250watt PSU Idea

[Markgm]

Hi,
This idea is originally for a general-purpose PSU for class D amplifiers, with a couple of knick-knacks. I think this idea might make a good, more-so public project idea. (As can be seen from posts below, this is my first look at an electronics project beyond cursory exposure, and it's got ac mains in it, so I might not finish it without losing my eyesight or burning to a crisp) and you may have to finish it without me. :-)

This collection of ideas is based upon looking at a market for a plug-in variety of what is now available in battery form - a good, inexpensive PSU for today's class D amps.

In looking at the requirements for this, it occurs to me that as power is going up, is attention to current slipping? Because I think that when there are a couple of hundred watts involved, and the speakers are down to, say, 4 ohms, then the current should hit 10 amps on a DC cable. I'm seeing people using 18 gauge out there. We'll get to that.

Here are some of the general design points:
1) Economy: $150 is a center price for retail total - lower is better. That means subtract labor from kit form price, subtract and get parts cost (you know it changes on big parts buys for whoever wants to do that). But here we're talking $100 parts bought individually, I think.
2) Power: Clean DC power, in two flavors. The first flavor is what is described here, which goes up to a reasonable amount for general use. The idea here is 250 watts to the amp. The second flavor would be something going up to twice that and would have two power output boards in it - a separate, larger product.
3) Function: Batteries with chargers are one alternative that people either will or will not take, I think in the end. The functions include the versatility to use on different DC devices, at different voltages, such as to move from one amplifier to another, as well as grounding (described below).
4) Signal Quality: Power quality (noise) has to be targeted for this application - its the best amplification on earth. It has also been made available at very low prices. I'll kick the ball off - would less than 1 V p-p ripple be too much, or what would be achievable in this regard, as well as other types of noise from interferences.
5) Stable for this application - does what it says it does and is stable.

Put in a list in another way, here are the design philsophy points...

a) A banded EI laminate type transformer would be cost-effective and would suffice for this application, due in part to other parts of this design (such as b).

a2) Attention would be made to the transformer, to keep it away from saturation in the application. I don't think this calls for more than 250 watts. Some might think it calls for less wattage than that, which I would be curious to hear about, if it saves.

b) The power going into the transformer should be filtered on both sides before it gets there to help keep it clean, there.

c) Filtering of the DC output. The original idea here was to cut cost by taking the extra effort of using a larger number of smaller caps for decoupling and maybe to lower ESR.

--- filtering means both ways, above, b & c combined - positive and negative. Always impedance downstream and none upstream, and proper ground management. --- (stable)

d) Physical layout: Transformer in the front, board in the back, connectors on the back. Back panel has a clear lid that comes down over it, lockable - is that possible? A variable PSU in the home sounds dangerous when hooked up. Behind the lid: DC out connector, ground connector, and voltage control.

e) The DC umbilical: I mentioned above that I think this is 10 amp territory, with that or more than that also being handled in an audio application by the caps. These are such quick peaks. However, I would like to ask to consider this... Some effort has been applied getting clean power to the output of the PSU. Then it will go through a cable, and then, generally, to a big tank in the receiving amp. With a big tank at both ends of a cable, won't it hit high current quite a bit between the two? So I think this cable could get bigger than what people tend to think, and this gives rise to "connector wars" - what is the best connector in a world where people run umbilicals from DC PSUs to amplifiers? This is current for two channels. How about something that takes 12-gauge, and while we're at it, a ground connection, as part of a |new| 3-conductor cable format - which connectors should it be? And put those on one end, and a cable with the less-than-needed on the other end for an amplifier that doesn't receive with this new format. One could serve up a ground line connection from the amp to the PSU, couldn't they? Here's a question... Would you like that?

f) A voltage control: Something that can be used to protect against too much voltage going to a given amplifier or component.

g) Grounding: A sub-ground connection point for people who make or design or tinker with class D amps. A ground with its own transformer winding as the source of it, running at 2 volts below the minus of main, DC output.

h) soft-start

Other considerations:

Something that calls itself a "300-watt power supply" might be read as something that can be driven at 300 watts all day long. Audio doesn't do this. More like 30 watts on average, peaks hitting 300, when it is loud (on inefficient speakers). This should not be designed as a PSU for the use formally described, but as the latter. This unit does not need to continuously produce 300 watts for some number of years.

