advertisement


Pre-amp Nirvana…?

M

Mr Tibbs

Infinitely Baffled Member
For me, things have moved on since ‘pre-amp fun and games’ – well, they have since yesterday, anyway.

‘Fun and games’ put forward the notion that you can only get so far with super-supplies and regulation. The other half of the supply (the 0V return half) needs at least as much attention, if significant progress is to be made. Ultra-low impedance 0V power supply to pre amp links were tried, followed by the split 0V rails. I think it’s safe to say that anyone who’s tried the split 0V mod and those who have gone further using multiple transformers with multiple 0V returns, would agree that there is much performance to be gained by looking at this area.

Lately I’ve been thinking that something more radical needs to be tried- something that may actually provide the ultimate 0V return path – no path at all! So, how can this be done? Simple. Let me introduce to you this most wonderful of creations, the DC-DC converter.

Put simply, this device is a miniature switch-mode power supply. It takes in DC, chops it up at high frequency, and passes it through a tiny isolating transformer, finally rectifying (and usually regulating) it back to DC. The bit we’re interested in is the fact that it completely isolates both legs of the input DC, from the output DC. This means that if the device is placed close to the star 0V point in the pre-amp, the output –Ve can be directly linked to this point, creating the ideal 0V point. This is the ideal power supply scenario – where the power supply 0V and pre-amp 0V are one and the same. It’s the equivalent of having the whole power supply right inside the pre-amp.

There are other benefits too. As these devices ‘regenerate’ DC, they are pretty much immune to the quality of supply feeding them. No need for a Hicap or better, just a simple linear supply will do nicely. Something like a single rail Snaps would be ideal. Also, poor mains quality will not be an issue.

In my 102, I’ve fitted two converters, one per channel of line stages. Each converter feeds the existing three local 1086 reg’s per channel. The model of converter I’ve used is the Traco Ten 5-2423. It can operate over a range of IP voltage, and outputs +/- 15V, though I simply take 30V from across the outer legs to power the 1086’s. The primary 24V supply I’ve used is a simple one – probably more basic than a Snaps. The important thing to note here, is that the 0V of the primary supply must be taken directly to the converters, and not connected to the pre-amp in any way. This arrangement has made redundant a heavily modified Hicap, fitted out with six SR’s and split 0V.

How does it sound? Really, shockingly good. Everything about the music is better. There is an obvious extra layer of detail, with instruments and voices sounder purer and more ‘real’. The dynamics are bolder from top to bottom, with cymbals sounding clear and sharp and kick drum being felt firmly in the chest as well as the ears. Every piece of music I’ve played has struck me as being better emotionally portrayed, than I’ve heard before – so it’s not just a ‘HiFi’ thing.

I can’t recommend enough that you give this a try. It requires you to leave aside some well established ideas about what makes for a good power supply, but keep an open mind and give it a go.

Mr Tibbs
 
Mr Tibbs,

These DC converters sound relatively small, and yet sound like they have v. low output impedance, even with a modest supply. Is this so? I believe your ears but it sounds like we are getting something for nothing.

Regards,
David
"let a hundred flowers bloom"
 
David,

Yes, they are small. The type I'm using have a regulated output, so (in theory anyway) the output impedance is low. The primary supply used will (within reason) make no difference to the function of the convertors as they use SM technology.

I tried them as a means of solving the 0V return riddle. Naim spent a lot of time designing preamps with very extensive star wiring, only to blow much of the benefit by having the PSU 0V point three feet away from the preamp. The DC convertors eliminate this problem.

On the +Ve side, as far as regulation goes, the converters will not be as good as say SR's, but there is nothing to stop you using SR's along with the DC convertors.

I have deliberately kept things simple, to see if solving the 0V problem could reap big rewards. I now know it does.

Mr Tibbs
 
Mr Tibbs,

That's very interesting, would you be kind enough elaborate on how you have wired the converter up. I have attached a picture from the spec sheet to help.

