Forward Biased

Lab-quality bench power supply (part 2)

So.  Now we have a nicely-idealized concept of a power supply (kind of like that ideal op-amp we’ve read about) and a list of 19-plus features to include.  All but the most inexperienced among us have rolled our eyes—or at least shaken our heads—a minimum of twice already at the thought of actualizing all that into a working prototype.  Even a non-working prototype.

First, as already mentioned, but worth a re-mention or two, is the fact that almost all the designs you’ll (eventually) find here will be in a form that allows easy omission of specific features that you don’t particularly care to include for whatever reason.  This power supply may well be one of the better examples of this concept, because there are several such features that you may have never before seen in a bench-top power supply, and see no reason to include here, especially if you already have similar functionality elsewhere in your own lab—e.g. the AC mains voltage monitor.  You may already have such a thing mounted on the shelving at the rear of your bench.  Even if so, you may not have a mains T.H.D. monitor, but would rather build it into a separate enclosure (perhaps with your mains voltage monitor.  Or not, as you can’t really see the point of that sort of thing anyway. 

If you choose to make use of the ideas posted here for a lab power supply, then you may well be in the majority if you choose to leave out some things, even if for sheer practicality.  The least likely person to do so would be me, and by the end of the day even I might decide to leave out a few.  No worries.  This typifies my approach to a lot of things.  Sometimes I entertain ideas that are, well, out of the mainstream, or ideas that I otherwise think are very unlikely to be chosen, but I still listen or otherwise consider them, usually to increase my confidence that I haven’t inadvertently left something out that I might otherwise have chosen.  Or something like that.

In other (hopefully better) words, the fact that I’ve included such a long list of features that I’d like to see in my own prototype in no way means that I will put undue effort into including them all.  I’ll be evaluating this continually as I go—and as I hear from you, of course.  I could, in fact, consider building the whole AC voltage/T.H.D. monitor thing as a separate project.  The Mains filter is another good candidate for a separate enclosure, as it can then be used in other places.  I’ll probably include some easy way to connect the T.H.D. meter to  the raw AC mains, then to filtered AC, to see what difference the filter makes under different circumstances.

See?  I’m about to talk myself out of including all of these in the power supply itself.

---

Next day:  I think I succeeded (see last sentence).  The line filter assembly takes up a fair amount of space, since its method for reducing common-mode noise is to use two power transformers back-to-back, each capable of the full available power post-filter; i.e. if I want a 500VA-capable filter, there must be two 500-VA transformers.  This is better built into a separate enclosure, perhaps along with both AC line monitors (“before” and “after,” as described above), than included in the power supply.  That way, if I do want to feed the power supply through the filter I can, but I can also use the filter with other things as well, since it should have several AC output receptacles built into it.

That takes care of items 16 and 17 on that list of features in the last post, and I’m going to have to think long and hard about no. 6, the 4.5-digit metering based on the ICL7135 (or MAX7135).  3.5-digit voltmeter modules are really inexpensive, and can be easily adapted to be current meters.  The 7135, on the other hand, once all the support circuitry is included, isn’t so inexpensive—in terms of either time or money.  It has better performance, sure—but is that performance really needed in this application?  I mean, enough to build eight freakin’ copies of it?  Or, if I want a really accurate and precise 4.5-digit bench volt/current meter, am I better advised just to build one separately?

Well, we have time to mull that one over, and in the meantime we’ll just refer to “metering” without being specific.  I am quite sure that whatever the choice is there will be eight of them, or at least one volt- and one current-meter per channel, if there ends up being less than four channels.

After reviewing the list of features again, I think the rest are pretty much there to stay.  For now.   So, herewith is a preliminary block diagram of the whole thing.

---

 

This work is licensed under a Creative Commons Attribution No Derivatives license.

A very highly-stable, very quiet, laboratory-quality bench power supply (intro)

And now for something completely different, ladies and gentlemen.

(Actually, I plan to generally have several subjects going on at a time, simply because not everyone will be interested in every category. But don’t worry—of course, I’ll be coming back to each one until finished, or until it’s obvious that no one is interested.)

All electronics experimenters, hobbyists or designers, even if all they do is build other people’s designs with no modifications, needs a good, reliable bench-top DC power supply, and there are more designs for this basic piece of equipment on the web now, probably, than for anything else. So if I’m bringing it up for discussion here, there must be some reason I think we can do better than basic—for those people who need, or just want, things better than basic.

Besides, at least for me—and I’ll bet for a lot of you—the motivation to design and build something myself comes at least as much from the fact that, if I just put in the time, care and parts, I can make it work much better than anything I can buy, or most designs I could build, as from any other factor. That one’s a real biggie for me.  Personal satisfaction and all.

And there are a lot of directions one can go with a power supply. Linear or switching? How many channels?  2?  3?  4?  Do you include metering, and if so, what kind?  At what resolution?  Do you want one channel to be able to track another? Floating or ground-referenced? Any other functions? I could go on and on with this, and so could you.

But I won’t. because I’m about to choose a few directions, and, if you know me, they won’t be hard to guess.