the schematic.

How would an electrical engineer troubleshoot a circuit like this to make it work?

 

How would an electrical engineer troubleshoot a circuit like this

He would check for a wiring error of the pot wiper connection since that's what is not working, especially since an output of about 16V is what one would get if the pot wiper was not connected.

 

He would check for a wiring error of the pot wiper connection since that's what is not working, especially since an output of about 16V is what one would get if the pot wiper was not connected.

Thank you very much for your help. I checked the pot to ensure it’s working before putting it in the circuit. I’ll check it out a bit more then.

 

Thank you very much for your help. I checked the pot to ensure it’s working before putting it in the circuit. I’ll check it out a bit more then.

Not just the pot but its connections to the rest of the circuit.

 

He would check for a wiring error of the pot wiper connection since that's what is not working, especially since an output of about 16V is what one would get if the pot wiper was not connected.

Thank you kind sir, it’s working now. I checked the pot out like you recommended... I was just wondering, how did you know the output would be 16v without the pot?

I’m studying electrical engineering at the best uni that offers it in my country but they fill us up with tons of theory that I’ve yet to make use of. That’s why this summer I decided to do I few hands on projects. I started out with the experiments volume on this site where I built and recorded most of the exercises. The power supply is my first real circuit and I’m going into my 3rd of uni. What can I do to grow faster?

 

I would ged rid of R3 and Q1, if you need more current put more Q2 transistors in parallel.

 

it’s working now. I checked the pot out like you recommended..

So what was the problem?

how did you know the output would be 16v without the pot?

If the pot wiper wasn't connected, then the ADJ resistance to ground is just the parallel resistance of the 5kΩ pot and the 2.2Ωk resistor, giving ≈1.53kΩ.
That, with R5=120Ω will give an output voltage of about 16V.

 

What can I do to grow faster?

Keep experimenting with circuits.
That's the best way to learn how the theory applies in the real world.

Also experiment with a Spice simulator, such as the free LTspice download from Analog Devices.
It has a somewhat steep learning curve but the tutorials and sample circuits will help you get started.
Such a simulator allows you to easily look at circuit operation, such as plotting the currents, voltages, and component power dissipation anywhere in the circuit, and easily change the circuit part values or design to optimize the circuit.

 

So what was the problem?

The pot originally came with 6 terminals, 2 of everything you find on a normal 3 terminal pot. Basically the bottom 3 got damaged when I was installing it somehow, so when I removed it and tested the resistance the meter returned 0L. After switching to the top 3 terminals and taking extra care when soldering, the circuit started working.

 

Keep experimenting with circuits.
That's the best way to learn how the theory applies in the real world.

Also experiment with a Spice simulator, such as the free LTspice download from Analog Devices.
It has a somewhat steep learning curve but the tutorials and sample circuits will help you get started.
Such a simulator allows you to easily look at circuit operation, such as plotting the currents, voltages, and component power dissipation anywhere in the circuit, and easily change the circuit part values or design to optimize the circuit.

Multi-sim was the only simulator I had any experience with before I switched to proteus. The reason why I switch was due to the component library. Proteus had the 4047 ic which multi-sim didn’t have. However, i’ve been hearing about LTspice a lot, is this what the pros use?

P.s thanks again for helping fix the problem earlier and all your advices, I really appreciate it.

 

I would ged rid of R3 and Q1, if you need more current put more Q2 transistors in parallel.

Thank you, my transformer is 10 amps but the circuit only utilizes 6 amps. I might make that adjustment in the future if I need more current.

 

some miscellaneous points, in no particular order:

• The output transistor will need a substantial heatsink, especially for high output current and low output voltage
• A single output transistor is not adequate for 6 A for low output voltage or short-circuit survival. Two, with emitter ballasting resistors, in parallel might be enough. Study the "safe operating area" curves in the transistor datasheet.
• The current under overload or short-circuit conditions is not well defined.
  
• The bridge rectifier will likely need a heatsink at 6 A output.
• 6 A is about the maximum DC current from a transformer rated for 10 amperes AC. This is because the ratio of RMS to average current ratio is quite high due to capacitive filtering.
• For 6 A, you will probably need more capacitance for the filter. This is a combination of ripple voltage and "ripple current" in the capacitor. Most of the major manufacturers of capacitors publish applications information that will help you understand ripple current ratings and how ripple current influences the lifetime of electrolytic capacitors.
  
• Simulation is a good way to look at things like the capacitor ripple voltage and current, RMS current in the transformer, power dissipation in the bridge rectifier, etc. These things are rather difficult to calculate directly and some are hard to measure in a physical circuit, but all are important. One thing that is necessary is an approximation of the resistance of the transformer windings. Most manufacturers don't publish this, but you can measure the resistances of the primary and secondary windings. In simulation, experiment with changing the resistance values to see the influence on RMS currents. Also experiment with changing the capacitance.
  
• Be sure to use a suitably rated fuse in the AC input to the transformer. Typically you would use a time-delay ("slow-blow") fuse so it will survive the initial high surge current when the filter capacitor is first charged.
• The LM317 requires a minimum load current to regulate accurately. You can add a permanent load resistor, but you need to consider the power dissipated in the resistor as the voltage is changed.
  
• Note what the fixed resistor in parallel with the variable resistor does with regard to linearity of the set point. A simple spreadsheet is well suited for calculating this.

Like crutschow says, keep on experimenting! Build circuits, measure things, and most important, think about what you observe and try to make sense of observations based on what you know of theory. If you've never done other things that demand manual skills, building circuits helps a lot at becoming good with using your hands, which is really important in "real life" in engineering.

I strongly recommend wearing eye protection when working on power circuits. Sometimes things go bang and bits go flying. It really spoils your concentration if you have to stop to dig a bit of a transistor out out your eye.

