I am working on a project that needs 100VDC @ 1A from standard 110/120 AC wall voltage.

I would like to be able to vary the voltage, not necessarily all the way to 0 but to at least 50VDC. I cannot find such a supply anywhere, aside from very expensive lab supplies, so I have resorted to building one myself.

The output does not have to be super clean as it will be running through DC-DC converters later on.

A full wave bridge rectifier and some filter caps is obvious, but how could I make such a high voltage variable? I have experimented a bit with using power transistors but I cant seem to come up with anything that works reliably. I have some TL783 regulators, which are adjustable, but their max input voltage is 125V, whereas I am getting almost 170VDC out of the rectifier unloaded. At the very least I could step down the AC first with a transformer so I get a low enough voltage out of the rectifier to run into a TL783, but ideally I want to do this without a bulky transformer.

Any thoughts or ideas would be great.
Thanks.

 

First, Transformerless power supplies are banned on this site. Second, you actually need a transformer. Buy a transformer. Any further discussion will not happen here unless you do that.

 

Any suggestions on where to get a 2 or 3 A transformer for this purpose? I was thinking a step down to 80VAC.. But Digikey's only 120-80VAC transformer is 500ma.

 

Your best bet is one of these, or one like it.
http://www.jameco.com/webapp/wcs/stores/servlet/Product_10001_10001_121276_-1

It gives what you ask for, but it must be fed by an isolation transformer to be safe. /since it's an autotransformer one output leg is common to ground.
What that means is it can/will shock you unless you power it from an isolation transformer. You would have to rectify the output to get DC

 

I'm stuck at home today so I'm gonna poke about and see if I can get you a better transformer.

Edit: The suck factor on this is significant.

 

Hello,

If you can find an ILP distributer, you will have plenty of choises.
Have a look at the attached datasheet.

Bertus

 

So..here's an idea. Buy 3 transformers at $19.84 each. This gives you the right voltage for 100 volts out, 50 volts out, and a couple of in-betweens. That will lower the heat you have to get rid of if you want to operate it like that. Mouser part number and datasheet included.

 

VDC = 1.4 x VAC
125AC = 175DC as you have seen already.. So a 70V transformer would be just about perfect.

 

Old habits...figure the 105 volt low limit for the power line and guarantee the output under that condition. (28 times 105/115 times the square root of 2) minus 1.84 volts for the rectifier loss and you have 106.6 volts. That leaves 6.6 volts for ripple voltage and the pass transistor loss.

 

You can buy isolated variacs, but they are a little pricey.

 

So I've decided on a different approach for the circuit I am building and it no longer needs the 100V supply to be adjustable. I just want a constant 100V out. Below is a simple schematic I have come up with to keep the output voltage constant regardless of load (within limits of course). The AC voltage source will be a 1:1 transformer (I just don't have it in because transformers don't simulate very well in the software I am using.)

Any obvious flaws with this design? The load regulation is what is tricky as I want it to maintain 100V out from 1A down to no load. Any suggestions on how to add to this circuit to achieve that would be great.

Thanks.

 

You could still use LM317 regulator and attach the adjust pin resistor to a 70v to 90v Zener. The DATASHEET clearly states this is possible and only limits the delta in voltage in to output, not the out to ground. I haven't figured out the power dissipation but it may make things easier.

 

Thanks for the reply.. Not sure I fully understand what you are getting at with the LM317.. I see what you mean about the datasheet stating input to output differential 3 - 40V.
Are you saying it is possible for an LM317 to output 100V?

 

I am working on a project that needs 100VDC @ 1A from standard 110/120 AC wall voltage.

I would like to be able to vary the voltage, not necessarily all the way to 0 but to at least 50VDC.

With that spec, even I couldn't build you one and I designed PS's for a living for about 25 years.

Do yourself a favor and figure out exactly what you need before you start screwing around with lethal voltages.

Also did not say if it has to be current limited.

 

You could still use LM317 regulator and attach the adjust pin resistor to a 70v to 90v Zener. The DATASHEET clearly states this is possible and only limits the delta in voltage in to output, not the out to ground. I haven't figured out the power dissipation but it may make things easier.

I know it says that on the data sheet but it's not a reliable design. The part can get hit with excessive voltage on turn on transients.

 

I will explain my project a little further to clarify.. I am building an underwater camera and lighting system that runs over a single coax cable. I need the 100V supply at the top end to get power to the other end of the coax. The high voltage is to compensate for the large voltage drop experienced over the coax cable (1000m+). I am feeding this voltage into a linear regulator and then a DC-DC Buck converter at the bottom end to end up with 20V @ 3A final output (60W). My first prototype design I was varying the current as a brightness control for the LED lights, but I plan to use PWM for this newer design.

The reason for the adjustable 100V was to account for the length of the coax cable. Depending on the length, the voltage drop will change. Too short of a cable would not drop enough voltage for the regulator at the other end and therefore the system would no longer work.

Perhaps I could put in a few selectable dummy loads for shorter cables? Worst case I will just have another supply for shorter cables, say 48V.

 

What about sending a constant current down the cable? That way, as long as the compliance voltage is high enough, you won't have to readjust things for different cables. Similar current loop applications are 4-20mA sensors and airport runway lighting.
/mike

 

Good idea, n1ist. A 3 amp driver with enough voltage available will push 3 amps through the LEDs, regardless of the cable length, without any need for humans to adjust anything, and eliminate the need for a buck converter.

 

But if the current source has to have 100V of compliance, and the load is only dropping, say, 50V, then the current source is dissipating 150 Watts, unless it's a switcher.

What kind of coax are you going to use?

 

Probably one which is not rated for 100v underwater. Probably one which comes on 100m reels, or maybe 250m reels at best. Probably one that also has to carry composite video as well as high current power on the same wire?

Sorry to sound like the voice of doom but it looks like one of those threads of "the blind assisting the ignorant attempting the impossible"...