Another power supply thread

[GreenBall]

Hi all!

I need a decent power supply for my electronics workbench, and I'm considerig building my own.
At first I figured I'd find a decent schematic, order the parts and get to work, but after some reasearch I found the first part to be rather hard.

I have a healthy respect for the dangers of AC current and trusting my life on a random schematic from the web dosen't sit well with me.
Looking at the first forum pages here I found a thread where a diy design from a magazine was shown to be outright dangerous!

So where can one get a safe schematic for beginners?
The threads I looked into all seemed to point at flaws in designs, but surely there must be some well known designs one can safely use even though it's not flawless in efficiency?

I'm tempted to go for a simpler way out, avoiding the AC stuff by getting a 20v adapter with sufficient amps as input and dedign a circuit to scale it down as I wish.
Would that be viable?
I figured 0-15v would be a decent voltage range for my needs, so I assume the above input would suffice and thus avoid the dangers of handling AC directly myself.

Any thoughts about this?

[kizzap]

Basically what you are suggesting is exactly what a lot of people do, and exactly what happens in a number of power supplies themselves.

The vast majority of those plug pack devices you see are just a transformer (and hopefully a fuse) with the output you use being the secondary. Most power supplies just include this transformer in the enclosure for a number of reasons.

However what you are planning on doing sounds like a great idea. 

[Whales]

My desk power supply is a modified ATX (computer) PSU.  They're basically free in a lot of situations and can deliver more current than most laptop-style PSUs, but generally deliver 12V max.

[homebrew]

Welcome Greenball!

Great you ask before building highly dangerous thinks as many others do ...
It general it is a very good idea for beginners to avoid messing around with mains stuff and use an of-the-shelf ac adapter. However, please DO NOT USE cheap ass china switch mode supplies. They might be inherently dangerous due to their often low (sometimes even illegal) build quality, especially with respect to insulation and creepage distances!

So if you consider building a supply connected to the mains, here some suggestions:

1) Metal case + solid mains earth gounding. Also ground the transformer core if applicable.
2) Minimize the mains wiring by using an off-the-shelf Power Entry Module (like http://www.farnell.com/datasheets/1676792.pdf). These modules contain the connector, a fuse holder, a switch and an EMI filter. So you only have to connect your transformer to the module, apply some heat-shrinking for insulation and you are done. Make sure everything is sufficiently insulated even when the case is open, as you want to be able to make measurements in a safe way.

Regarding the circuit it highly depends on your needs. You selected the voltage range to be 0-15V. That is ok, but most simple LM317-type designs do not go below 1.25V (internal reference voltage) i.e.: http://www.electroniccircuitsdesign.com/power-supplies-circuits/lm317-adjustable-power-supply-electronic-project.html

If you really have that requirement you might need some negative supply voltage, like this design. Here is a good thread: www.eevblog.com/forum/beginners/zero-volts-from-a-lm-317/

If you would also require current regulation, things might get some more complicated ...
In general it is not so easy to build Voltage/Current regulators that are stable under all load conditions. And that is a core requirement for any good lab supply ...

But apart from that, these are all linear regulators. Therefore they will 'burn' a lot of power and require large heatsinks. Just imagine, you are regulating to 5V with an input voltage of say 25V. Drawing 1A, your regulator would drop 20W! Some soldering irons use less ... Heatsinks are specified in K/W, describing the rise of temperature in Kelvin (=Celsius scale for any delta) per Watt of energy dissipated. Hence a 3.2W/K heatsink (already nearly 100x50x30 mm large) would rise more than 60!!! degrees above ambient. Hence either a larger heatsink or a fan would be useful ...


Just my two cents ...,
Pete

[Vito_R]

But apart from that, these are all linear regulators. Therefore they will 'burn' a lot of power and require large heatsinks. Just imagine, you are regulating to 5V with an input voltage of say 25V. Drawing 1A, your regulator would drop 20W! Some soldering irons use less ... Heatsinks are specified in K/W, describing the rise of temperature in Kelvin (=Celsius scale for any delta) per Watt of energy dissipated. Hence a 3.2W/K heatsink (already nearly 100x50x30 mm large) would rise more than 60!!! degrees above ambient. Hence either a larger heatsink or a fan would be useful ...

Just my two cents ...,
Pete

I much prefer Linear power supplies vs smps for bench testing for the main reason they produce clean power.  SMPS produce a lot more noise and this could have an affect when you're working on radio or audio circuits for example.  I don't care if i'm losing power to heat, its only for testing not running it 24hrs/day.

