DIY Power Supply

[Joshua]

Hello Everybody,

I am very interested in making my own bench power supply. I know Dave said in one of his blogs that he built I think three(?) of his power supplies.

I've seen and investigated converting an old ATX computer power supply, but I am not really interested in that.

I would like to do this project for a couple of reasons.



I feel like this could be a good learning experience to learn more about power supplies.
I would rather it be more 'custom' than a 'pre-fabricated' design.
I would like to document my project for possibility of putting on a blog(a blog that doesn't actually exist yet though)
I enjoy making my own things, and doing this kind of work.



Unfortunately, I have no Idea where to start. Almost all searches on google for this return most results for the converted ATX type.

Does any one  have any information, schematics, links, helpful hints, etc... for going about this?


Thanks,
Joshua

[JohnS_AZ]

There are a number of questions that need to be answered first about what you actually want.

- A fixed value supply, or a variable output unit?
- Do you want/need current limiting?
- Do you want a linear supply, or a switching unit?
- What kind of max current do you want?

My sense is that most of the folks here are fond of switching supplies, and/or buying pro/semi-pro bench supplies. Personally, I've gone many years with two homebrew supplies I've built. One is a 5V 5A, and the other is a -12/0/+12 also 5A. Both are very simple and have always met my needs. I also have a variac driving a 50V transformer that I've used when I need an odd-ball voltage.

The internet is awash with power supply projects and schematics. If you define your requirements it should be pretty simple to narrow down the choices.

[alm]

Agreed with John about stating some requirements. Plenty of designs for linear adjustable power supplies. I like the design with two op-amps for voltage/current, an OR with diodes, and a constant current source connected to the base of the pass transistor. Other well-known designs are based on the LM723 (constant current mode is not great) or two LM317's. Most designs need a negative voltage below output low to regulate down to 0V, which is an important feature IMO.

Below 3-5A or so, linear is usually fine. Above that, or for high voltages, cooling will be bulky and expensive, so switching might be a better option. SMPS is more complex, however, and usually noisier output signal.

[Mr J]

A great inexpensive and rugged P/S kit is the RSR3010 from http://elexp.com/tst_3010.htm .I use the 3010 in my class room and the kids beat on them for years and when they do break it's very easy to fix.

If you want to make your own I would suggest fixed 5 volt, Adjustable Plus Voltage (+1.25 to 15V or so) and Adjustable Minus(-1.25 to 15V or so) Voltage. This will cover most electronic project like Digital, Op-Amps, Basic DC. I've attached a very simple triple power supply schematic. It uses a LM317, LM337 and LM7805.

There is a trick however if you want the LM317 or LM337 to get to zero volts. If you refer to my schematic instead of the pot (voltage adjustment) going to ground of the LM317, pull it to -1.25 volts, this will offset the voltage to zero when the pot is adjust to minimum volts out. Thus for the LM337 pull the pot (voltage adjustment) going to ground  and pull it to +1.25 volts. To do this you can use a couple of zener diodes or another 317 and 337 or voltage reference IC's. In my experience I really haven't need the need to go below 1.25 volts for basic electronic circuits. Anywho Let me know if you would like a schematic of the modified version and i can print one out for you.     

[Mr J]

Here is one i used with a switcher

[saturation]

Excellent construction.  Kudos!  Do you know what the ripple and noise is from the output?

[Mr J]

The top one is a linear p/s i just measured the following:

15.005V no load less than 2mv p-p (lowest my scope can go)
15.005V 150 Ohms 2.5mV p-p
14.795 full load 17 ohms 100mV p-p
% load regulation is about 1.39%

Sorry I don't have the switching p/s (second one pictured), gave it away to a student.

