advertisement


Naim style rectifier: why?

jpk

pfm Member
I try to understand why Naim power amps have transformers with two centre tapped windings. If I understand correctly the Naim configuration looks like this:

368rkZa.jpg


So two centre tapped transformer windings (or two transformers), ground returns into the transformer directly, and two diodes each for the positive and for the negative rail. It seems to me that in this configuration only half of the transformers power is used. So I tried to sketch a different configuration with only one transformer:

8k0a1X3.jpg


Ground returns into the transformer directly as well, and again there are only two diodes each for the positive and for the negative rail. I made a crude simulation which you can see in action here.

But I don't understand: what's the benefit of using two transformer windings as Naim does?
 
Last edited:
Interesting thought, but I think you are incorrect.
I think they are configured like your second diagram. Perhaps there are some amps like your first circuit, but why?
Lots do have 2 primary windings, that gives 120 volt or 240 volt options.

The bottom circuit is a 'standard bridge', the transformer centre tap gives 1/2 the voltage between either the +ve or -ve rails. You can therefore use a bridge rectifier package.
The top circuit does not allow the use of a bridge rectifier, you need individual diodes to make the two 'full wave' circuits.

I'm sure someone with more Naim experience will be along to explain different transformer configurations if any.
 
I do seem to recall from a NAP250 I repaired years ago that the OP is correct. Called the Bi-phase rectifier it was the most common in valve amps as only a single rectifier valve (two diodes per envelope) were required for full wave rectification. It does rather under utilise the transformer.... I recall thinking "why the hell have they done that?" at the time... I can't immediately think of any other power amp that uses that set up but I'm sure it can't be literally unique. This was in a 1979 NAP250 BTW and IIRC there was a row of 4 stud mount discrete rectifier diodes on a small "L" bracket.
It was not unknown in valve days to use a 250 0 250 transformer which was readily available with a Graetz bridge to get 500V rather than 250V (which would have been the normal output with a Bi-phase rect) for small transmitters...
 
There is a subtle difference that I think Naim still regard as important. The positive and negative ground return paths to the transformer centre taps are completely separate. A lot of very high frequency current pulses pass this way. Perhaps it isn't a good idea to make them share what is to them a significant impedance path and in this way modulate the shared ground's voltage? It can be measured. I might be able to dig out the paper where it was investigated.

This regime also prevents any net DC from ever going through either core though I don't expect this to be a significant risk in an AC coupled amplifier.

Deffo not a cheaper option for a given wattage of amp so I doubt it was done without good reason.

John
 
There is a subtle difference that I think Naim still regard as important. The positive and negative ground return paths to the transformer centre taps are completely separate. A lot of very high frequency current pulses pass this way. Perhaps it isn't a good idea to make them share what is to them a significant impedance path and in this way modulate the shared ground's voltage? It can be measured. I might be able to dig out the paper where it was investigated.

This regime also prevents any net DC from ever going through either core though I don't expect this to be a significant risk in an AC coupled amplifier.

Deffo not a cheaper option for a given wattage of amp so I doubt it was done without good reason.

John

It's tenuous... but just about possible. It's also a possibility that they copied a PSU from something made when rectifiers were expensive... not realising it was an inefficient and pretty obsolete technique... and just never changed it...
 
There is a subtle difference that I think Naim still regard as important. The positive and negative ground return paths to the transformer centre taps are completely separate.

I modified the second schematic in the first post so that it also has separate ground return paths.
 

I remember seeing a series arrangement like this in a 200W Hiwatt guitar amp, except the ground was at the bottom instead of the middle. I think they were running +700V to a bunch of 6550 or KT88 output valves, then used the center +350V to power the preamp section.
 
It's tenuous... but just about possible. It's also a possibility that they copied a PSU from something made when rectifiers were expensive... not realising it was an inefficient and pretty obsolete technique... and just never changed it...

Way back in the late 80's when I was at Uni I actually asked a a Naim Engineer who was at a HiFi show in Heathrow why Naim did this and that was the reason he gave to me. I remember it well because I was building my own NAP250 clone from the lifesize picture in their advertising blurb and had just made a note of the transistor types from the one they had opened up on display at the show!

John
 
I actually asked a a Naim Engineer who was at a HiFi show in Heathrow why Naim did this and that was the reason he gave to me

With "that was the reason" do you mean they copied a PSU or your earlier mention of positive and negative grounds returning completely separate? If the latter: I can't see any difference in my schematics regarding the grounds...

OK I will ask a silly question: as a small amount of ripple remains (before regulation downstream), could that ripple "pull" the GND a bit up and down depending on the phase of the AC in the transformator windings, so that in one case the phases would double up and make GND "shaky" and in the other case would cancle themselves out and make GND steady...?
 
