New guy to the forums needing some help making a 0-18VDC regulated variable power sup

[Mike86H]

Ok, so heres what mine looks like on the inside:
**broken link removed**
Im assuming that it has all the extra circuitry due to the digital voltmeter on it. I ran it for about 3 hours the other day and the only real problem I had was that I could hear the hum of the current through the box. Im assuming that its probably a shielding issue and the fact that the components are so cheap. But surprisingly it didnt over heat.

The other circuit I found that in my mind resembles the schematics posted on here is:
**broken link removed**
The only difference I see is a lack of resistors and the addition of the analog voltmeter.

I tend to be more of a visual learner, thats why I figure if I put these pictures up on here and yall can look between the schematic and the picture and tell me if its close, I'll be able to have a better grasp of what needs to be done.

 

[KeepItSimpleStupid]

That second pic loos like what we have been discussing. xformer, 10 turn pot, linear regulator and heatsink.

You can purchase a


Here is a nice supply based on the 723 and a pass transistor
: https://www.jameco.com/webapp/wcs/stores/servlet/Product_10001_10001_2130431_-1

This, is en essence, what we discussed building. An LM350 could be dropped in and changing the diodes sand increasing the capacitor. But, at least you would have a PCB.

You have to add transformers in both cases. The ua723 has been around for years and was probably one of the first power supplies intruduced. It wasn't complete and requires support components.

You can use just the PCB and put the transistor on the case, or just mout the PCB near the case with the proper height standoff.

Agian,

I'm going to ask:
Do you want to go 1.25/Vout or 0/Vout for the switched input? For the LM350/317 design, it rquires two to 3 components..

Two possible designs:

1 resistor A, LM334 current source, resistor B to set current source, optoisolator
2. resistor, optoisolator.

 

[Mike86H]

1.25/Vout is fine. It doesnt need to go all the way down to zero, as Im seeing more and more power supplies truly dont.


I think I may have found a very similar power supply, in production, as to what we have been discussing here. The only problem is their $250 price tag.

The tech specs are:
Features:
Single or multi-turn potentiometer
Auto ranging input voltage 85–265Vac 47–440Hz
Variable intensity of output LED
Fan cooled for optimum performance
External fuse with spare (3.0Amp)
Auto-over temperature protection
Auto-over voltage protection
Built in EMI filter
Removable heavy-duty power cord
One-year warranty

Specifications:
Input - 85–265 Vac 47–440 Hz (auto ranging)
Output - 1.2–23 Vdc @ 2.5 A continuous, 4 A peak
Weight - 1.6 lbs.
Dimensions - 6.8” x 4.7” x 2.2”
Finish - Natural Aluminum

**broken link removed**
**broken link removed**

Im betting that this is probably very close to what we have been discussing simply based on what I know of power supplies so far, and what yall have told me regarding the LM350/317 IC. Im sure they are probably using one of those for the voltage regulator. I know some of those visible parts that I can see. But I would assume that the internal parts are along the lines of what we have discussed.

My big question with that power supply is heat dissipation. Since I dont see much in the way of a heat sink, and all I really see is the fan on the back, is there a chance that they are using the aluminum box itself as a heat sink? Thats the only way I can make sense of them using such a small box with no external heat sink and only a fan, but pretty much knowing that they are in fact running some sort of voltage regulator.

 

[ronv]

The first one and the last one are switching regulators. That is why they have so many more parts. But they are more efficient (read that less heat) The second one looks like what we have been discussing - a linear regulator. But I doubt it has the same specs as we have been talking about or the heat sink would be larger. If you don't like the big heat sink we can do a switcher, but they are best built on a printed circuit board where the linear one can be built in free space.

 

[Mike86H]

The switcher being like the power supply cord that was posted on page 1 or so, correct?

 

[KeepItSimpleStupid]

The 85-265 VAC is the dead giveaway that it's a switcher. The fan should make it really reliable.

We never discussed switchers here to my knowledge. Not easy to manyally build.