[semaphore]

How about this one? Look kind of cool.
http://www.microsyl.com/index.php/2010/03/31/bench-power-supply-0-25v-0-5amp/

[firstolast]

This might be useful. It is a simple enough circuit. I cant say how it would compare to others out there..

schematic: http://us1.webpublications.com.au/static/images/articles/i306/30603_8mg.jpg

which came from this Silicon Chip article, http://siliconchip.com.au/cms/A_30603/article.html

I got it from a post coming from "admin" (Dave Jones?) in the comments of this vlog, http://www.eevblog.com/2009/09/11/eevblog-30-jaycar-bench-lab-power-supply-review/

I just thought I'd share because I'm in the exact same boat as PianoKid1994 and this is what I've decided to try and build. Good luck =)

[saturation]

Thanks a bunch, that rounds out the full design, from schematic, build, to finally, it performance specification.

The top one is a linear p/s i just measured the following:

15.005V no load less than 2mv p-p (lowest my scope can go)
15.005V 150 Ohms 2.5mV p-p
14.795 full load 17 ohms 100mV p-p
% load regulation is about 1.39%

Sorry I don't have the switching p/s (second one pictured), gave it away to a student.

[Jon Chandler]

Here's a link to an approach I took.  I used a wall-wart switcher supply and the old standby the LM317 variable voltage regulator to come up with a decent design.  Nothing too fancy, but it provides stable voltages with fairly accurate justs just by turning the knob.

[DavidDLC]

I'm wondering why nobody put this link before:

http://hackaday.com/2010/12/09/atx-psu-turned-into-an-adjustable-voltage-bench-supply/

[Time]

Because he said hes not really interested in an ATX based supply.  That project crossed my mind too.

[DavidDLC]

That is true. 

[Bored@Work]

I'm wondering why nobody put this link before:

http://hackaday.com/2010/12/09/atx-psu-turned-into-an-adjustable-voltage-bench-supply/

Because they are dangerous shit, as discussed here already in the past.

[firewalker]

One of the best DIY power supply I have found is the following one.

http://www.tuxgraphic.org/electronics/200506/article379.shtml

It works perfectly in my bench for about 5 years now.

I have made some changes to the code (4 lines lcd, removed the pc communication part, added power consumption indication). Also I used a plain prototype board. The design is really neat and easy.

There is also a better design but unfortunately the schematic and the pcb are sold as a kit only.

http://tuxgraphics.org/electronics/201005/bench-power-supply-v3.shtml

[alm]

Using an AVR with an extremely slow ADC (a few kS/s or so I think), a brain-dead control algorithm (one 10-bit DAC step for every ADC sample) to replace a jellybean LM324 is a very stupid design, IMO. The bandwidth and slew rate will be absolutely horrible (the very first op-amp in history was orders of magnitude better), I believe it will barely be able to track 50/60Hz ripple. It appears to be a design that uses a micro just for the sake of using a micro, instead of choosing the most optimal solution. Yes, it's probably the simplest MCU-controlled PSU, but proper regulation is IMO much more important than digital control. A computer controlled PSU isn't that useful in a home lab.

A proper design would use op-amps to control the pass transistor(s), and feed the signal from the DACs to one of the inputs of the op-amp as reference. That's how all the big guys like Agilent do it. I also don't see the point of building your own R2R DAC when you can get them for <$1 in onsies, except educational value.

But it will probably be fine as long as the input is fairly clean and the load is fairly constant, and you don't care about noise. The kind of requirements that almost any power supply will satisfy.

[firewalker]

Maybe David is willing to do some review-testing of some "famous" bench DIY power Supplies.

[nothing551]

I am trying to build the dual tracking power supply from Silicon Chip.

I live in the US and we get 110 AC. The transformer in the schematic is a 60 VA, 0-30 V at 2 A. I couldn't find a transformer that is 110 on the primary that has the same power and secondary voltage. I found a 56 VA, 0-28 V at about 2.4. My question is can I use this transformer even though the schematic says 30 V or should I just get a 110 to 220 converter?

[IanB]

I am trying to build the dual tracking power supply from Silicon Chip.

