The idea is they are the raw power. The regulator comes after, and switching power, while noisy, is also very regulated.

 

The idea is they are the raw power. The regulator comes after, and switching power, while noisy, is also very regulated.

I know. There is still much more AC component in them than a linear. Check out the specs on them. Some claim a 30mV PP ripple! And that's what they claim too. Closer to the truth it could be 50mV !!!

I have just done a schematic for the supply as you requested:

 

I make some silly mistakes but I always pick up on them myself usually.

Can you see any problems?

 

Yeah this is correct. Sometimes I blow it up a bit to make sure I am not going blind.

 

These supplies are arguably unsafe for an experimental supply powering circuit boards on your desk, because they're not isolated.

Balderdash!
They most certainly are isolated.
Just because you don't understand the technology is no reason to bad-mouth it.

The power isn't as clean either. Not suitable for audio, and after all this will be a dual rail supply.

You can hear 100kHz switching? My, what good ears you have.

You should look up Class D audio amplifiers. They are essentially forward converters using feedback modulated with audio that has been PWM'ed at a high frequency. Unlike class A and class AB amplifiers, they are very efficient, and dissipate relatively little power in the amplifier itself.

Here's a link to an article over at Rod Elliott's ESP site where he built a 350W switching dual rail 30v supply for his car audio system:
http://sound.westhost.com/project89.htm
Rod is no slouch at this.

 

Getting closer to a final:

 

Balderdash!
They most certainly are isolated.
Just because you don't understand the technology is no reason to bad-mouth

I hold an 90% pass grade with DC power supplies that I studied back in 98.

Go and have words with my teacher if you disagree with what I am saying on here.

 

Yeah this is correct. Sometimes I blow it up a bit to make sure I am not going blind.

You have "Cal Trimpot" as a 200 Ohm pot. Not good. If you accidentally set the 200 Ohm pot too low, you would burn it up, as the LM317 will source up to ~2.2A short term attempting to keep Vref (OUT to ADJ) at a nominal 1.25V; ±50mV.

If you go "by the book", it should be a 120 Ohm fixed resistor; or Vref*100 (it can be lower if necessary, but the idle current subtracts from max output). Using 120 Ohms or Vref*100 guarantees a 10mA output current which is required for guaranteed regulation. If you can otherwise guarantee that there will be at least 10mA output current, you can increase the value up to perhaps 330 Ohms; after which the Iadj (50uA-80uA) error term becomes more significant.

 

You have "Cal Trimpot" as a 200 Ohm pot. Not good. If you accidentally set the 200 Ohm pot too low, you would burn it up, as the LM317 will source up to ~2.2A short term attempting to keep Vref (OUT to ADJ) at a nominal 1.25V; ±50mV.

If you go "by the book", it should be a 120 Ohm fixed resistor; or Vref*100 (it can be lower if necessary, but the idle current subtracts from max output). Using 120 Ohms or Vref*100 guarantees a 10mA output current which is required for guaranteed regulation. If you can otherwise guarantee that there will be at least 10mA output current, you can increase the value up to perhaps 330 Ohms; after which the Iadj (50uA-80uA) error term becomes more significant.

I do this so the pot can be adjusted to the scale on the artwork as a "once off" thing on the board. It is a trimpot which I don't have in my Circuit Maker 2000.

I do agree though, and there's a 47R in there now:

 

Any other things to change / improve on while I have it in my mind?

 

Just an appearance issue in the schematic; I generally put the 100nF/0.1uF cap just to the right of the R1/R2 divider to suggest it needs to be very close to the regulator on the board; and on the board, keep it close to OUT and GND, and the leads short.

Using a 1uF cap from ADJ to GND will help eliminate the overshoot that occurs during startup, and will also help a great deal to reject transients. If you use a cap here, add a diode from ADJ to OUT if Vin > 20v.

A 10uF cap on the output will "buy more time" for the regulator to deal with transients; the LM317 is somewhat slow to respond. 1uF is kind of minimal.

 

Might want to consider increasing the cap on the output to around 100uF for load transients.

 

Just an appearance issue in the schematic; I generally put the 100nF/0.1uF cap just to the right of the R1/R2 divider to suggest it needs to be very close to the regulator on the board; and on the board, keep it close to OUT and GND, and the leads short.

Using a 1uF cap from ADJ to GND will help eliminate the overshoot that occurs during startup, and will also help a great deal to reject transients. If you use a cap here, add a diode from ADJ to OUT if Vin > 20v.

A 10uF cap on the output will "buy more time" for the regulator to deal with transients; the LM317 is somewhat slow to respond. 1uF is kind of minimal.

Done.

 

Might want to consider increasing the cap on the output to around 100uF for load transients.

No way. It won't be responsive unless the load is actually something enough to discharge it quick.

 

Now take those images and start a completed project thread. It is nice to see it done correctly now and again.

Thing is, I can see it being duplicated many times over. This is a beginners site after all.

 

Now take those images and start a completed project thread. It is nice to see it done correctly now and again.

I sometimes do 100 drafts before I am happy.

 

Now take those images and start a completed project thread. It is nice to see it done correctly now and again.

The simple, but safe and sane power supply has been submitted.

 

Hello,

You just forgot to draw the ground connection at the minus connection of the 1000 uF capacitor.
It will be non connected at the moment.

Bertus

 

Thanks Mr Pussy Cat.

 

Balderdash!
You should look up Class D audio amplifiers. They are essentially forward converters using feedback modulated with audio that has been PWM'ed at a high frequency. Unlike class A and class AB amplifiers, they are very efficient, and dissipate relatively little power in the amplifier itself.

All modern Cel phones use Class D amps for audio, even for the headphone/music jack and speakerphone. They sound quite good and can be very loud, especially the alarms ringtones. Relative to the meek "beep beep" of cel phones circa 2002.