Rectifiers

[Oliver]

Are you looking to rectify an AC wave from a transformer?

To understand the rectification process, think of the diodes as controlling current direction. The output of the transformer is floating, i.e. it is not referenced to 0V earth potential, unless you ground one of its electrodes.

Are you designing a power supply?

[Oliver]

Can't you buy a miniature switch mode supply for this?

Do you know how much current is required for each line?

Your approach will not work as planned. A transformer will only transform AC to AC. Running DC through a transformer will only overheat the primary coil. If you want to get these AC volages, I suggest you use a multitap transformer, then rectify the AC to get unipolar waves. You can find designs for this type of scheme online, or in introductory DC power supply text books.

Have you got some regulators in mind? The rectified waveform will need to be filtered and regulated for the computer. Compuer supplies are usually done with switch mode regulators. You may need some more detailed help to get this designed well.

[Oliver]

Hi Moto,

The supply linked to by JLF65 was the type I was thinking of. If that is not small enough to fit into your plug pack, then I suspect you may be out of luck. If you are building a basic linear regulated power supply, then I doubt you will make it any smaller than the Pico PSU, though you may be able to, with a really good shoe horn.

Here is a data sheet with a basic +-12VDC supply design, in figure 18. There are various ways that this design could be improved, with transient filters and blockers, better filter caps, over current/heat/voltage protection, etc.

There are several things to note about this kind of design.

The regulators require an input unregulated DC voltage a bit greater than their output voltages. If their input voltages are too high, the regs will suffer from unnecessary heating / poor efficiency. To get the calculations right for this, a quick search turned up this link, which I think has most of the info you will need.

The regulators will need heat sinks, though probably not forced ventilation.

If you want really good handling of fault conditions, then the design gets a little more involved. Simple handling of fault conditions is not too difficult, but less robust. Faults on the AC mains side can be handled with a good outboard line filter. Small and cheap line filters are fairly limited, and usually have poor failure modes - i.e. if the fault is beyond the range of their components to handle, then the full fault is experienced by their load circuit. If you have a good quality mains supply, then cheap filters will be sufficient for nearly all faults. If you have a really poor supply, then it may be worth using a more expensive filter. If you really want to learn electronics in depth, then you could try making your own multistage, guaranteed failure mode circuit protection. There are many options and combinations of solutions available in this regard. One thing you can keep in mind, is that the regulators have some fault handling capacity, though a robust supply will handle regulator failure (they can open up in failure, and provide over current to the load). Designing fault handling is not extremely difficult, though it can be confusing for a beginner, and it is also not something which I have a lot of experience with.

To get a +5V supply, the usual means is to have an additional winding, or tap on your mains transformer, which is also referenced to the grounded center tap. You than have an additional regulator stage equivalent to your +12V supply. Alternatively, you may place a +5VDC regulator after your +12VDC supply stage. However, such a regulator would have to handle an input voltage quite a bit higher than the output voltage, and would experience degraded efficiency and higher heat dissipation. You could use a 7805 with a decent thermal mating surface, such as a TO220 package.

There are a many options for your +-12VDC regulators. Some dual tracking types could probably provide a high grade of regulation, though may be a little more complicated to design for. I suggest using linear regulators throughout, as they are much simpler than switch mode, and have fewer components, and less maths involved in their application.

Let me know how you go, and if you need some clarification on any of the above. I suspect it will still be simpler/smaller/better to adapt something like a Pico PSU for your application. Of course that may not be suitable, for whatever reason.

Cheers,

Oli


edit- PS: I hope you are really competant with electrical safety, and understand how to comply with your electical codes of conduct etc. It would be a real shame to put someone's life in danger, or burn a house down, over a miggy power supply. Also, understand that building this device may infringe on conditions of some of your insurance policies.

[Oliver]

pyrre wrote:
...link

I think that is a bit big for the intended application. One can use a DC-DC converter on a single chip, and with external components, the entire stage would fit on a medium sized coin. I haven't recommended this as they are a little more complicated to design for. A very simple linear regulator is fine for this, though a small heat sink would also be required.

On the subject of heat sinks, it should be possible to have all regulators mounted to a single heat sink; it is a common design feature to have a grounded heat sink on voltage regulators. Verify with data sheets, of course.

[Oliver]

pyrre wrote:
Why not build a pico psu inside an a500 psu box...

I think the point was to eliminate the need for a sizeable external PSU box, by placing all components in a plugpack.

Zeners do not make very good regulators. They are small and cheap, and can be useful for low power loads. They can be improved with use of output transistors. It also requires a little more effort in design to include good failure modes. In terms of voltage regulation, the zener provides a reference output which sits on its reverse breakdown IV load line, which performs less well than a linear regulator.

There is pleanty of information about zener regulators online, including various designs. I still use zener regs for some applications.

[Oliver]

motorollin wrote:
...given the possibility of it failing I'll have to just leave it I think.

Lots of people design these things without any regard for failure modes. Most of the time it is fine, too. Personally, I think if you can understand how to do so, the small expense in additional components is well worth it. If you are willing to read an electronics text, then low power linear PSU design is not an unmanageable goal. I wouldn't adivise making one without at least a basic understanding of what is involved, though.

[Oliver]

motorollin wrote:
My degree is taking most of my spare time...

Yeah, that's pretty much what I had expected. I was actually a little surprised that you would be taking on a design project at this time.

All the best for your studies.