A little about myself: I'm learning as I go. I think this sounds like the kind of PSU that is needed these days. The next thing you know, you've got a PSU you would rely upon for a pre-amp, DAC or any other thing, with a voltage control. As I say, an idea for a bigger unit would have two output boards - each would have its own voltage control. Or, twice as much from one. Imagine a 700-watt PSU with an amp and a pre-amp hooked up to a pair of DC outputs.

Cheers,
Mark

[Markgm]

Some market-related thinking on this....

Begin stream of consciousness...

I've been a hifi enthusiast for decades, and I've owned equipment costing thousands of dollars over that period of time. It has become clear to me (and it is a well-shared sentiment) that class D amplifiers are driving hifi right into the ground. On price, that is.

The ground is a little farther down than mid-fi. I don't know how many companies I've seen in mid-fi who appeared, with hifi products at lower prices, then sold some, and then went out of business. I think the major contributors to this are:
1) The demand for hifi is finite (and much smaller than the general marketplace).
2) Hifi products don't turn over - they last. People don't throw this stuff away.

Mid-fi is an amp costing $500 10 years ago, or maybe today as well. It's been underdone.

The two-channel revolution has arrived. In 1928 it was proposed that either 3 or 4 channels would be a way for everyone to have stereophonic sound, well, within a practical price range. I think we've got damned good sound today, with a little homework, within a wide price range, and it is including good quality stereophonic sound (using two channels), more channels is more money, but you can do that, too. I think two damned fine channels is as widely available as it has ever been, and sufficient to call the general objective completed. I've certainly heard better than what I've got altogether, but these are not the advancements in sound quality so large behind me, as the smaller steps forward from here.

On the market, we have those two factors against us, which translate into saying that when a certain quality is reached, then a certain scale of manufacture should not be considered, especially when it gets too cheap, and there is no (larger) hifi demand to go with it. That demand takes years to develop, through exposure to it.

This marketplace occurs, as always, at the level of parts, kits and retail. Today, there is Internet-direct (manufacturer to customer). When I say "slammed into the ground", it is because I think that the price of hifi has gotten so low even a midfi manufacturer will have trouble. I mean, "slammed into kit price".

Like this... Anybody who wants more labor and less cost, does that. Want to pay more cost for less work? Do that. Whoever wants to do so just do so. The idea here is to put together ideas towards a general PSU of the day, with what it needs to be something people find useful, and would do for the money involved.

I think the money involved can come down. I also said 1 volt peak-to-peak, should be more like 0.1 V p-p. And do you have ideas about this? This is for the PSU kit that "slammed into the ground". Hifi at the most basic price.

As far as the other ideas... Grounding. Would you like to have a vacuum-cleaner ground connection lying around when looking for grounding problems? Well, this is the kind of thinking. Also, for the future, these connections don't have a ground, but the amps don't plug into anything else, either.

So price and quality are good questions, in terms of how that works out. What existing PSUs would you look at today for class D amps needing them?

Cheers,
Mark

[Markgm]

After all of that about market (i.e. 'demand')... A thing that can be known by its name is handy for that reason.

It would be something valuable that few people would make. Those are the quality and price points I am thinking of. Cheap, sufficient and long-lasting. Something that could never be made on a large scale. Ever!

When something is "black-boxed", like the digital component, or like these amps, I think it means that a large share of the market calls what they have "sufficient". That's what i mean by "black box". A sufficiently performed function for general use. Question: Could good audio power quality cost little enough to do this, and at what price, for this function?

I should have said (um, a lot of text up there), by all means, I've had a bit of a stream of consciousness. Please let me know what you think are good ideas and bad ideas.

Cheers,
Mark

[DF96]

It has been suggested in another thread that you may be trying to run before you can walk. Could you confirm that despite the above comments you have yet to design, build and test your first DIY mains power supply? Do you know how to calculate and measure supply ripple, for example? Or PSU output impedance?

Bear in mind that high current low voltage supplies carry their own safety problems.

[Markgm]

Originally Posted by DF96

It has been suggested in another thread that you may be trying to run before you can walk. Could you confirm that despite the above comments you have yet to design, build and test your first DIY mains power supply? Do you know how to calculate and measure supply ripple, for example? Or PSU output impedance?

Bear in mind that high current low voltage supplies carry their own safety problems.

