VC Buck Converter, will it work?

Gasboss775 said:

One thing that I've thought of but haven't tried yet is using an LM337 to provide the -1.25 V reference so the LM317 can go down to zero, two diodes would be cheaper, however I do happen to have a few LM337.

Click to expand...

See post #32 GB.

spec

spec said:

system difficult to stabilize in the frequency domain

Click to expand...

Sounds like you have built power supplies before. LOL

Most of the internet schematics have temperature instability and add too much voltage gain to the loop.

My circuit is not the best at temperature. (Vref + Vb-e) But many internet examples are (3 X Vb-e)

For those why don't know what we are talking about:
When look at a unknown voltage, The circuit on the left is more of a voltage comparitor than a amplifier.
If the input is less than the Base to Emitter voltage times 3 then the transistor is off. More than = on.
We all know the Bbe varies with temperature. The voltage gain of the transistor and phase shift adds confusion to the IC.

The circuit on the right also has a Vbe in it but not 3 X Vbe.
If the two resistors have the same value.....
The input voltage-Vbe is on the bottom resistor and across the top resistor.
We are comparing Voltage-Vbe to the temp stable Vref inside the IC.
There is no voltage gain and little phase shift to confuse the IC.

ronsimpson said:

Sounds like you have built power supplies before. LOL

Most of the internet schematics have temperature instability and add too much voltage gain to the loop.

My circuit is not the best at temperature. (Vref + Vb-e) But many internet examples are (3 X Vb-e)

For those why don't know what we are talking about:
When look at a unknown voltage, The circuit on the left is more of a voltage comparitor than a amplifier.
If the input is less than the Base to Emitter voltage times 3 then the transistor is off. More than = on.
We all know the Bbe varies with temperature. The voltage gain of the transistor and phase shift adds confusion to the IC.

The circuit on the right also has a Vbe in it but not 3 X Vbe.
If the two resistors have the same value.....
The input voltage-Vbe is on the bottom resistor and across the top resistor.
We are comparing Voltage-Vbe to the temp stable Vref inside the IC.
There is no voltage gain and little phase shift to confuse the IC.

View attachment 102526

Click to expand...

Hi Ron,

I think you may have designed a few power supplies too.

It is quite true what you say about temperature effects, but for a 'constant' voltage across the pass element controller, precision is not required. Reliable operation and frequency stability are the main criteria (and simplicity), I think anyway.

spec

spec said:

See post #32 GB.

spec

Click to expand...

Hope you didn't think I was stealing your thunder!

You have taken the idea a step further in providing a means of ironing out any variations in the 1V25 references of the LM317 & LM337, I like this!

To avoid any issues with temperature or frequency instability I was thinking that the pre regulator could be just switched in or out by a relay, say at around 15 Volts output. Such that for outputs < 15 V the preregulator could drop the 32V input down to 18 volts. Above 15 volts output the 32V would be fed directly to the lm317 circuit. Given that the maximum current the PSU will provide is around 500mA, the inefficiency won't be a massive issue.

BTW the whole point of this project was to find something useful to do for one of my many plugpack type PSU's from obsolete equipment ( I have a whole box of them! )

Gasboss775 said:

Hope you didn't think I was stealing your thunder!

Click to expand...

Not at all GB. It may seem extravagant to use an LM337 to stabilize the negative supply, and so it was when terminal regulators cost around £5UK and £5UK would buy you six pints of beer, a good meal, a packet of fags and a tank of juice (gasoline) for the car (automobile) but, nowadays as the price has dropped to a tenth of that, and probably a hundredth in buying power, you can sprinkle three terminal regulators around at will.

By the way, I have been meaning to do an article about a simple, and frequency stable, dual-tracking power supply using LM317/LM337s, and a single variable resistor. Would you be interested in experimenting with that?

Gasboss775 said:

You have taken the idea a step further in providing a means of ironing out any variations in the 1V25 references of the LM317 & LM337, I like this!

Click to expand...

Yes that is the fiddly bit.

Gasboss775 said:

To avoid any issues with temperature or frequency instability I was thinking that the pre regulator could be just switched in or out by a relay, say at around 15 Volts output. Such that for outputs < 15 V the preregulator could drop the 32V input down to 18 volts. Above 15 volts output the 32V would be fed directly to the lm317 circuit. Given that the maximum current the PSU will provide is around 500mA, the inefficiency won't be a massive issue.

Click to expand...

This is the approach used on many power supplies. With a transformer/bridge rectifier/ reservoir capacitor arrangement, multiple taps on the transformer are used and, often, you can hear the relays clicking in and out as you change ranges. The other approach is to have a range selector switch that selects different taps on the transformer.

Gasboss775 said:

BTW the whole point of this project was to find something useful to do for one of my many plugpack type PSU's from obsolete equipment ( I have a whole box of them! )

Click to expand...

Or even just having fun with electronics.

spec

I tried this idea this morning. That is using a capacitive dropper ahead of the bridge rectifier & smoothing capacitor as a crude sort of pre regulator.

In the diagram shown the changeover switch would in fact be the contacts of a relay. At lower output voltages from a subsequent linear regulator ( not shown ) the capacitors would be connected in series between the AC source and the bridge rectifier. At higher output voltages the capacitors would be bypassed giving the full unregulated output voltage to feed the regulator.

At low currents the unregulated voltage with the capacitors in place would be closer to the non capacitor value but as the current is lower, power dissipation would be less of an issue.

I used 3A schottky diodes ( 1n5822 ) to keep the reverse voltage across the electrolytic capacitors to a minimum.

_____________

spec said:

Or even just having fun with electronics.

spec

Click to expand...

YES!

Here is a simulation of the capacitive dropper feeding a simple voltage regulator: