Simplest Linear PSU?

I am a fan of small chokes to eliminate the HF noise generated by most series regulators. 

Don’t be too critical of SMD. For semiconductor devices, the exact same dies are available in old fashioned THT and SMD. Difference: packaging. Packaging may affect heat rating, but for power applications you will find the devices are THT. In many cases newer devices with better performance are ONLY available in SMD.

Look more deeply.

As noted, if you understand the stuff, build your own. I do and they outperform pretty much anything you can afford to buy.

I'll also take issue with "simplest" implying that simpler is better. The best are not that simple. Mine are not that simple.  I use combinations of active and passive noise elimination.  Simpler would be nisier.

@clustrocasual

Obviously buy what you like, and what makes you comfortable.  But the purpose of these sites is to share info, and I'm trying to counter what i see as knee-jerk reactions to modern technology.  Some SMD IS bad.  But some can be very good. I want as many as i can impact ot understand that.

I found your reply hard to follow -- full of vague analogies (some off base) and such. I design with both and have listened extensively and find pretty much the opposite. The short, tidy connections i can achieve with SMD, combined with more up to date components, are (and its a tiny difference) superior. Some packages demand THT - for size, weight or simply historical reasons. So my boards are most often actually hybrid.

In an LPS, in particular, some components simply require size and mass (thermal mostly). But i doubt any serious design tries to use SMD for high power levels. So i’m not sure if you have some deep secret knowledge, or are just defending your position... if its the former please share, if its the latter, you might benefit from being mroe open minded and looking into the details of each design and what is appropriate. I had reservations (in some instances still do) but have worked through to find what better (most) and what’s not (long random list). discrete R2R dacs for example would be impractical on many counts without SMD. resistor tolerances, size, complexity of assembly, bad noise immunity (exposed surface area/difficulty in shielding) to name 4.

-G

SMD, not SMP = Surface Mount Devices

THT means through-hole devices

often the same devices, different leads. Which is why the criticism is misplaced, IMO. There are bad components of all types. And also good ones.

On the topic of SMPS (switched mode power supplies) the problem is noise.  Noise CAN be controlled, but it is difficult. That said notice some of the finest test gear now uses SMPSs.  I however, do not. And the ones we can afford are just nasty.

Guys - the amazing level of confusion continues, post after post.

The OP asked about power supplies that were made with SMDs. Maybe he mistyped, maybe he has no idea what he’s typing, but he did. SMDs are surface mount devices. Note this means the same old components in a different package, - no through holes. His question was for LPSs with

"the least amount of miniaturized SMD and semiconductor nonsense?"

This means he wants his transistors to have three long legs rather than three short tabs. Period. End. Nothing more to see here, move along. The dies are the SAME, except that sometimes you can buy better transistors int he modern parts and th eolder ones have been discontinued. e.g.: 2SC389, J201.

Somehow people moved on to Switched Mode Power Supplies and then combined the two. I mean, really? Does anyone have any clue what they are typing about?

Quick answer: not apparent.

The problem is that there can be issues with both. But they need careful clarification. I took the OP at his word (letters?) and discussed that the linear power supplies he desired could be made, maybe should be, with at least some surface mount devices for cost, size and the ability to robotically assemble (meanings cost and quality). Other parts (pas transistors) likely are better off THT for heat dissipation and current capability.

If someone wants real data from someone who designs, professionally, with all this stuff I’ll jump back in but at the moment its just too painful.

We also seem to have moved to R2R DACs! The connection, except that i used them as an example of why tiny, cheap surface mount technology is essential to audiophile grade modules, is pretty tenuous.

Transformer, Bridge Rectifier and filter caps.....

yes, that's the simplest but likely not the best.  For constant current draw stuff, multiple filter stages can be very very good, and sound remarkably right. but for anything requiring specific DC levels (like any ICs) or current surges, it has major issues and regulation is required. regulation brings its own issues and noise sources.

This is why i somewhat push back on the most  basic tenets espoused in this thread.

I should add that the link above describes what appear to be very well thought out supplies, at a not-insane price considering what's inside them.  I pay little attention to the stand alone LPS market since i design LPSes for use within products. And when i need my own personal outboard LPS i just build it ( i have genero PCBs that i fab and can adapt to most uses).

@lewm 

Regulation, if perfect brings noise to zero.  Nothing in this world is perfect.  Most regulators generate high frequency noise which can have nasty audible effects (likely beats, but it really doesn't matter - they do). I agree that overall they are the place to start, and are essential for many dynamic loads. But filtering can help.

Since i have way too many 100s of hours into research and trial/error/success i wont go into detail, but its worth looking at.

Everyone needs to keep in mind that psycho acoustically the signature of noise/distortion is often more important than its magnitude. We LIKE large amounts of consonant distortion in some tube amps.  But we HATE tiny amounts of dissonant distortion. This is music theory 101

@clustrocasual

So many things to unpack.

Look i dont really care what you believe, but i want you to understand, and every reply proves more and more that you are not following my points.

1. I don’t like the sound of SMD components, so seek to reduce as many as possible.

No really valid. there are garbage SMDs and great ones. What you are not getting is that as a design engineer i KNOW that many are identical, and some better.

2. Therefore, SMPS are definitely out. Hence the title - LPS desired.

SMPS and SMDF have nothing, repeat nothing to do with each other. I also dont like SMPS. See my comments above, apparently missed. Easy in this mess of a thread.

3. Many LPS still have extra circuitry and pack in a ton of SMD components. If they were at least through-hole, I could upgrade to audio grade (Audionote, discreet regulators, etc).

Whatever. Confused. (me)

4. KISS concept is the goal, and I intentionally left more details about my application out to research recommendations from people. Thanks for all the recommendations.

I agree with KISS. I also agree with Albert Einstein, who wryly said "everything ought to be made as simple as possible, and no simpler"

Please, please please uncouple SMD from SMPS. I have a SMPS attached to a commercial DAC down in my LR. Actually disconnected since i replace d it. There is not one SMD component in it.

I contract design for some of the high end houses.

If i may add - most SMD components are as cheap and crappy as possible. But one can get transistors, resistors and capacitors that are either identical or better than THT. They cost more. A ***lot*** more. They are hard to get at mega Chinese fab houses. But you can, and i do. And in the end lot more for SMD is much cheaper than THT after assembly, error rates, etc.

SMD provide much better QC and better assembly quality. They are cheaper, allowing me to spend money where it actually matters, rather than where amateurs have heard it matters. Sorry, tough but true.

Lighten up and get a good power supply. That, i agree, is a brilliant idea.

I had this argument with an old engineering prof of mine, oddly. Old (very!), cranky, but still sometimes brilliant. He finally did his home work and came around.

Back to our friends and the party, and some outstanding wine and music.

In my understanding, source components for obvious reasons require the most silent power supplies.

Absolutely. I go nuts getting really quiet power for moving coil RIAA phono stages. Think about the impact of even a small amount of noise on something that may be amplified 20,000X by the time it reaches the speaker. How do i get that number? In farmer’s math, my MC stage has a gain of 66dB, 2000X, my power amp about 20X, preamp about 3X, but of course the volume is mostly turned down significantly, lowering the effective gain. BTW when i say "my" they are both what i use but also designed by me - so i know them well. My daily drivers are all prototypes of some form, some from long ago.

Given the above 1/10,000th of a volt of input noise becomes two volts of output noise. Which is normal listening level! It would be all noise! A signal-to-noise ration of....... 1. :-) Puts things in perspective eh?

DACs are actually much more forgiving, since the voltage levels on a DAC are nominally much higher.  But still you are right - whatever goes into the beginning of the signal chain gets amplified.