Saturday, May 21, 2016

The Naim Audio Game...

For the last couple of weeks we have been catching up on what the world of Hi-Fi has been doing for the last 2 or 3 decades.  Now, I have been involved in Studio Sound, or at the least across it for all the same time, and find the differences rather striking.

One brand that has been constant from the mid 1970s till now is Naim Audio.  It seems they kind of invented the current Hi-Fi's networked Streaming Server products.  I've been playing stuff from a computer for so long I hadn't noticed.

The Power Amplifier I currently have in my small studio is a 50W per channel GainClone I put together from modules from ebay with a few modifications. Some $24 for the supply board and 2 stereo amps.  I added to that Zobel networks on the output and a relay speaker protection de-thump circuit. This is based on the LM3886 power opamp and very similar to the  very expensive, Audiophile 47 Laboratory Gaincard amplifier. Sounds great, but only cost around $100 to build using the other parts I had on hand. Well, I have experimented with power amplifiers and their  design since the early 1980s so had heatsinks and other hardware, but not the 2U rack case needed.  I did a blog post on this a year or so ago.

I happened to notice the recent Naim NAP 100 Power Amp , 50W per channel,  (equivalent to my own GainClone) and that these products still seem to have a good reputation like the old NAP160, 120, 140 and 250 amplifiers.   Making things sound "musical", which in the Audiophile world actually doesn't necessarily mean accurate.  I always liked the way they mechanically designed them. Very clean.

Naim Nap 120
I then saw that in Sydney the Naim NAP 100  sells for AUD$1,850..... wow.  How could that be?

Naim NAP 100 rear
Naim NAP 100 Front
Naim NAP 100 internal view

A bit of googling found the type of designs they use and that supply regulation was a key philosophy to Naim. By the way, the NAP 100 doesn't have that additional supply regulation.  All Naim power amp designs use a Quasi Complimentary output stage, based on the original RCA application note, with the Baxandall diode.
NAP 250 Power Amplifier Schematic
They also only ever use single power transistors, never wanting to parallel up devices to obtain higher power handling.  This very traditional design dates from the time that only NPN power transistors were good enough, many decades ago.  The output stage Vbe multiplier transistor is also not mounted on the power transistor heatsink, so thermal runaway is possible. An interesting design choice, but really, nothing special or magical.  They currently make a fuss about special custom power transistors, but they aren't that special, More marketing than reality.  It would seem the multiple regulated supply rails they use, and that for lots of their products they put the power supply in a completely separate box, is what sets them apart from other brands.

Their highest powered Naim NAP 500, some USD$21,950, uses 2 of their power amps in bridged mode per channel, 4 total to make a stereo amplifier to get around their self inflicted need to not parallel output stage transistors.  That is probably a far more complicated approach.

None of these design choices justifies the price, but I think that is part of the wonder of the current world of Hi-Fi.   Prices, claims and reality are only indirectly connected. As much style as technology.  Still, makes any Naim a desirable product.  They are cool and have that X Factor.

Not better than what I have built myself though, for much less $.  You can even buy clones of the NAP 140 power amplifier PCBs on ebay, for very reasonable prices.

And a random comment comparing the GainClone and NAP 140:
Just finished a gainclone type power amp with LM3886 chips and bog-standard circuit as in the National Semiconductor data sheet, plus a 280 VA power supply. With a simple double potentiometer.
Connected it up to a Naim CD3 and speakers and lo, wonder and horror at the same time, it sounds better than my Naim 72/140. In particular, the bass is just as strong, but less woolly, more detailed, more "tactile" if that makes any sense. Like an image in sharper focus. While the treble is gentle and sweet. This is after a few hours use.
Fascinating. Truly fascinating.

We can be contacted at Art & Technology


tropicalontour said...

I've been a big believer in MOSFET class D amplification for a while. Call me a heretic. :)

Then again, I was an early adopter of switched mode power supplies in audio amplifiers.

Most people can't hear anything worth a damn above 10kHz anyway. Those that claim they can and do are often revealed to be fooling themselves, when careful evidence is gathered.

Adrian Bruce @Megacurve said...

I have a 50W channel Class-T Amp here myself. It needs a speaker protection circuit, as these things sound great, until they do DC. I must say it does sound different from my LM3886 based unit. Not better or worse, just a "little" different.

Crispin said...

Can anybody guide me how to repair the right channel on my Naim NAP100? Answers other than “send it to Naim for £230+VAT service and repair”, please :-)

kraaiestaart said...

On the subject of quasi complementary output, this "older" technique is allso employed by the lm3886. It seems because it's hard to diffuse power pnp on chip ? Anyways , it doesn't seem to affect the perceived sound quality of this chip amp , taking into regard the many raving reviews of gainclones. Probably no one hears the difference between a well designed quasi or fully complementary output stage, I don't for sure.

Christian Thomas said...

Actually their power transistors are quite special indeed. They have some crazy specs like a power rating of 250W+ and 70A or 80A continuous, and hold up their hfe incredibly well, rather like the MJ14002 which are still at x100 at 10A and x50 at 20A (x70 @ 15A). I suspect they do rather better than this, given they have been through a number of iterations. The MJ14002 performance is probably where they started. And one of the ways they achieve this is to have two devices on each transistor, so you have perfect thermal tracking in a way you wouldn't get with two separate power transistors. So although Naim notionally object to paralleling devices, that is actually what they have, though without current sharing resistors.

I've been trying to find the commercial equivalent of these devices for some time now and haven't managed to, even though I know who the manufacturer is. There are modern devices that will nowadays hold up their hfe just as well, or better, but they're still not anywhere near as beefy.

Incidentally I don't think these devices were used in the smaller earlier amps like the 160 and 110. I seem to remember those used BDY62 or something nearby.