[Azhar]

I was planning to build a power supply from scratch for myself, then I decided to just buy two of these ebay LM317 based Step Down modules:
http://www.ebay.com.au/itm/231295358801?ssPageName=STRK:MEWNX:IT&_trksid=p3984.m1439.l2649
and two of these V/A panel meters:
http://www.ebay.com.au/itm/380984804829?ssPageName=STRK:MEWNX:IT&_trksid=p3984.m1439.l2649
and install everything in an enclosure with a decent 25V transformer!
But I am not sure how reliable are these ebay power supply modules, still waiting them to arrive. 

[hamdi.tn]

For years i used a cheap power supply for satellite receiver give 3.3/5/12/24/33 V not much power but enough as a bench power supply. put in a case with couple of switch selector for output and a simple led indicator. total cost 15$ maybe.

[Syntax_Error]

I, too notice the large amount of homebrew power supply threads here in the beginner's section, but perhaps from a different perspective. As a beginner myself, I am on the side of having a strong interest in a homebuilt power supply, and I read every thread that comes up about it.

For the more experienced members here who comment on these threads, thanks for answering the same/similar questions over and over again. I do detect a certain amount of frustration at seeing the same threads over and over again, and this is understandable from your perspective. From mine, it is understandable that the threads keep appearing, because I have often thought that no one thread (or even all of them combined) has ever answered most of my questions or put the "simple" issue to rest, so another thread always seems warranted, again from a beginner's perspective.

Because of these things, I actually do ask for a good how-to sticky, or instead of a how-to, a pitfall and what's-what type thread, covering how's and why's, and maybe pointing to some key resources for further learning. I'm trying to walk a line here between asking for spoonfeeding, which nobody wants to do, but getting enough info out into a single point that everyone is satisfied. I myself have done a lot of reading, and think I can handle building a linear supply without burning the house down, so I am not saying it's sticky or bust. We can find the info ourselves, but it would be *nice* to have a sticky as a perfect "jumping off point" to start the learning journey.

Have you noticed how the same "it depends" answers are given? I get it, trust me. It's because IT DOES DEPEND on your intended uses. A sticky that covers this single aspect succinctly would pre-answer a ton of questions. Covering various applications and their associated input/output parameters that a designed must choose prior to actually starting a design. How do I know how much current I want? Voltage range? Why constant current? How many outputs is really needed? How many variable outputs? Can I power them from the same input section? I could go on forever. Once these things are determined, people can provide the information required for a really good answer.

[magetoo]

I'm in the same situation as Syntax_Error, and second the request for a sticky.

Just some very basic information, like a known good schematic of a LM317 based design and a few pointers to more information could probably answer a lot of the questions that people have.  I'd suggest linking to Dave's first few iterations of a PSU design too.

Ideally I'd like to see several suggested designs / topics (like one for a standard lab PSU, one for a split supply for audio, etc) as well as some theory, component suggestions, safety information, etc, but just a "TL;DR: build this" plus tips on where to go next would be useful to start with.

[jinzo]

I have been thinking about writing my own opinion on this topic. There're several things that I find interesting about this topic. There're a lot of info/topics/schematics/instructions floating on the internet but it's hard to find something that at least some portion of the people agree on. As I'm looking for a small power supply that I can use for my tasks (mostly ECU r/w and other chip r/w) - so up to ~12V (and a bit over, with some thinking about future ~24V). I will be not using it a lot (currently using hacked together ATX power supply) so I'm looking for something cheap to buy/build. As most of the old stuff is expensive like it's from pure gold I'm mostly looking for self-build stuff or kits - but I have yet to find a comprehensive design that I can easily use/build upon. There're some candidates (I'll try to compile the links at the end) - but nothing pops out after reading a ton of forum posts/websites (but I also don't want to take just some design I found on the internet - without proper documentation/design confirmation).
Another thing that bothers me in this field coming from software programming is that there're a lot of designs with unclear/no licenses and I would like to avoid that. So preferably something that has an open source design/documentation/schematic. I know I probably want too much - but with the plethora of questions and anwsers about this topic I would expect there is something like a "gold standard" to point people to for study/reference. Hopefully I'll find something, someday

[1] http://tuxgraphics.org/electronics/201005/bench-power-supply-v3.shtml
[2] http://dangerousprototypes.com/forum/viewtopic.php?f=19&t=6674&sid=29916b2353eb743f0e9c0549e41dabbd
I can't find the third one that looked really good - dual output, nice design, ...

[RogerMc]

I hope this isn't considered hijacking the thread, but ...

I am also in a situation where I'd like to build a PSU and have been reading as much as possible.

Been considering this item (in kit form)
http://www.testequipmentdepot.com/elenco/power-supplies/xp720.htm

Here's a direct link to the product manual
http://www.testequipmentdepot.com/elenco/pdfs/xp720k.pdf

Would it be possible for someone with experience to please comment on the schematic & design?