 

[Mike86H]

What if I took a switching unit, such as that power cord, broke it open and took out the internals and put them into another box to make one self contained unit. Then since the fan should make it more reliable, since Im assuming that the heat from them when used to charge laptops is what causes them to die eventually, put a fan in the box to aid in cooling. Then I would still run the voltage regulator correct?

 

[KeepItSimpleStupid]

To tell you the truth, I think there are two reasons:

1. The cooling aspect. Cooling makes a huge difference.

2. The other reason is the lack of surge supression. Internal surge supression costs the company lots of dollars whne the quantity of production is high. A $1.00 part really contributes to their bottom line.

I've had one laptop switcher die and it was surge related. I had another wall wart die, but I haven't disected it, but a neighbor had a surge the same week.

Laptop suppiles are fixed voltage. With a FAIR amount of work, one can sometimes reverse engineer the supply to find out what resistor to change. Power dissipation might go up a bit a lower voltages.

 

[Mike86H]

So it would be a resistor change to fix surge suppression? There isnt another circuit that can be built in order to help with that issue? Would it even be recommended to remove the power supply from its plastic enclosure and move it into one that would allow for better cooling or should I just leave it be?

 

[KeepItSimpleStupid]

No, a resistor change would just make it variable.

Removing it from the case has one caveat that I can think of. There are no screw holes and they are not easy to add. We would have to find a card edge support that works.

A laptop supply is considerably more compliated than thie https://www.instructables.com/id/Changing-the-output-voltage-of-a-cheap-power-suppl/, but as you can see the guy successfully reverse engineered and changed the voltage.

 

[Mike86H]

Well the one that RonV posted on page 1 was good up to 24V, then he mentioned something about adding the regulator and pot to adjust the voltage down from there to where I need it. If I went that route, would I still have to change resistors in order for it to work? Cause from what I've seen the majority of the power cord power supply things have some sort of adhesive or caulking stuff on them that I would think could be a pain to get off.

 

[KeepItSimpleStupid]

The first problem is that a linear regulator requires a source of DC. That COULD be provided with a laptop brick or a transformer and capacitor.
A power brick would have to be FORCIBLY taken apart with possible case damage because we don;t ave the ability to x-ray it.

Changing a resistor to a POT or POT/Resistor could make a BRICK a variable power supply. This project is relatively difficult. Then the linear "POST regulator" would not be necessary.

Think of it as 3 these 3 choices:

1. xformer+rectifier+filter + linear regulator
2. DC power brick + linear regulator
3, A modified computer power brick could give you a variable supply. This may be beyond your skill level.

 

[Mike86H]

Ok, I looked around my house and the only thing I have that converts AC input to DC output is a 6/12V trickle charger. I've taken it apart to find the transformer, that I cant identify, the switch to swap from 6V to 12V, a couple of decent sized caps, and a BDW93C transistor(?). I dont know that I would be able to salvage much more than some caps off it (3300uF, 1000uF, & 10uF), but I guess thats a start.

 

[ronv]

Don't know why I didn't think of China right away.

You could use something like this switcher. Build the transformer, bridge and filter then add this. You would need to remove the adjustment pot and replace it with your 10 turn pot.

**broken link removed**

 

[Mike86H]

Once I switched my method of thinking from "tattoo power supply" to "LM317 based power supply", I found this.

**broken link removed**

It looks a bit detailed but said that its possible to be made from about $25 worth of components, but Im assuming that means without the transformer, case, etc. Most of the cheap Chinese power supplies are coming equipped with 317 ICs as oppose to 350 ICs, so I dont know that it would be a bad thing to use them here. Im sure with quality components it would be a huge step up from what I have now. Although if I were to up the Amps, would the 350 be easily swapped in instead of the 317? It wouldnt hurt anything correct?

 

[KeepItSimpleStupid]

The LM350 could basically be swapped. Some minor changes. Transformer and possibly some resistors in the current limit),

The design you picked out has a strange swith in the transformer It would be thrown when the voltage is below a certain value to reduce power dissipation.

It's also a wider range than what you wanted.