Regards,

Mus
 
Mus,

Input from the supply;

Pins 2&3 negative in. Pins 22&23 positive in.
Again, the input should be taken straight to the converter(s) - DO NOT connect the negative side of the input to the preamp ground. Also, DO NOT use the incoming supply to power any other circuits in the preamp (like the output relay for instance), but power them from the output of a converter via a suitable regulator.

Output from the converter;

Pin 11 negative out. Pin 14 positive out. Site the converter(s) as close to the preamp star 0V point as possible and connect pin 11 to the star 0V point - a short link wire here is what you are after - we are trying to achieve zero impedance between the preamp star point and the converter negative terminal. The positive lead(s) out to the local reg's are not critical.

Pins not used; 9&16 - these are 'common' for when a split rail output is required.

HTH

Mr Tibbs
 
Other considerations;

Switched mode supplies are relatively noisy! They generate up to 50mV p-p, at the switching frequency (300kHz), and with harmonics. They spew out this noise in all directions. A simple regulated primary supply should stop the noise getting as far as the mains. On the output side, a stage of regulation should keep the noise from potentially de-stabilising the amp stage you are trying to feed. If you are not using a reg after the converter, a simple LC damping filter should be used in addition to the existing Naim RC filter. The type of converter I use has a shielded metal case - I would not risk the plastic covered type.

The local LT1086 reg's I use are connected direct to the individual stages (no RC filter) and with no apparent problems. I have yet to try powering the MC phono boards from the converters - I will report on this as soon as I've tried it. It's early days yet, and I will probably try out some low value inductors here and there to build in a safety margin, even if there are no apparent problems.

Linn (and others) have been using switched mode technology for years. If they can do it, so can we. Time for some music.

More later...

Mr Tibbs
 
Cheers Mr. Tibbs!

I have been wondering what have you been tinkering with.
Now we know. Very interesting, very interesting..

At least now we can say you have re-built your 102.
Put the superregs back there and we will have another
rave results!

Oz
 
I've got 50 copper stakes in my garden , each an earth for bits of my HiFi , keep tripping on the wires though , was thinking of just burying the lot , that'll solve earthing problems for good :D
 
Hi Mr. tibbs,

Indeed fabulous idea!,


The most surprising isue here is that the NAC sounds better with the noisy DC-DC converter.

This means that the low impedance from the power supply to the pre-amp is maybe more important than the supply noise (at least to some level of noise).

At 300kHz the local LM1086 regs. have only ~20db Ripple rejection!, Have you measured the output noise of the 1086 at 300Khz?. I am sure the Super regs here as local regs will be much better since thay have huge advantage V.S the 1086 in the noise figure in high frequencies.


Mr Tibbs said:
Linn (and others) have been using switched mode technology for years. If they can do it, so can we.
Mr Tibbs
Click to expand...

If switched mode supply good for the NAC.. Maybe those Class D amps are
not so bad ... ;)

Cheers,

Avi.
 
Mr. Tibbs,

I could not locate the 5-2423 converter from the Traco site. How can I find the right one? How much it cost?

Oz
 
Avi said:
The most surprising isue here is that the NAC sounds better with the noisy DC-DC converter.
Click to expand...

Avi, Yes, it seems to run counter to the accepted rules, but what you say next, is (IMO) the reason why;

This means that the low impedance from the power supply to the pre-amp is maybe more important than the supply noise (at least to some level of noise).
Click to expand...

And it's not even the +Ve rail that matters, it all hinges on having the -Ve of the converter connected right at the preamp star point.

martin said:
I'm not quite sure if the improvement comes for the reason you give - I'm still stewing on this - but the overall principle I have no problem with at all. A dab of LC is probably a good idea, but the much-smaller-box payoff is elegant (says the man with a 6VA preamp supply )
Click to expand...

Martin, ha ha... you want proof - I can understand that ;) In fact, it's easy to demonstrate (at least to myself) why and how this works. Insert about 300mm of wire between the preamp star point and the converter and the difference is easy to hear.

I saw that article you mentioned BTW. What they did was take the (raw) output from a hicap, then fed it through a pair of converters and back into a spare socket in the hicap - so that they could quickly swap between normal hicap and converter outputs. This does nothing at all to improve the 0V system, so they are entirely depending on the DC converters outputting better DC than the reg's of the hicap.