The functionality is well suited to my needs, but a big concern of mine is safety (mine and my project's safety).  There seems to be so much emphasis placed on constant current capability and this Elenco design is lacking the feature.  Am I likely to regret not having it?

Thanks and I hope this is considered a contribution to the thread.

RogerMc

[Vito_R]

I hope this isn't considered hijacking the thread, but ...

I am also in a situation where I'd like to build a PSU and have been reading as much as possible.

Been considering this item (in kit form)
http://www.testequipmentdepot.com/elenco/power-supplies/xp720.htm

Here's a direct link to the product manual
http://www.testequipmentdepot.com/elenco/pdfs/xp720k.pdf

Would it be possible for someone with experience to please comment on the schematic & design?

The functionality is well suited to my needs, but a big concern of mine is safety (mine and my project's safety).  There seems to be so much emphasis placed on constant current capability and this Elenco design is lacking the feature.  Am I likely to regret not having it?

Thanks and I hope this is considered a contribution to the thread.

RogerMc

Get an LM350K, mount it on a large heat sink, and you can build a simple but reliable 3 Amp variable power supply with only a few extra components:  fuse, transformer, bridge rectifier, resistor, a few caps and a POT.  It has built in short circuit and current limiting.  I built one when I was a kid over 25 years ago. It was my first electronics project.  Still using it today, never failed on me.

[David Hess]

Would it be possible for someone with experience to please comment on the schematic & design?

It is lacking some protection features which should be included on a bipolar supply but looks sound.  Integrated regulators are pretty foolproof; they included current limiting on the current boosted 5 volt supply which is good.

The functionality is well suited to my needs, but a big concern of mine is safety (mine and my project's safety).  There seems to be so much emphasis placed on constant current capability and this Elenco design is lacking the feature.  Am I likely to regret not having it?

The power level is low and the integrated regulators have short circuit current, thermal, and safe operating area protection built in so adjustable current limiting is not as necessary although it is certainly a nice feature.

[Vito_R]

Get an LM350K, mount it on a large heat sink, and you can build a simple but reliable 3 Amp variable power supply with only a few extra components:  fuse, transformer, bridge rectifier, resistor, a few caps and a POT.  It has built in short circuit and current limiting.  I built one when I was a kid over 25 years ago. It was my first electronics project.  Still using it today, never failed on me.

schematic for lm350 circuit

just add fuse, transformer, bridge rectifier and 10 000uf filtering cap + 100uf before the input

[rdl]

schematic for lm350 circuit

just add fuse, transformer, bridge rectifier and 10 000uf filtering cap + 100uf before the input

Wouldn't the LM350 benefit from a bypass cap on the adjust pin like the LM317?

Also, I usually add protection diodes on LM317 circuits, to avoid any load capacitance from dumping charge back through the regulator, and on the adjust pin bypass cap.

[Vito_R]

schematic for lm350 circuit

just add fuse, transformer, bridge rectifier and 10 000uf filtering cap + 100uf before the input

Wouldn't the LM350 benefit from a bypass cap on the adjust pin like the LM317?

Also, I usually add protection diodes on LM317 circuits, to avoid any load capacitance from dumping charge back through the regulator, and on the adjust pin bypass cap.

Yes it's just not on that schematic I just pulled it off the web to post it quickly.  In my P.S I built over 25 years ago I put a 10 or 22uf cap across the pot.  I'm sure I did that.

[jinzo]

Thanks for this - looks nice. But if I may ask - what's the difference between P.S. like this and the cheap brand (or offbrand for that matter) P.S. that cost quite a bit more than the cost I quickly summed up for a setup like this?

[Vito_R]

Thanks for this - looks nice. But if I may ask - what's the difference between P.S. like this and the cheap brand (or offbrand for that matter) P.S. that cost quite a bit more than the cost I quickly summed up for a setup like this?

Well if you're saying the P.S I listed above is cheaper to build then buying a cheap brand power supply then yes I agree.  The advantage is 1- you save money and 2- you get the satisfaction of building it yourself and learning something.  You can also make sure you use quality parts like Japanese capacitors unlike the cheap garbage that a cheap brand p.s will most probably use cheap chinese caps that will have a limited life for example.

As I do obviously use larger and more sophisticated power supplies today like BK precision or others, this old home made power supply from my childhood still sits on the bench as a reminder of simpler times.  It also reminds me that you still can build something of quality that will last you a lifetime, when you do it yourself.

[Vito_R]

Just to note I also had to replace all the capacitors in my BK precision when I discovered they used cheap crap Capxon capacitors in them.    I was very pissed off when I saw that and replaced them all with Nichicons.  I also had to add a heatsink on a voltage regulator / transistor that was not properly heat sinked.  So even some of the better known brands that you think are higher quality also try to shave on costs with piss poor parts sometimes.  So when you build it yourself you know exactly what you're putting in there.