I have to agree with you. I haven't designed and built my own ps yet. I have been reading about doing and how to do exactly the things you are talking about. And I am running before I can walk. It's true.

I am very excited as a customer. I am trying to describe what I think there is simply a market for. This is a developing need. I know how to be a customer. I've been one of those for decades. I think some of the things being said are market-wise.

If someone wants to lead in circuit design - good. If they didn't like what they saw otherwise, then leading the show is what they needed to do anyways. I know there can be a lot of disagreement about what is a best way of doing things, but I also see a more singular need here.

If one started off with a conventional transformer, a case, a cord and fuse-holder and a switch, and add in a CMC on the AC in, and the price of a couple of transistors or a rectifier bridge, then that is a start in defining this project. What are the number of dollars required to perform a given task? I am simply excited about this.

With a parts total that is $100 or less, and I would be happy to make a project out of cleaning up a simple design to start with. Using math. Using a scope. Making measurements and listening. Only saying what's real. "I read x".

I'd be happy to tell you what I am doing. What would you say to me about this idea, based upon your experience?

Part of what is happening here is to stop looking at this like something I will do, and start thinking about it as something that consensus will do. I don't know how. But I intend to. That's funny to put it that way, but I finally openly accepted that this is a project for a design that if anyone wanted it, they could just do it themselves if they felt that strongly about it, or to make kits, or to make finish products. Or to make mods. I enjoy accepting this way of framing the affair.

When I found it couldn't happen with a unregulated supply, I found it couldn't happen with what I've read so far. I saw parts and cost and requirements met. This thing needs to exist.

At this point, I have to drop to a knee, at the very least, and say that I haven't read about the internal operation of SMPS or linear supplies. I've read some general lines.... "unregulated is cheapest, good dynamics, noise is a problem"... "SMPS is next cheapest - deal with the switching noise" ... "linear costs the most - watch for dynamic restriction". When I saw unregulated as a hifi choice, that got me excited. So I started studying that one. How muddied is this original motive by a voltage control? Add in sub-ground on cost/hassle/convenience as well. The statement would be "cheapest, best for that, for everybody", and less function due to the basic resolution: a bare-bones PSU. That being an unregulated supply at lowest cost meeting (say 75% of) audiophile power quality requirements. (That was the original idea)

When factoring in batteries, I think something that plugs into the wall is the other half of what the market needs to offer. Some will do it, some won't.

I also came to the conclusion that this is something that should come from contributions here, because it is an idea for something that arises here in price. The basic unregulated PSU had the beauty of being resolved - of making a statement - of establishing a cost/quality point.

It picked up a couple of complications to make it more useful in a component marketplace. The sheer simplicity of the original idea is lost. I will ask this... How hard would it be to achieve consensus with a voltage control, and quality, and price?

People will be open to debate over the virtues of various things to filter. How much is debate, and how much can be resolved? I would take the rulings of experienced parties, because that is what I do - I'm reading. And so now it comes full circle. I'm putting this idea here.

I've enjoyed manufacturing of control panels for industrial equipment, so I wouldn't mind doing labor on PSUs for people who would rather pay the labor than do it themselves, you see. But it's not a company. Its DIY. It is a DIY power supply, I think. I think that's what happened.

[DF96]

OK. You may have noticed that power supplies are a debating point, and people don't even agree on exactly what the debate is about. If you are working at the level of evaluating other people's views, then you are likely to conclude that at least 50% of people are talking nonsense but you won't be able to decide which 50%.

One fairly clear dividing line is between those who believe that power supplies make an important contribution to sound, and those who believe that this is a sign of poor amplifier design. There is also a divide between those who aim for 'good sound' (which is not a well-defined concept) and those who aim for 'good engineering' (e.g. low ripple, low output impedance).

I fear that at present you are in the same position as a sociologist of science trying to determine if string theory is correct or not, simply by listening to particle theorists talking to each other. Good luck!

[Markgm]

The original idea, at one voltage, a basic, unregulated PSU, was what I was learning about how to do. It had its own virtue, in that it was that 'black box' idea. That would be the prohect, to make it clean and up to the task.

The voltage control, for example, was where I went past what I was studying. Did it lose its virtue, I wonder? It seems like something that would really make it a better use of parts and materials, don't you think? Well, did it lose its ability to resolve general concerns (at its own price)? Did it lose its ability to be comparable to anything at its own price?

