Discussion in 'Power Amps' started by Rotijon, Jul 17, 2017.
Good idea! I'll get you some measurements next week.
We don't want the measurement activity of the Vidar to distract from the release of the inevitable Aegir. Maybe measure an Aegir, too while you're at it.
Unfortunately, I'm kinda slammed this week, in-between the Aegir changes and about a half-dozen other things. So I'm not able to get you a complete suite of measurements. However, in terms of output power, these two graphs should clarify the power output question:
Vidar is about 1% THD at 35V RMS into 8 ohms, both channels driven, or 153W RMS.
Vidar is about 1% THD at 64V RMS into 8 ohms mono, or 512W RMS. Which also equates to a power output of 256W RMS into 4 ohms stereo, both channels driven.
As I said, comfortably above the ratings.
+50W w.r.t. either load. Very nice...
Now monos in 4 ohms before going into protection...
I would have done this as I have the 8-ohm brick resistors with the heatsinks everywhere, but I still don't feel comfortable feeding 35V at 8-ohms into the AverLab.
And only just above 0.005% THD? Disappointing. Should be 0.00001% THD.
You know, some really dumb people on the Internets are going to measure it under only mono configuration where the THD is just above 0.01% THD and proclaim to be defective and therefore "not recommended".
Thanks for posting these. I have a potentially stupid question about "power doubling." As has been discussed in this thread and others, many speaker manufacturers (including Magnepan (link)) recommend using an amplifier that doubles its power at 4 ohms. I noticed on the Vidar description page on the Schiit website that this power-doubling phenomenon is mentioned, and the rated specs show this power doubling with 100W RMS at 8 ohms and 200W RMS at 4 ohms. Magnepan's write-up says this phenomenon is directly linked to the quality/adequacy of the power supply, which makes sense to me.
From an engineering perspective, how does that power-doubling phenomenon relate to the measurements posted above? If power was truly doubling, then either the 8 ohm RMS measurement would be 128W RMS (256W RMS at 4 ohms divided by 2) or the 4 ohm RMS measurement would be 306W RMS (153W RMS at 8 ohms multiplied by 2), but this does not appear to be the case. (Again, not attacking just trying to understand.) Is the doubling just a matter of rated power, but no amplifier truly doubles per measurements to 1% THD at 4/8 ohms? Or does it depend on topology somehow?
The power being measured is well past what the Vidar is rated for. After that there is no real gaurentee that power doubles at 4 ohms. From was Schiit wrote it does the following at 1% THD:
153 watts @ 8 ohms
256 watts @ 4 ohms
Maybe you could push it to 306 watts, but distortion would get a lot worse.
Oh boy, I get to kill Santa Claus!
Or, "Yes, Virginia, there is no power-doubling amp."
Why? Internal impedances. When you draw more current with a lower impedance load, you lose volts through:
1. The transformer internal resistance (typical of an unregulated linear supply, or "most power amps").
2. The output transistor emitter resistance (used in nearly every power amp to ensure power-sharing between multiple devices, and thermal stability).
3. The internal transistor output resistance (unavoidable).
Considering that the output rails can sag 3-5V under heavy load, and considering the losses in the output emitters can reach 20+W at low impedances, and the fact that you're always going to have some resistance in the output transistors, this means that pure power doubling simply does not happen in the real world. A perusal of old Stereophile reviews of rated "power-doubling" amps will confirm this.
So, how close can you get to power doubling?
You can minimize the effect #1 by using a regulated output power supply, but this will either (a) give off massive heat if linear, or (b) be a switching supply. Even then, regulation isn't perfect.
For #2, you can cross your fingers and use no output emitter resistors. Good luck with that. I would never sell an amp like that. Some people have stronger stomachs for risk. Some people were Colin Chapman.
And to mitigate #3, you're looking at moving to an alternate universe where semiconductors are actually perfect and have no on resistance. Again, good luck with that one. Let me know if you make it there. Also please let me know if the ground planes are actually 0 ohms there. That would be very helpful.
So yeah, sorry. No true power-doubling amps. A switchmode-supply, no-output-emitter-resistor device could get closer. But it still won't be perfect.
So why rate "power doubled?"
Because it means you're designing for the rated power into half the impedance, which implies a stouter amp. Imagine, say, an amp that makes 100W into 8 ohms and only manages 110 into 4 ohms. This means some major sag, or actual current-limiting. It is also not unrealistic in the context of, say, AV receivers, which typically use current limiting, if they are Class AB. And, if you rated it as a power-doubling amp, it wouldn't be too impressive: 55/110W 8/4 ohms.
Hope this helps a bit! And sorry for killing Santa Claus.
