TGM10 - based on NAIM by Julian Vereker

[tvi]

Better quality but its a naim Klassique

 

[Bigun]

I remember few years ago I saw a Naim circuit (possibly their regulated supply schematic) that was shown with -ve rail about 0.7V-1V larger than the positive.

V. strange - sounds like they could have been toying with isolating the ground from the centre tap through a diode.

good photo TVI

 

[Bigun]

Temporary build-out

I re-purposed a chassis from another project, grabbed the transformer I had along with a standard bridge rectifier and put together a stereo box.

The volume control was hanging off some twisted pairs pulled out of some CAT5. I connected this box up to my PMC FB1 floor-standers using some bi-wired chunky speaker cables.

 

[Bigun]

Listening Impressions

I wired in my old Sony blu-ray player to spin some CD's.

Well now, it sounds very compelling.

Some impressions:

Adele - example of powerful female voice. Slightly gritty, no 'ess' and no irritations. The bass wasn't very present but the voice was powerful, just what you would expect from Adele.

Sade - example of a strong but smooth female voice. Punchy bass, superb rendition of her voice. Steel brush sounded as it should, cymbals had just the right amount of shimmer (not at all splashy).

Magdalena Kozena - example of operatic female voice. Superb recording, an energetic voice, sounding youthful and unconstrained, very enjoyable instrumental with textured oboe and 'in the room' piano

Blondie - example of female rock. Very clear, a bit of 'ess' on the vocal but everything was present.

Hooverphonic - example of a more processed sound, some slight 'ess' on female vocal. Very compelling, I didn't want to get up to change the disc, just keep listening.

Some generic comments. Overall impression was dominated by a vivid presentation, really easy to pick out every separate instrument in all the recordings (more so than most amplifiers I'm used to), slightly dry, really superb on any kind of percussion instrument. This is not a neutral amplifier and you don't feel the need to crank it loud unless you want a lot of energy from the music. Deep bass the only area that I would criticize at this point in time.

I'm really not sure that any more work is justified on this amp. But I need to try a better source, this CD player I'm using is limited in fidelity.

 

[dedegogo]

The collector resistor is 270 ohms (lower than the NAP140 1k because the LTP current is higher).

Ah! Thats close to my clone. Mine is 470 ohm with 2.2mA ccs. I think i wanted a high transconductance input stage due to fast output stage. I prefer a slower output stage but i wanted to go inline with Bigun's spirit of using fast output transistor. But this is also aimed to go with 12MHz 2SD718 that i have plenty.

 

[Ian Finch]

Well now, it sounds very compelling.

That's an enthusiastic rap, if ever I heard one. What you didn't say, with regard to the first impressions, was anything about the harsh sound you experienced. Can we assume that it's gone, you've become accustomed to it or it doesn't matter to you now? Is there something you would attribute any change(s) to?

 

[Bigun]

I'm not hearing any harshness now. But I can't say what caused it, could have been the benchtop regulated supplies, the B&W speaker, or the old tuner. With the unregulated supply, CD source and PMC speakers and good quality speaker cables it's gone.

 

[dedegogo]

Bigun, nice to hear that you like the amp. Hopefully you can judge later whether you can live with the amp or whether it is as good as TGM8 for music listening enjoyment.

I already love my design. This topology can actually made very linear, once compensation is carefully planned. Using fast output and 1mA-1.2mA LTP imbalance the distortion can be made very low (gain approx 20x, lower than Naim's) while maintaining low order harmonics domination.

The inherent bad phase performance of this topology is also solved by using lower gain than Naim is using.

 

[Ruwe]

V. strange - sounds like they could have been toying with isolating the ground from the centre tap through a diode.

OK, I found it. I knew that I've seen it somewhere. This is from Neil McBride's website and is about setting the regulated power supply that goes with the NAP135:

"The other trimmers on the reg board are to set the output of the regulator. The positive voltage rail should be set to +39.2 V and the negative rail to -39.8 V. Note that the voltages drift when the components are new, and also when they are cold just after turn on. So you want to have the amp sitting there for half an hour or so before doing your final setting. And also when you've run the amp in a bit, set the voltages and bias again."