My approach is very different in that (looking at the spec's) I expect the DC converters to be poorer than good reg's, but that the 'ideal' 0V path they allow when placed in the preamp at the 0V point, more than makes up for the shortfall in quality of DC.

Linnik said:
I could not locate the 5-2423 converter from the Traco site. How can I find the right one? How much it cost?
Click to expand...

Oz, use the link I provided earlier, then select 1-8W DC-DC. Scroll down and select TEN 5. I got mine from RS (part No 293-2886). They will set you back about £50 a pair. Of course, the usual rules apply - I like what this mod does for the music - you might not agree!

Late last night I delved into my John Lee Hooker collection. The house was quiet, everyone was asleep, so the volume had to kept low. I listened for hours (no work today thankfully) and was struck by how good the music was sounding at this low volume. Hooker has a habit of changing his voice and playing from a whisper to a roar in the blink of an eye, and this dynamic content is mostly lost at low volume. Now, the full effect is there at much lower volume than before.

Mr Tibbs
 
Avi said:
At 300kHz the local LM1086 regs. have only ~20db Ripple rejection!, Have you measured the output noise of the 1086 at 300Khz?. I am sure the Super regs here as local regs will be much better since thay have huge advantage V.S the 1086 in the noise figure in high frequencies.
Click to expand...

Yeah, no doubt most of the noise will be getting through at high freq's. But, in an analogue circuit with limited bandwidth, is that really an issue? Some simple LC filtering could substantially reduce the HF noise, so it will be interesting to hear what difference it makes when I try it. I would tend to think this noise would very definitely make itself 'heard' if applied to a digital circuit though.

Super reg's are here on standby - should the need arise. At the minute though, there seems to be no need - this may change when I try powering the phono boards via the converters. To be honest, I hope to keep things as simple as possible. This has the potential (if my ears are hearing correctly) to make very high performance possible without the need for exotic power supplies/regulators (and the know how to make them work). First and foremost comes sound quality - that can't be compromised, but I can't help but smile at the fact that the whole preamp can sound so good with just the addition of a couple of off-the-shelf components and a small external PSU. If we can use some modern technology to our advantage then all the better.

martin said:
...says the man with a 6VA preamp supply
Click to expand...

he he... I like it.

Mr Tibbs
 
Thanks for everyone for your links and tips!

So, are we expecting your trials for the minimum of PSU's then, next, Mr. Tibbs.

It would be very exciting to hear how it will do it with a wallwart?
Or even with a switching wallwart?

Are we standing on the shore of Ultima Thule now?

Oz
 
... In fact, it's easy to demonstrate (at least to myself) why and how this works. Insert about 300mm of wire between the preamp star point and the converter and the difference is easy to hear.
Click to expand...
Well that's all that need be said then... I started drawing out the before and after schematics to have a think about impedances and side effects and quickly ran out of space. Simple demonstrations are good...

he he .. I like it
Click to expand...
Thought you might:
11%20DPA%20PSU.jpg
 
it all hinges on having the -Ve of the converter connected right at the preamp star point.
Click to expand...
Don't you get the same effect by having the -ve sense of an SR at the preamp star?

FWIW I think a couple of SR on the output of a Traco or other off the shelf SMPS with some filtering would make a very interesting and very thin (most important) PSU for a Naim pre. There's something extremely inelegant about using way oversized transformers and capacitors.

Paul
 
Okay Martin,

Tell us more about it all!

How is your power distribution and regulation built for the preamp
when you do use such a PSU sized for an electric toothbrush?

Oz
 
Oz - don't worry, it's not naimy at all.

It's the offboard supply for my Deltec (DPA50) preamp ; +/- 21v, plus a rail for the muting relay. There are further regulators inside the preamp itself, as well as a lot of fresh air:
14%20Inside%20the%20Pre.jpg

16%20SMT%20porn.jpg

Gratuitous SMT porn: the 'opamp' is a 55 x 25mm 4-layer board deposited on a ceramic substrate, and a reputed 500MHz GBW product :)
 
You must log in or register to reply here.


advertisement


Back
Top