Regulated Power Supplies for Low Power Amps

PTL60

Member
2002-02-01 6:30 pm
I need some information and opinions about regulated power supplies for low power amps. Here's my example: Headphone Amplifiers.

I know headphone amps are very low powered and I'm sure using a regulated power supply is a good thing,but; Do headphone amps really need elaborate power supplies? Isn't a simple on-board regulator good enough?

Some of these headphone amp manufacturers claim that these elaborate beefed-up supplies actually make major improvements in sound quality. I'm not saying that these supplies are not good,but do you really need them for a headphone amp? I think it's just a marketing scheme on the part of these manufacturers to make people beleive that it is better so they can sell amps at higher prices.

Why would you need a $3000 headphone amp? or even a $1000 headphone amp?
 

paulb

Member
2001-06-01 4:53 pm
Calgary
I started researching headphone amps back when I noticed I had room in my power amplifier case for 'something else'. The headwize.com site has lots of information.
My understanding is that if the headphone amp is simply based on an opamp (the OPA134 is a typical candidate), the power supply isn't really that critical because of the high PSRR of the opamp. But a Class A design can pass through all the shmuck from the power supply and the resulting sound quality is very much dependent on the supply voltage quality.
My personal view is that regulation isn't all that important but a low-noise supply is, so my Class A headphone amp uses a capacitance multiplier but not a regulator as such.
 

paulb

Member
2001-06-01 4:53 pm
Calgary
Another point: one of the posts on the headwize site was somebody who had cut open the box of a fairly expensive Grado headphone amplifier (I think a $200 or $300 amp). What was inside?
One dual opamp, a few caps and resistors. That's it. Maybe 3 dollars worth of parts.
I was in a nice wooden case though, that I'm sure contributed to the sound quality.
 

PTL60

Member
2002-02-01 6:30 pm
Well, that's what I mean about these super expensive headphone amps with super elaborate power supplies.I mean,it looks nice and seems,WOW,this amp has got to be awesome!! For example,the Grado amp that you mentioned. It has very cheap parts. It looks nice on the outside with the mahogany wood cabinet,but so what!!

I think these amps are marketed towards the so called "Audiophile" community. Why would someone pay $350 for a block of wood with $10 worth of parts inside? On the other side,look at the headphone amps from Creek Audio; for less than $350 you can get a Class A headphone amp with a regulated power supply and built quite well,including an Alps Blue Velvet Pot. That's not bad if you can't build your own amp.

But also, why would you spend thousands of dollars on a headphone amp that puts out less than a Watt of Power?? Only a few milliwatts will burst your eardrums.So,I beleive that a simple on-board voltage regulator should be just fine for an amp with such low power.
 
Hi all....

I with you on 'getting the most out of the best... ' and the headwize mosfet headphone buffer of (szekeres... or something like that ...).. will do just that...

I just finished my third version of that amp.. I liked my first (the most simple) the most... the circuit has inherent good PSRR but the input biasing needs good attention (filter) unlike the published design.. but I still get very good results with that design

one comment though: 32 ohm Headphone sound loud even at 30mV .. that why very good noise free supply is needed.... 1mV of rimple may be audible?

goodluck,
Thijs
 
Guess what I'm building right now? A headphone amp, and I've researched it VERY thoroughly. A couple of hints. If you are going with op-amps(which I did) then buy the best you can, because, remember, you are only buying only two. I think the op-amp is probably the single most important component in the amp. Don't skimp here. I used Analog Devices AD797 which has REALLY low distortion. If you want high-quality amp using op-amps, then a well regulated supply is a must. The better the power supply, the better the op-amps will preform. My project total so far(everything but the enclosure) has cost $75can (about $45usd) and is expected to perform VERY well.

I did a crapload of research, so if anyone has any questions about building a headphone amp, then maybe I can help, or at least point you in the right direction.

p.s. I'm not surprised that commercial units were crappy scams. There are WAY too many "audiophile" components and accessories which are bogus.
hmmm... maybe i'll start another thread about that...
 
Many people use the output of a regulator straight into the circuit.
I tend to use a regulator into a large capacitor bank. The more capacitance you use, the more the 'sound' of the capacitors tends to dominate the 'sound' of the regulator. I would think that in a case such as this where the current draw is going to be fairly small, you could use a few uF of pretty good capacitance with a fairly simple regulator and end up with a power supply that sounds like a million dollars.
That said, I'm a big fan of discrete regulator circuits, rather than chips. I do sometimes use opamps as the "brains" of a regulator, but still with discrete pass devices.

Grey
 

PTL60

Member
2002-02-01 6:30 pm
I guess it's true that a lot of these audiophile companies have components that are way overpriced. Even though most people cannot build their own amps,you can still buy a very good one at an affordable price. I don't beleive you need to spend thousands of dollars,especially for a headphone amp!!

Most people that buy headphone amps usually buy the ones that cost between $200-$300 and find them to be great. So, why do they sell and make models that cost thousands?; because it caters to the "Audiophile Mind" and the end result is;the company will make more money.
 
Nothing exceeds like excess!

What possible advantage could there be to feeding the LM317 and LM337 regulators twice their output voltage? "Margin for error and undervoltage", sure, but you're p***ing away half of the input power in the form of heat at the regulator! This strikes me as excessive, even above and beyond the normal standards of audiophilia.

Then again, to quote the designer: "The ultra-regulation of the power supply (figure 3) is so over the top and unnecessary that most, if not all, people building this amplifier would not even notice the difference."

I think you could cut the input voltages down to -+/- 36VDC and never notice the difference.

BTW, what about supply bypassing within the amplifier itself?
 
At first I thought that Gilmore's circuit was totally over the top for a headphone amp. Then, having had time to reflect and really think hard about the factors that have driven this elaborate, 317-frying design, I conclude that I my initial instincts were not incorrect. Is this the result of a lurid fetish for electronic components? Perhaps this mollycoddling is an indication of the frail integrity of the amplifer on the end of it! Good grief.
BAM
(having a bad day :mad: ;) )
 
As far as regulators for preamps go, i agree with paulb on this one. Each different amp circuit will have a different level of sensetivity to the power supply, and various combinations of active regulation and passive regulation will produce different results.

With opamp based circuits, I have had good results with plain old 317/337 regulators followed up by simple passive circuits like C-R-C filters, even just with small resistances, and the addition of ferrite beads etc. For really low noise supplies, the opamp with external pass transistor designs work well.

seangoesbonk: you might also want to consider the OPA134 and OPA627 opamps. The OPA627 is about the closest thing to a technically perfect opamp as I have seen. IMHO the Burr-Browns sound a little nicer, and i think you'll find that once you factor in current noise, these opamps are also quieter (depending on source impedance of course) and have higher input impedance. At the moment, my favorite preamp design uses an OPA134/627 with external discrete SE class-A MOSFET buffer inside the feedback loop to drive current to the laod. I'd point you to my website, but it's STILL down... grr. gotta fix that soon.