 

i’ve been hearing about LTspice a lot, is this what the pros use?

Not necessarily.
Several of us use it on these forums because it's probably the best free Spice simulator available, and we're too cheap to pay for the non-free ones.
It doesn't come with a large library of parts, but there are numerous libraries available online that can be added, such as ones for the CD4000 and 74HCxx series digital parts.

 

Some more comments on building things:

Be sure to keep photos of your work, if not the actual circuits, even if they were just experimental. Keep notebooks. When you graduate and go looking for employment, I think you will find potential employers who will be quite impressed. It not only shows that you have developed skills that other candidates may lack, but it demonstrates that you have an absolutely essential requirement for being a good engineer - the desire and ability to learn on your own.

 

I have a lot of these lm 317 s sitting around doing nothing ... can they be wired in parallel to control more current ? ...

I understand there's a danger of current hogging but what do you think 12 of them in parallel could safely control without heatsinks ...

I'll be using the most basic circuit .

 

I understand there's a danger of current hogging but what do you think 12 of them in parallel could safely control without heatsinks .

In the TO-220 case without heatsinks, they have a junction-to-air thermal resistance of 38°C/W, so each can dissipate about 3 watts at a 25°C ambient without exceeding the maximum 150°C junction temperature..
But at that point their cases will be running close to 140°C .

The power dissipated by each regulator is (Vin -Vout) * Iout.

 

It looks like when the current limit transistor (Q2) turns on, it turns on the output transistor harder, rather than limiting the current. This would mean that when the "current limit" is reached the circuit goes into self-destruct mode. You need to move the current limit transistor to the emitter of Q1 rather than the input of the LM317.

 

some miscellaneous points, in no particular order:

• The output transistor will need a substantial heatsink, especially for high output current and low output voltage
• A single output transistor is not adequate for 6 A for low output voltage or short-circuit survival. Two, with emitter ballasting resistors, in parallel might be enough. Study the "safe operating area" curves in the transistor datasheet.
• The current under overload or short-circuit conditions is not well defined.
  
• The bridge rectifier will likely need a heatsink at 6 A output.
• 6 A is about the maximum DC current from a transformer rated for 10 amperes AC. This is because the ratio of RMS to average current ratio is quite high due to capacitive filtering.
• For 6 A, you will probably need more capacitance for the filter. This is a combination of ripple voltage and "ripple current" in the capacitor. Most of the major manufacturers of capacitors publish applications information that will help you understand ripple current ratings and how ripple current influences the lifetime of electrolytic capacitors.
  
• Simulation is a good way to look at things like the capacitor ripple voltage and current, RMS current in the transformer, power dissipation in the bridge rectifier, etc. These things are rather difficult to calculate directly and some are hard to measure in a physical circuit, but all are important. One thing that is necessary is an approximation of the resistance of the transformer windings. Most manufacturers don't publish this, but you can measure the resistances of the primary and secondary windings. In simulation, experiment with changing the resistance values to see the influence on RMS currents. Also experiment with changing the capacitance.
  
• Be sure to use a suitably rated fuse in the AC input to the transformer. Typically you would use a time-delay ("slow-blow") fuse so it will survive the initial high surge current when the filter capacitor is first charged.
• The LM317 requires a minimum load current to regulate accurately. You can add a permanent load resistor, but you need to consider the power dissipated in the resistor as the voltage is changed.
  
• Note what the fixed resistor in parallel with the variable resistor does with regard to linearity of the set point. A simple spreadsheet is well suited for calculating this.

Like crutschow says, keep on experimenting! Build circuits, measure things, and most important, think about what you observe and try to make sense of observations based on what you know of theory. If you've never done other things that demand manual skills, building circuits helps a lot at becoming good with using your hands, which is really important in "real life" in engineering.

I strongly recommend wearing eye protection when working on power circuits. Sometimes things go bang and bits go flying. It really spoils your concentration if you have to stop to dig a bit of a transistor out out your eye.

Note what the fixed resistor in parallel with the variable resistor does with regard to linearity of the set point. A simple spreadsheet is well suited for calculating this.

Thank you for your reply, you made me aware of concepts I didn’t know that existed, like ripple current. I can rap my brain around most of the pointers except for the last one. What do you mean by that?

Another thing, I’ve been studying transistors since the start of this summer but I still cannot understand how the amplify current. I know a larger current is allowed to flow through the collector-emitter when the base is biased in the saturation region is that right? I got an assignment to build an inverter circuit last semester and my power mosfets kept overheating. Btw can both bjt’s and mosfets be use to amplify current?

The problem at time was that I didn’t know that you can’t use the mosfet as a switch and a current amplifier at the same time. I’m not even sure if what I know about transistors is true. I did all the experiments on semi conductors on this site but apart from a few basic principle of operation I know nothing...

I really want to learn more about transistors and how they can be use to amplify current. Do you know of any resources/tutorials that can help me?

 

It looks like when the current limit transistor (Q2) turns on, it turns on the output transistor harder, rather than limiting the current. This would mean that when the "current limit" is reached the circuit goes into self-destruct mode. You need to move the current limit transistor to the emitter of Q1 rather than the input of the LM317.

Q2 is not a current limit - it is how the external transistor shares the current with the LM317.

 

It looks like when the current limit transistor (Q2) turns on, it turns on the output transistor harder, rather than limiting the current. This would mean that when the "current limit" is reached the circuit goes into self-destruct mode. You need to move the current limit transistor to the emitter of Q1 rather than the input of the LM317.

The pnp-npn transistor was originally meant to maximize current not limit it. Wouldn’t a 5 amp fuse stop the cuircuit from being destroyed if too much current was being drawn?