I live in the US and we get 110 AC. The transformer in the schematic is a 60 VA, 0-30 V at 2 A. I couldn't find a transformer that is 110 on the primary that has the same power and secondary voltage. I found a 56 VA, 0-28 V at about 2.4. My question is can I use this transformer even though the schematic says 30 V or should I just get a 110 to 220 converter?

Note that the US has a nominal 120 V AC, though the actual voltage you find at your outlet may be lower than this depending on where you live.

I have in front of me an 18 V 2 A transformer and a 12 V 2 A transformer, both inexpensive and found in Fry's (the brand is Philmore). These could be put in series to get 30 V and 2 A. In short, keep looking and be a bit creative. If you can't find a suitable transformer inside the USA you are not looking hard enough.

[nessatse]

One of the best DIY power supply I have found is the following one.

http://www.tuxgraphic.org/electronics/200506/article379.shtml

It works perfectly in my bench for about 5 years now.

I have made some changes to the code (4 lines lcd, removed the pc communication part, added power consumption indication). Also I used a plain prototype board. The design is really neat and easy.

There is also a better design but unfortunately the schematic and the pcb are sold as a kit only.

http://tuxgraphics.org/electronics/201005/bench-power-supply-v3.shtml

I suspect the general consensus here may be that this is one of the worst possible designs for a bench PSU.  In fact it is so bad that it could be used as an good example of how not to design a PSU.   

[BravoV]

Can I put this psu design analogy with this example ? 

Imagine driving a car with all those fancy displays and gadgets, and there is a big computer screen in front of you replacing the front window, yeah, you can't see whatever outside of your car directly through the glass window. A camera mounted in front for the car streaming it's content through wireless network to one of the video streaming server located thousands kilometers away, and you're watching that video, again through wireless public network to have a clue what is happening in front of your car, and there is no direct connection what so ever from the camera to your screen locally in the car.

Now drive it home safely with minimum speed of say at cozy speed at 40 Km/h or approx 25 mph, not too fast right ? 

Just want to describe this psu in layman's terms to the "right" audience, and feel free to correct me if I'm wrong. 

[Joshua]

I am trying to build the dual tracking power supply from Silicon Chip.

I live in the US and we get 110 AC. The transformer in the schematic is a 60 VA, 0-30 V at 2 A. I couldn't find a transformer that is 110 on the primary that has the same power and secondary voltage. I found a 56 VA, 0-28 V at about 2.4. My question is can I use this transformer even though the schematic says 30 V or should I just get a 110 to 220 converter?

Look at      www.antekinc.com

I bought a big toroid transformer from them. Came decently quick in the mail, well packed, good shipping price. I just haven't got around to finishing the supply yet. Between high school studies and dual enrollment courses, I haven't been able to devote much time to my hobby. Yes, every spare moment I have is at my bench but I must keep my gpa (4.0) where it's at so I can get scholarships to go to college this fall. That, and I have access to a reasonable psu- so I'm getting through it. Not sure how I got off on all of that, but keep us posted on your build.

[Bored@Work]

I am trying to build the dual tracking power supply from Silicon Chip.

I live in the US and we get 110 AC. The transformer in the schematic is a 60 VA, 0-30 V at 2 A. I couldn't find a transformer that is 110 on the primary that has the same power and secondary voltage. I found a 56 VA, 0-28 V at about 2.4. My question is can I use this transformer even though the schematic says 30 V or should I just get a 110 to 220 converter?

It should be no big problem. You might need to adapt a few other components a little bit, to reduce the maximum adjustable output voltage.

[Mechatrommer]

why people like to complaint on "built and proven to be working" stuff? instead of providing better idea? have they actually build themself one similar or have experience with the part used in the link? iirc, avr adc can do KHz level of adc sampling (even faster using dedicated adc chip), human eye can notice slower than 30fps moving pictures, some front panel have to slow that down even further to avoid annoying jumping up and down figure on 7 segment display. and iirc, the very adc chips are used in Dave's psu design? for feedback loop? i maybe mistaken

One of the best DIY power supply I have found is the following one.
It works perfectly in my bench for about 5 years now....

...reply...