Hi Guys,

I want to built a bench power supply of max 36V and 5A. I have been researching a bit and it seems the most common way to do it is with a LM317/8 or LM338. The problem I have is that this allow for voltage to only drop to 1.25V approx. I want my power supply to be able to drop all the way to 0 (zero) volts.

Can I achieve this by simply dropping some volts using diodes/resistors? Or is there a better way to do this?

Thanks in Advance

 

Apologies, I meant LM350 not LM338

 

Hello Andre and welcome to the forum.
You can make the three terminal regulators controllable down to zero by referencing the voltage control resistors to a negative 1.23 volts instead of ground. Here is an example I found on google.

The stability of the source for the -1.23 volts will impact the stability of the output. I would use a precision reference diode. Anyway you get the idea.

 

Welcome to AAC, Andre. Experimenting with circuits and building your own equipment is an excellent way (if not the only way) of gaining knowledge and experience in electronics.

A power supply that can deliver 40V@5A is no simple task. That's 200W.

Here are some excellent examples of power supply design and construction on the late Tony van Roon's website:

http://www.sentex.ca/~mec1995/circ/ps3010/ps3010a.html

http://www.sentex.ca/~mec1995/circ/ps4002/ps4002.html

The question is, can you build a high wattage adjustable power supply at lower cost than what you can buy off the shelf? Probably not.
May I suggest that you start off with a simpe 5V@1A supply for digital experimentation and +/-12V@1A for analog circuit design and prototyping.
After that you can choose to tackle your original goal of building a hefty power supply.

Here is a simple dual 12V supply for starters. A single 5V supply follows along the same design:

http://www.sentex.ca/~mec1995/circ/741-ps.html

For more of Tony's electronics circuits, see this:
http://www.sentex.ca/~mec1995/circ/circuits.htm

 

Hello Andre and welcome to the forum.
You can make the three terminal regulators controllable down to zero by referencing the voltage control resistors to a negative 1.23 volts instead of ground. Here is an example I found on google.

The stability of the source for the -1.23 volts will impact the stability of the output. I would use a precision reference diode. Anyway you get the idea.

I had the circuit in my computer, the R value should be recalculate,
because 0<Vout <3V.

 

Thanks for the input guys.

MrChips, thanks for the advice and links. I might just take your advise on building something simpler. I'm probably a bit ambitious for my experience level.

I think I'll try a 12V 1A. That should be sufficient for 90% of the projects I have in mind.

 

Excellent idea. Glad I can be of assistance.

Here is a starting point:

http://electronics.stackexchange.com/questions/73863/cap-value-for-full-wave-rectifier-circuit

http://www.circuitstoday.com/12v-5v-combo-power-supply

http://www.circuitstoday.com/regulated-dual-power-supply-circuit

 

I still have the bench supply I made back in the 70s. It was made entirely from 78xx regulators.

 

Yes. The first supply I ever built as a teenager is a simple 5V@1A supply using a 7805 three-terminal regulator.

I still have and use this supply once in a while.

 

Mine are in my cars to act as USB chargers. Another trickle charges my boat battery over the winter when it's not being borrowed to power a project.

 

I still have the bench supply I made back in the 70s. It was made entirely from 78xx regulators.

The one I built in the early 70's used a uA723 and a 2N3055 pass transistor. Still use it all the time. It has only had to be repaired twice. Once from age and the other time from applying 170 volts to the output.

 

I too still have a high power supply I built in the 1970s using the uA723. It had fold-back current limiting added using an opamp (likely a '741). To limit power dissipation, it has a Variac to control the unregulated voltage ahead of the pass transistors. At any given output voltage, I can manually set the unreg. voltage to be only slightly higher than it has to be to keep the regulator from dropping out...

 

Is this the nostalgia thread?

I have a dual tracking +/- 15V supply I made in 1973 from discrete transistors!
Then another from about 10 years ago with all LM78 and LM79 regulators.

In my blogs is a circuit that goes to 0.0 volts.
Many ways to skin the cat.

http://forum.allaboutcircuits.com/attachments/723float-png.72328/

 

I have yet to find a regulator chip as useful as the uA723. I find it very good when I need to limit currents such as a test stand where the unit being tested may well be bad.

 

This is most definitely a nostalgia thread.

Does anyone remember the MC1466 precision regulator by Motorola. (Obsolete in 1980s) Constant voltage and constant current and zero volts out. I built, in my kitchen, a PCB for this chip and made a few bucks selling them. I wish they were still made. Great chip.

 

I would not use a 3 terminal regulator for an adjustable power supply. In my experience they may go into current limit mode even if the heat sink is large.
This because most of them have internal input output differential voltage/dissipation monitoring. I have found this to be true on the LM338, Lt1083,
and even the LM317 if used near 1A. I still have the LM723 in stock and recently built a 12V 8A linear supply with it. I have also found to use a darlington transistor as a series pass element and it takes less current to control it, or better yet use a power mosfet.

 

Here's a 3A linear regulator adjustable down to 0V output.

 

crutschow, that is a very cool regulator. Good find.

 

Another interesting feature of the LT3083 is that it has a low dropout voltage of about a half volt or less if a voltage approximately 1V higher than the IN voltage is fed to the Vcontrol input. This higher voltage can be generated from a transformer bridge supply by using a separate single Schottky rectifier-capacitor from the transformer to generate a slightly higher voltage than the bridge output going to the IN terminal. The Vcontrol input takes only 80mA maximum so you can generate a voltage close to the transformer output peak voltage with a 1A Schottky diode and a large filter capacitor.

 

Thanks a lot for all the input guys. Appreciate it.