Jason I don’t think Audio Santa Claus exists on SBAF - he was assasinated the minute he pulled up the sleigh on Marv’s roof, or possibly ripped apart by dogs.
Please keep dispelling the audio myths and tell the truth. We can handle the truth!!
Thanks (again) for taking the time! Very helpful explanation. With regard to the portion of the post quoted above, you reminded me of a white paper I had bookmarked a long time ago re: a regulated power supply for the Sanders Magtech amplifier: https://sanderssoundsystems.com/technical-white-papers/161-the-magtech-regulated-power-supply-wp. It looks like Sanders is making the same claims as you do in (a) and (b) above. Rather than use a switcher, it looks like he developed a circuit that uses two power supplies to achieve regulation: (1) a low-voltage "ride" supply that behaves like the power supply in an unregulated amp and (2) a higher-voltage "boost" supply that adds power to the ride supply when needed, each monitored by digital control circuitry.
I'm not an engineer, so defer to you and others as to the merits of the approach. However, running the numbers:
153W RMS @ 8 ohms
256W RMS @ 4 ohms
i.e., 83.7% of the way to true power-doubling
500W RMS @ 8 ohms
900W RMS @ 4 ohms
i.e., 90% of the way to true power-doubling
So even with a complicated, regulated power supply that is supposedly 100% efficient, one only gets a little over 6% closer to the impossible target of true power-doubling. And at much greater expense. (This isn't a knock on Sanders' product as I have no experience with it, was just curious.)
* These are from specs posted on the Sanders website, not bench measurements, which I could not find any of.
Yep, one shouldn't take recommendations from a snake oil salesman, but then the internut is filled with ignorance.
Santa Clause joined the Father, Son and Holy Ghost on the last train to the coast, affixed with a California 65 warning label.
Joke's on you. I don't have any testosterone left after being in this hobby for years.
Rather than worry about power doubling going from 8-ohms to 4-ohms load, just look at the power going into 4-ohms (say when using Maggies) and pretend the the 8-ohms figure doesn't exist.
Power going into 4-ohms is power going into 4-ohms.
Amp#1 is 200W into 8-ohms, and 400W into 4
Amp#2 is 385W into 8-ohms, and 550W into 4
All other things equal, go for amp #2 for maggie, despite it not "power doubling" like amp #1. Actually, go for amp #2 for everything.
Audiophile concerns and logic are convoluted and makes things unnecessarily complex. I blame Benchmark whitepapers, Paul McGowan, and Head-Fi.
Maximum power transfer theorem helps to explain *power doubling*
Is there value in measuring power at 1 ohm?
Before getting a Vidar I was trying to find an integrated that could dig deep enough to power some ATCs, as they have a particular reputation for being power hungry, with a measured sensitivity 3 or 4 db lower than the official 85db. It seemed though that even a 110w RMS per/chan Exposure integrated wouldn’t quite do them justice. But I noted two things about Naim amps; one individual, elsewhere, claimed a Naim Nait XS integrated drove his ATCs surprisingly well compared to his monoblocks, along with various comments re. the Supernait handling demanding speakers well, and that Naim amps simultaneously have lower official power ratings whilst having bigger transformers with more filtration by them than similarly priced integrateds. Their marketing materials also state that there are additional windings on their transformers to increase their current reserves.
I came across some measurements in a group test (Source, Mirror) that show that even though a Naim XS2 was rated quite a bit lower at 8, 4, and 2 ohms. At 1 ohm they put out 130w of dynamic power compared to all the others which drop to around 30w. According to other measurements Hifinews magazine have published it was only big power amps that could handle 1 ohm loads without dynamic power dropping off a cliff, the only other exception being a Marantz PM8006 at 78w.
As a laymen this implies it would have more grunt, handling current demands of difficult to drive speakers with more authority and less strain. I would appreciate any clarification on it’s true implications, or indeed whether there are any whatsoever. As with the ATCs at least they’re measured at Stereophile as not dropping below 2 ohms, and the other amps rate much higher down to that impedence.
EDIT: In practice which entry below would you consider most ideal, or which better for what scenario?
Dynamic power (<1% THD, 8/4/2/1ohm)
1. 175W / 325W / 555W / 30W
2. 105W / 178W / 275W / 130W
3. 155W / 285W / 480W / 25W
4. 76W / 138W / 228W / 50W
I suspect Naim would have better slam, because its has better power output at different Z loads.<--- Edited. It doesn't break down when something strange happens. When some low freq transient needs to be created there is energy in the source to create it.
Considering that a transient consists of different frequencies, when the lower freq part is not powered enough you won't "feel" it.
I am interested in what the audio field experts have to say about that subject.
SBA 12" driver with nominal impedance at 8 ohms. (check more examples on next page)
Separate names with a comma.