 

[ronv]

The one I posted from E-Bay is a switching regulator, The 317's and 350's are linear regulators. The switcher needs no additional heat sink. The 317 and 350 need a heat sink.
The last picture you posted has current adjust as well as voltage adjust. I don't think you need that. It also has some protection against shorting the output or putting a voltage on the output. None of which you need. Take that "stuff" out and you have the one from post 36 except we put in a 350 which can supply twice the current of the 317.
It seems like something about the 350 supply bothers you. What is it?

 

[Mike86H]

Theres really nothing about the 350 that bothers me. In fact I like it better than the 317 for the increase in amperage, which is just about ideal. I just figured that with the 317 and the drop in amperage it wouldnt require as much in the way of heat dissipation. But Im still unsure about the amp output. Lets say, for argument sake since I know the specs off the top of my head, I went with the RadioShack transformer thats 25.2V/2A. If I went with the 350, only 2 amps would pass through? However if I went with the 317, only 1.5(?) amps would be sent through. Since heat dissipation in wattage is multiplied off the amps, if its only going to be a .5A difference in current, but a substantial difference in heat, I guess its just easier to wrap my head around that. Although Im sure there are tons of other factors besides just heat that go into yalls suggestion. Im really just trying to process all of this information, sometimes "out loud".

I wasnt sure if that other stuff was needed or not in that last schematic, but the magazine article that went into detail describing it seemed to make it out to be a grand design that would solve all problems. I knew it looked to just be the schematic you posted Ron, but with alot of add-ons.

One concern I do have is over-complicating this using more parts than I are absolutely essential, and making trouble shooting should something go wrong that much harder. Im afraid that if I use a switching power module, plus a regulating module, plus a transformer and rectifier set up, that with the increase in components, when something goes wrong, its basically throw it out and start over. Whereas, with as few components as necessary, when something breaks it should be fairly easy to troubleshoot and fix the worn out component.

And plus with using parts that I have to order in from overseas, should that component fail and I dont have a back up, this unit would be out of commission until a replacement arrived, and shipping from China can take up to a month. That stepdown converter doesnt seem easily repairable should something in it fail. At that price I would assume that its a throw away component as a whole anyway.

 

[KeepItSimpleStupid]

You, I think, have a misconcepton, Your house could have a 200 A service and your car battery can supply 400A. but it just means it can supply that much at the same time.

So a 2A secondary means it CAN supply 2A if necessary. The large capacitor can supply a "BURST" of current. The 317 will go into current limit when more than 1.5A is drawn from it, so increasing the capacitance beyound a certain value means nothing.

You could use a 1 A xformer and a 40,000 uf capacitor and you could easily draw 3A surges for a long time at a price of dips when the power supply is first turned on.

There is also a whole bunch of stuff that one might go though to make a robust design, ie. specs lie 95 to 130 VAC input.

 

[ronv]

Ahh, I get it now.

The current depends on the load, in this case it is your machine. Either regulator would try to deliver whatever current you pen asked for. That's why we picked the 350, because the initial spec was 2 amps which is more than the 317 can do.

The power is kind of counter intuitive. Lets say the bulk supply (bridge and filter) puts out 22 volts and you set the voltage out of the regulator at 18 volts and your pen draws 2 amps. The regulator drops 4 volts(the difference between 22 and 18) times 2 amps. 8 watts. But lets say we set the voltage to 10 volts and the pen still draws 2 amps the regulator drops 10 volts times 2 amps or 20 watts. Doesn't make any difference which linear regulator you use.

The switching regulator on the other hand only turns on (not quite true, but close) when the voltage is to low so there is very little power loss.

My personal feeling is that I think you will have better luck with the 350. I have a picture of this big old heat sink with the fins on the back of your metal box. If you had a hole in the back of the box the 350 would be on the heat sink, but inside the box. On the front an on off switch, fuse, adjustment pot and maybe a power on light. I don't think you would even need a board to hook the stuff to. I think you could mount everything to either the bridge or the regulator.