[DF96]

I think the issue was that the voltage control idea made it clear that you were trying to operate/debate way past your current understanding of power supplies. Even the unregulated supply may have been beyond where you are, given some of the questions you were asking.

As well as your reading, can I suggest you download PSU Designer 2 and play with it.

You need to learn about half-wave and full-wave rectification, RMS and peak voltage, the difference between reservoir and smoothing capacitors, RC and LC smoothing, ripple calculation. Your aim should be to be able to estimate the output voltage and ripple of a PSU, simulate it with PSUD2, then explain the differences between your estimates and the simulation results. At this stage you will be beginning to understand PSUs, and will actually know more than some of the people who talk nonsense in the forums!

[Markgm]

Originally Posted by DF96

I think the issue was that the voltage control idea made it clear that you were trying to operate/debate way past your current understanding of power supplies. Even the unregulated supply may have been beyond where you are, given some of the questions you were asking.

As well as your reading, can I suggest you download PSU Designer 2 and play with it.

You need to learn about half-wave and full-wave rectification, RMS and peak voltage, the difference between reservoir and smoothing capacitors, RC and LC smoothing, ripple calculation. Your aim should be to be able to estimate the output voltage and ripple of a PSU, simulate it with PSUD2, then explain the differences between your estimates and the simulation results. At this stage you will be beginning to understand PSUs, and will actually know more than some of the people who talk nonsense in the forums!

Wow! Thank you for your candid details. To define the task - to measure it (the output) and explain it. What am I in for, doc? Thanks for the way you put that (for me).

I do love the idea of each part's impedance, inductance, capacitance, electricity traveling through conductors, and the puzzle of the collection of values, and going for a good collection.

Well, I think it's clear I have overshot where I started. I think that it is obtainable for me to get the point of measuring the output, and getting it to be acceptable. Part of that thinking is that a simple PSU (like the antek) is presently being used (out there), and modifications would improve its output.

But then, it would be deeper, in doing the math for designing the same kind of thing from scratch, due to some parts changes, such as a conventional transformer. As long as there will be filtering, which I am psyched about working on, might as well go a general, low-cost quality psu for (my) modern needs. I have a couple of class D amps now needing DC PSs. What's up with that?

I like the idea of good grounding - part of the original project. It seems to me that the voltage control is where I can't stay in the above areas, or, where I am going to have to do quite a bit more.

And it appears I've been shadow-boxing demons, by suggesting a done PSU (as described) wouldn't be achievable. That's got to be ridiculous. Existing designs are acheivable, and I think that will probably work. I think getting all the numbers right in an existing design (for this application) has sounded more to me like what to do. So I have been fearing false apparitions.

Regarding the ideas now on this page... So I own a couple of DC amps. What is the industry standard for connecting audio DC PSs to DC audio gear? I don't know of a standard for it. For connectors? What about the ground - wouldn't that be a good idea? Of course, you would need a connection for that on both units. Thought for the day: Who are the manufacturers here? Has anybody done this? Will anybody do this? Is this a DIY affair?

Cheers,
Mark

[Markgm]

Let me take the opportunity to say, I think we have this problem in software development. I have spent years in that, and some software on motherboards and other devices, and work on projects with millions of lines of code.

One way to do this is to just start writing code without any design. "Ad hoc" - seat of the pants programming. Also, test programming. Code that is not intended to be kept, or to be organized to begin with. Why? Ad hoc programming.

When the ad hoc program grows up, it reaches a size where it collapses under its own (unstructured) weight. It loses in its ability, by bad programming practices and failure to know and follow the rules, to survive with any comprehension of its interdependencies. Programmers become afraid of touching it, for fear of not knowing what will happen.

Honestly, I opened up a can of worms here that is so much larger than... "Modifications to existing designs"
That sounds like unstructured programming; that sounds like it isn't the rules section first.

Well, it is what it is. Spoken in appreciation of your rendering of what needs to be done to do this.

Programming starts with a simple set of rules, but large program development rightly involves a deeper and more structured set be followed - more piles of stuff to be on top of. Electronics seems to have about 1.4x10^5 times as many simple rules. (so to speak) They're not rules, they're the knowledge that is needed. You have to know things.

What do you think of the approach of, with existing designs, to learn all about them? To learn until knowing them intimately.