It means an amp where there's by design no sag on the supply rails irrespective of the output demanded by the load. A balanced CCS-loaded SE classA would be one example.
@Barleywater - what do you mean by 'wonderful' here?
Commercially it is not particularly viable, Rick. When you draw the bottom line, you discover that all of it costs quite a penny, and it's not easy competing with cheap Far Eastern contraptions which look nice but sound from poor to very poor (at least the ones I came across).
This is further accented by the fact that cans are so anaylitical than any mistake you make along the way is immediately shown up for what it is. So, no quick'n'dirty job can cut it.
For example, I ended up with two 30VA toroids for a true dual mono setup, two shunt power supplies, each with 2 heat sinks and two boards, again each with a heat sink - pure class A does that, you know.
The reason - I think - why they tend to be so analytical is that they are indeed an easier load to drive, but they also do cover one's ears, and thus all but eliminates much of the extrenal noise and disturbance in general it becomes much easier to focus on the sound as there's less distraction. I believe this to be true even for open back phones, let alone closed back cans.
Also, they are generally single driver jobs (although there are multiway models out there from AKG, etc), so no crossover complications and losses and much less phase shifts.
However, and I can't stress this enough, because they are so revealing they depend heavily on whatever's driving them. They also make it very important where you draw your signal from. In that, I am in luck, as both my CD and tuner have dual outputs, fixed and variable, so one set goes to the preamp, and the second right into the headphones amp - as direct as possible. There is a source A/B selector on the cans amp.
If I may, I believe Rick was referring to fully regulated power supplies, the kind which varies by something like 0.1% or less from zero to full blast, and that assuming "full blast" is actually above what the cans amp might actually require.
Essentially invariable.
Perhaps a bit like the PSUs I used. They are fed by two 30 VA toroidal transformers, each feeding a shunt power supply fior each channel separately. The electronics are are then ran at class AA, meaning that if the whole circuit needs say 45 mA of current to stay in pure class A even with low impedance (say, 8 Ohms) cans, one then biases it at twice that, in full knowledge that there's no way under the Sun for it to ever approach such levels.
Theoretically, this should all add up to the nearest thing to perfect there is, but while that may be so, of course one is still left with the electronics. While all this helps, it's still no guarantee of excellent sound if the electronics are not really top notch.
Easy Dutchie , everyone on the Nog tonight ..
Yeah, even I am drinking full fat milk tonight ...
Let me know when i can get one of those ....
Wayne, I've been doing some calculating and adding (of prices and resources). You would truly need mucho dollaros for what you want in sheer materials.
The way I see it, you would need AT LEAST 5 pairs of high power devices per channel, assuming your power supply lines are no more than +/- 40V for the current stage. And some MIGHTY heat sinks, although a full calculation would probably be better off using forced air cooling.
But I did solve the problem of what to name it - Current Factory.
However, it can be done, we're looking at something like 800 Watts into 1 Ohms, even if that is a tall order.
Actually, I should say it HAS been done. Elektor, the European electronics journal (published on English, Dutch, German, French, Polish, etc) did do a project exactly like that some years ago. I will now assume you are interested and will look it up, it was a complete project, meaning it had its full complement of printed ciruit boards (which you might still be able to buy) and bills of materials.
How about that?
Yep, I reckon you're right - if it were viable then there would be someone out there making them.
I don't find this totally convincing but its certainly a possibility. Why I don't find it convincing is our brains/ears do a decent job of filtering out stuff according to the direction the sounds come from - the famous 'cocktail party effect'. With headphones I don't get the directionality the way I do with speakers - the sounds appear within my head, rather than laid out in the distance behind the speakers. The 'within the head' effect is telling me that somehow my brain's confused how to interpret what's being presented to it. Perhaps the fact that my brain localizes the sounds within my head is part of the reason there's an 'analytic' quality to headphones. Or perhaps the analytic question is a mix of all of the above.
I managed to make my amps sound closer to how my headphones sound in terms of dynamics by adding more caps to the supply, that's what led up to the hypothesis that perceived dynamics have a lot to do with the stiffness of the supply. In particular the LF impedance. In theory a regulator gives excellent low impedance at LF, in practice it doesn't sound like it does - in the sense that adding caps makes the LF snappier still when theory says the impedance should be dominated by the reg.
Yep, definitely.
Incidentally I wasn't talking about 'essentially invariable' supplies - making the supply super-stiff, rather I was thinking in the other direction. That of how to make the amp draw a perfectly constant current so supply stiffness didn't enter the picture at all, at least to a first order. Which means the supply sees just CCSs and all the speaker current is diverted between two EFs.
Last edited:
As you may have gathered, talking to me about stiff power supplies is knocking on an open door. I could not agree more.
I just ordered custom toroids for my Centurion project. A nominally 100/200 WPC power, these will be 500 VA transformers. They will be backed up by centrally paralleled 10,000 uF caps, two per supply line, 40,000 uF per channel, plus 6 locally mounted 2,200 uF caps per channels. All this put together means 26,200 uF per supply line, 52,400 uF per channel.
Given that my speakers are the exact opposite of Waynes, hitting an all time low of 6.5 Ohms
, all this is absolute terms probably an overkill, but then I believe one can't overdo it that easily. So I'm thinking about a new central power supply using 3 rather than 2 10,000 uF per supply line. This would, in my view, enable the amp to be effectively limited by PSU voltage only, not the current.Heck, it might even work for Wayne.
Have you ever listened to Etymotic Research ER 4S? Tonally, they are almost indistinguishable from AKG 240DFs, the 600 ohm version now discontinued and the most neutral headphone I have ever heard, but I haven't heard many Staxes. I know when I am recording and I take the ER4Ss off to listen to the real thing (orchestra and choir, usually, Royer SF24 microphone) there is no significant difference above around 60 Hz.
- Status
- Not open for further replies.