Variable Power Supply

[SantaCruZin]

I need to create a power supply that will provide 18-36VDC from a variable 0-5VDC input. The output load will not exceed 600mA. The supply does not have to be precise or finely filtered.

Any tips?

 

[VEBill]

Things are going to get a bit tricky when the variable input power varies down towards 0 VDC.

 

[SantaCruZin]

Good point. I can just as easily make it 1-5VDC.

 

[SantaCruZin]

Perhaps I could run a DC power supply with a constant output of say, 48 or 50 volts (whatever might be common above 36 volts) as a source, and have some sort of electronic voltage divider condition the output?

 

[buzzp]

Your last post caused some confusion.
Do you want more voltage than the input voltage? If so, you need a switcher or converter.
If the supply voltage is more than the voltage you need, this is the easiest. You could use a standard 3 pin regulator (likely with a heat sink) if the input voltage is 30V or less (might go higher, dont know).

 

[SantaCruZin]

Sorry for the confusion.

Input: 1 to 5 volts, variable.
Output: 18 to 36 volts.

When I mentioned the ~48V power supply, I should have used the word "reference" instead of "source".

I'm thinking that with a constant DC source that's above the output I'm looking for, I can take the power supply's output as the input to a custom control circuit, where a 1 to 5 volt control signal will create a new output of 18 to 36 volts (or thereabouts).

 

[OperaHouse]

I'm still confused. Do you want a regulator that operates from a power source greater than 36V with a control voltage that varies from 0-5V and results in a regulated output of 18-36V?

This could be done with a simple op amp (zero and span)fed to the common pin of a three terminal regulator like a 317. Open collector like a LM339 would work best with maybe a 10V zener to keep it away from the high end limit of the device.

 

[buzzp]

I believe he has 1-5VDC and he wants to change this to be 18-36VDC and he does have a power supply available, if need be, to use in the circuit to transform the signal. In otherwords, a switcher is not required.
The adjustable regulator method should work. I have not attempted this circuit in the past (may Operahouse has) but this involves changing the normal ref voltage of the 317 from 1.25V. Not sure this is ok with the manufactured part. I do know they have resistance limitations on this part of the circuit. In any case, it is definately worth checking out.

http://www.fairchildsemi.com/ds/LM/LM317.pdf

 

[emmgon]

With an input voltage which is less than the required output voltage, the circuit might be very tricky.

It is impossible to use the usual LM317X and 78XX series because they require input voltages which is greater than the required output voltage (usually 2.5V higher). If you want to produce an output voltage from a 1-5V dc source, you may have two options:

Option 1: Design a circuit which involves the help of an pulse train, usually from a 555 chip.

Option 2: Look for an special integrated circuit (e.g. Maxim

http://www.maxim-ic.com

) where you can find chips for such applications.

--------------------------------
Emmanuel A. Gonzalez, E.C.E.
ECE Department
De La Salle University - Manila

 

[OperaHouse]

The major issue here is that you can't help people that post once and never come back. Who knows what this guy really wants.

 

[SantaCruZin]

Another major issue on forums such as this is dealing with PRESUMPTIVE PEOPLE. Who said I wasn't coming back? I work for a living, and have been extremely busy lately.

For further clarification of my requirements: I will be using a National Instruments DAQ board that has an analog output of 0-5VDC. A board I need to test is a DC/DC converter that takes an input from 19 to 36 volts. To test the thing, I need to provide both the lower and upper rated input voltages, and measure the stability of the output under 50% and 100% loads. This will be in a production test environment, with a LabVIEW GUI driving the DAQ card (again, the source of the controllable 0-5VDC).

I'll dig around the Maxim site, they do have many voltage conditioning devices to choose from.

Thanks for everyone's input, if anybody has any further suggestions, I'd enjoy reading them.

-SC

 

[analogkid2digitalman]

Why not skip the DAQ method and use a COTS programmable power supply with USB, GPIB, or serial port for communication? Still a LabView solution, and a lot quicker too, no custom circuitry.

If going the DAQ route, just a note that output of analog out is limited to about 10ma, I think. Another concern is electrical isolation- What if the UUT is defective, can it take out the DAQ card and the PC? Think production test environment.

Wheels within wheels / In a spiral array
A pattern so grand / And complex
Time after time / We lose sight of the way
Our causes can't see / Their effects.

 

[leonar40]

If you don't have a power supply that has GPIB or serial communication, then how about using a power supply into a FET, and then using the DAQ card with LabVIEW to turn the FET on and off when you want to test your circuit. The DAQ card can sink enough current to turn the gate of the FET on and off, I have done it many times.

 

[emmgon]

Hi there. I like the previous suggestion. Its a good idea to use an FET and control it with LabView.

--------------------------------
Emmanuel A. Gonzalez, E.C.E.
ECE Department
De La Salle University - Manila

 

[itsmoked]

Yeah I was a little confused there for a minute and I live in Santa Cruz!

SantaCruZin,

It sounds like you only need the low and high voltages, why not two supplies and a digital selection between them? Use foldback supplies so bad product doesn't even blow a fuse.

If you must have the variable voltage there are many PS with analog input control that will provide output current displays, fault current limiting, etc, that your LM137 type solution will not provide without extra effort.

 

[SantaCruZin]

Thank you everyone, for the great recommendations. I chose to go the route of the programmable power supply, I found one that'll control 0-40VDC at 3A, which is right up my alley. It uses serial control, an easy solution for LabVIEW. The power supply will come in handy for future tests, so the versatility was another sellling point.

Thanks again, this forum is a great resource!

 

[analogkid2digitalman]

Excellent. I try to minimize custom circuitry in production test systems if there is a COTS solution. May cost a little more, but usually faster and more reliable.

-Pete

Wheels within wheels / In a spiral array
A pattern so grand / And complex
Time after time / We lose sight of the way
Our causes can't see / Their effects.

 

[itsmoked]

Yes akid2dm,
It also prevents you from being the fireman for the test jig everytime a bad product toasts it. Good call.