 

[dedegogo]

OK, I found it. I knew that I've seen it somewhere. This is from Neil McBride's website and is about setting the regulated power supply that goes with the NAP135:

"The other trimmers on the reg board are to set the output of the regulator. The positive voltage rail should be set to +39.2 V and the negative rail to -39.8 V. Note that the voltages drift when the components are new, and also when they are cold just after turn on. So you want to have the amp sitting there for half an hour or so before doing your final setting. And also when you've run the amp in a bit, set the voltages and bias again."

This raised my eyebrow. Nice to hear that it is not from Naim. So the amp must have the negative rail with 0.6 volt lower??? I tried to reduce the negative rail of my design around 10 volts lower! And the performance is still acceptable! What i mean is i have changed various operating points or environments to see the amp linearity. To my surprise, variations didnt break the amp.

 

[nigel pearson]

Since we are investigating Naim tricks:
I remember few years ago I saw a Naim circuit (possibly their regulated supply schematic) that was shown with -ve rail about 0.7V-1V larger than the positive. My simple thought then was that it's just compensation for the Baxandall diode extra drop on the negative side and was there for symmetrical clipping.
Does anyone has information/confirmation about that?
Would it be only for better clipping, or it might be more underneath?
I've never seen other dual rail amplifier design with different supply rails.

Regarless of that I always let my VAS clip first. When yours ears are bleeding what matters that it is symetrical. That's why a Cascode VAS is worth a thought as the trade off is at monster volume. One can always tweak the rails to suit.

If making a bootstrap CCS check with a scope how it clips. Often a series resistor helps. Oscilloscope required.

I don't like the idea of 0R22 as a value with the 4700 uF. One should filter the VAS+LTP and feed raw DC to the dumpers. The drivers might like filtered also. If the dumpers are lets say at +5V up they work like the simple example voltage regulators ( with zener ). That means ripple reduction at this point is rather pointless and must degrade sound if fitted. In the " zener " bit it is far from pointless. Seldom does anyone seem to see this.

The reason this is important is easy to explain. If this amplifier was made class A by turning up it's bias it would almost certainly hum. When 50 000 uF fitted it still would hum. After fitting a regulated PSU hum would become excellent . The amplifier even at 10 watts becomes rather large. The NAP clone has this hum also. You never hear it because the music hides it. Mostly it's rather good how well it works. All the same it's there. NAP250? If you regualte up to the drivers mostly you get a free lunch.

 

[Ruwe]

This raised my eyebrow. Nice to hear that it is not from Naim.

I suspect it comes straight from Naim and their regulated power supplies. Not that we have to worry about it in the clones with unregulated PS. Just an interesting fact.

 

[nigel pearson]

In the day a Naim amplifer was considered to be 80% about the power supply. I would suggest if Julian Vereker was given a Denon PMA 250 inside a week it would have sounded like a Naim. Julian said to me they had found a few clones even then. I remember Julian saying how unlike a Naim they sounded, he seemed amused the could get it so wrong. That's why I said buy a Nait 2 and sell it at project end.

 

[Ruwe]

In the day a Naim amplifer was considered to be 80% about the power supply.

I remember a British hi-fi magazine somewhere in the early 90's ran a comparative (subjective) tests of 5 integrated amps.
One was Nait 2, the olive version. There was a photo of the internals and most of the case was occupied by the power transformer. Huge toroid (600VA according to the reviewer) for an amp of barely 2x30W.
The others were Musical Fidelity A1, Sugden and Cyrus - can't recall the models, but all were good old British classics.
The winner was Pioneer A-400, though. I wonder if anyone remember that one at all

 

[nigel pearson]

That Pioneer thing was a bit naughty. It was the right thing to do as the Gang of Four as they were called were too powerful. Naim, Linn , Rega, Nytech. Nytech the really interesting one. It became Ion and was always setting up new factories in remote areas.