I mean the later, i.e. that the reason was given as common mode interference on the combined 0v return to the transformer. These positive and negative rail charging pulses do not happen at the same time. The point in time at which the diodes conduct depends upon the instantaneous state of charge of the capacitor on that rail. I imagine heavy low bass amplifier demand (below 100Hz) in particular will effect the timing. The diode switch isn't clean as there is some ringing because of the transformer inductance and diode capacitance, so it isn't just a single spike of current through the capacitors and back to the transformers.

It is a long time ago now since I thought I fully understood this. I agree that the dual returns to the transformer do appear to minimise the issue. There will still be a finite impedance from the shared 0v to a single transformer winding which is a long way in a Naim Amp however. There is also the impedance in the winding itself which is shared and part of the ringing when the diodes switch. I think that the uncoordinated diode switches on a single winding will still give a shared 0v a voltage bounce and ripple which can effect the timing of switching on the opposite rail. A separate transformer winding is floating and apart from this interaction.

At the time and since I have on a number of occasions tried both schemes out, most notably on HackerNAPs I have built. Perhaps contrary to popular opinion it seems, I prefer the sound of a Class A/B amp with the Naim rectification method of the uppermost diagram and I am satisfied that it is better for me. I'm therefore interested in hearing arguments as to why it cannot improve amplifier performance.

I would suggest trying both and deciding which you prefer. I found two lower VA transformers with the Naim scheme sounded better that a single larger one without, so it isn't a cost issue IMHO.

John
 
I think a little lack of understanding on the very early amps got carried through and just left that way because it worked.
Interestingly the very early 160s and 250s had two positive only rails, and they were wired in series like a couple of battery cells to create a 0v between them and + and - 40v.
 
I think a little lack of understanding on the very early amps got carried through and just left that way because it worked.
Interestingly the very early 160s and 250s had two positive only rails, and they were wired in series like a couple of battery cells to create a 0v between them and + and - 40v.

I wonder, can you hear any difference between the two schemes?
 
I mean the later, i.e. that the reason was given as common mode interference on the combined 0v return to the transformer.

Thanks for the insight! Did you notice that in my second schematic the two windings are separate windings which share ground only at the starpoint near the "GND" terminal...?

Interestingly the very early 160s and 250s had two positive only rails, and they were wired in series like a couple of battery cells to create a 0v between them and + and - 40v.

I thought "positive" and "negative" are only terms in relation to each other, so both my schematics can be seen as positive supplies stacked on each other, with the 0V taken out at the middle...?
 
The very low impedance of the smoothing caps renders all other factors negligible anyway. Hence I regard transformers and rectifiers as having no influence on sound quality (within reason). I've even used 120VA transformers to power monoblocks capable of 200W... with 10 x 22,000uF smoothing caps on each one! I actually measured 214W from them. Obviously music has a high crest factor and is nothing like sine wave drive so even at completely mental volumes the average power was maybe 20W even if peaks of say 150W were being hit for very short periods (<1S) this could easily be supplied by that huge bank of capacitors.
It seems to more often than not be forgotten that the caps power the amp for >80-90% of the time with them being charged up 100 x a second during the remaining time. Whilst it beggars belief that anyone could think mains cables have any effect, this is the ultimate argument that they don't IMO.
 
Here's a question that's perhaps unrelated, but the topic made me think of it. If you have two separate CT transformers with the same rating, could you combine them as shown in the top diagram? Or is it possible only by using a single transformer with dual secondaries?

Also, am I correct in assuming that you could get double the voltage of each transformer? (The myriad configuration of transformers still throws me for a loop.)
 
Whilst it beggars belief that anyone could think mains cables have any effect, this is the ultimate argument that they don't IMO.

Try connecting your audio directly to the supply comming from the street and you will hear a big difference. And I say big. Even a simple guitar amp will sound totally different if you connect it directly to mains supply from the street (or closest possible to).

If you have two separate CT transformers with the same rating, could you combine them as shown in the top diagram?

Yes, it doesn't matter whether the windings are on the same core or on separate transformer cores in regard to rating / wiring.
 
Try connecting your audio directly to the supply comming from the street and you will hear a big difference. And I say big. Even a simple guitar amp will sound totally different if you connect it directly to mains supply from the street (or closest possible to).

As compared to... ???
 
Try connecting your audio directly to the supply comming from the street and you will hear a big difference. And I say big. Even a simple guitar amp will sound totally different if you connect it directly to mains supply from the street (or closest possible to).



Yes, it doesn't matter whether the windings are on the same core or on separate transformer cores in regard to rating / wiring.

Complete and utter bollox. You shouldn't be allowed to post on forums with lies like that.
 


advertisement


Back
Top