NVA was the outsider, yet good. Crimson and spin off Quantum. The big Armstrong 700 was almost a world design, it could be from the USA, Japan, and less likely UK. The owner of Armstrong came to see me with the prototype as he knew my old boss when at Oxford ( Armstrong was started when a student, my boss supplied components through his company Duval Electronics , Related to the imfamous Claude Duval from Kidlington, Duval become in a way SSL mixing desks ) . I think I was first to see it outside of their group of friends. Armstrong was his first car. 700 series used a Hitachi SIL chip for the preamp stages.

http://datasheet.digchip.com/000/000-1-HA12017.pdf

 

[dedegogo]

That's why I said buy a Nait 2 and sell it at project end.

Yes its always useful to have a comparator. But lets assume Bigun can hear certain quality aspect with the Naim and it is not available with the clone, so what?!

But i think it is not difficult to get a few % of that quality aspect as it is "part" of the topology. At least, from a few clone candidates i have, they all share the same beauty i rarely get from other topologies.

But yes i found that it is not easy at all. The easiest topology is a CFA. This one (quasi baxandall with LTP input) is the hardest imho. Im glad that i can figure it out. I hope this will be my best amplifier. All aspect is good except one thing: performance under squarewave loading. I dont know how audible it will be.

 

[nigel pearson]

Here is the Nytech we held to be so wonderful at the time. Richard Hay told me he was asked to do this amp for a tabacco company who wanted to beat B&O. He knew it was pointless and when given up he had a design ready made. He was ex Radford.

You might think this some sad old UK dinosaur. NO WAY. Richard liked capacitor coupling. JLH and me also. JLH went back to it in 1980. Notice no choke nor resistor to the speakers. Notice also the different Cdom ( tick ). Richard taught Julian about distortion in the current readings and how to do it. The input stage a perfect version of how Julian liked it except Julian wanted the simple DC coupled output. Even the tabacco company's 15 watt version had some magic. Richard felt NPN PNP good enough by then to not use Quasi all NPN. This is the work of a genius. Alas I doubt if anyone will agree. Richard was a real teacher. Alas gone.

 

[nigel pearson]

Yes its always useful to have a comparator. But lets assume Bigun can hear certain quality aspect with the Naim and it is not available with the clone, so what?!

But i think it is not difficult to get a few % of that quality aspect as it is "part" of the topology. At least, from a few clone candidates i have, they all share the same beauty i rarely get from other topologies.

But yes I found that it is not easy at all. The easiest topology is a CFA. This one (quasi baxandall with LTP input) is the hardest imho. Im glad that i can figure it out. I hope this will be my best amplifier. All aspect is good except one thing: performance under squarewave loading. I dont know how audible it will be.

Were it not for the title here TGM10- based on JV I would be happy to agree.

I guess the Datsun 280 was based on a MGB GT/ E type/ Mustang ? I was talking with the Renault guy who decides things ( can't say ). They would love to bring it back and not change too much. The fact it failed to be a MGB makes it so more interesting.

 

[Bigun]

How many people listen to an original Naim without a Naim pre-amp ? - does anybody really listen to the sound of the power amp without it being coloured in some way by the pre. The Naim pre has it's own distortions to add and has perhaps even more scope for technical 'improvement' than the power amp.

In most cases, with modern gear, we don't need any voltage gain in a pre-amp. Nevertheless, I read that many people who like the old Naim gear universally search out a system made up from a combination of pre-amp and power-amp, and if possible an external power supply for the pre-amp. The favourites seem to the 32.5, 42.5 and 72, but many are swayed by the desirability for matching chassis.

If feels like the journey is not completed without considering a Naim-style pre-amp.

 

[Bigun]

I found a couple of Chesky CD's, these guys can produce a quality of recording that I've never heard from another label.

I'm getting a bit of 'ess' on some tracks of female vocal but it's not a dirty 'ess', it sounds more like it's in the recording since it's limited to certain tracks and certain artists. Anyhow, the clarity of sound (at normal volumes) continues to impress me. These recordings lack no deep bass, there's no issue with the deep bass. It's vivid and energetic in a way that is slightly unnatural so you are forced to listen to the music - it's hard to day dream.