Trying to eliminate noise totally

[manisandher]

I didn't know where to put this, so thought I'd just place it in a new thread.

OK, I bought a new amp today. Yeah, my wife is ecstatic - I already have the BD-Design monos, the Sauermann monos, the Berning 300B and the Sanders Magtech... and now I go ahead and buy another 'expensive' amp. She's totally ecstatic. The only reason I'm still alive is because I promised I would sell at least one of my other amps. (Anyone looking for one?)

So why did I do it? Well, it's an amp that I've wanted for a while. It's another valve amp. It's a Tube Distinctions Soul 15 - a 15wpc single-ended amp with an ultra-wide bandwidth solid-state driver stage (>MHz, slew rate >600V/µs) and a single KT88 output valve per channel. Overall bandwidth is 7Hz-150KHz +/- 0dB. Yep that's correct, totally flat. Actually, the bandwidth has been limited by the designer because the amp is prone to picking up radio stations otherwise.

Unlike the Berning SET, the Soul 15 is a floating design - signal ground is NOT connected to PE.

But I've got a noise issue. With my 115dB/w horns, there is noise (just) audible from the listening chair. There's a little bit more noise than when using the BD-Design or Sauermann amps, both of which use balanced connections. Rather than being upset, I see this as a great opportunity to try to eliminate noise totally from my system.

I've identified a couple of things that affect the noise massively:

1. If I disconnect just one RCA cable (from NOS1 to amp), the noise virtually vanishes.

2. If I disconnect both RCA cables, the noise really does vanish. Nada, nothing, even with my ear inside the horn!

These two data points suggest to me that the amp itself is very quiet. (Remember, this is a single-ended valve amp driving 115dB horns!) But something happens when connected to the NOS1. The amp designer believes there is some sort of ground loop being generated in the NOS1 between the left and right channels, and hence why the noise is almost totally eliminated when just one RCA cable is disconnected.

I've tried my other NOS1 and both generate pretty much the same amount of noise.

Anyone have any ideas as to what I could try to eliminate this noise? Any help would be much appreciated.

(In case it helps, I'm using a 'regular' USB cable. I've also tried another PC with a 1m 'regular' USB cable, and the noise is pretty much identical.)

Cheers, Mani.

[CoenP]

Hi mani,

Your nr1 observation indeed points to a (ground) loopproblem between left and right channel.

This is about the internals of the NOS1, I guess Peter will jump in soon.

Dit you try to redo the offset of the NOS1?

Regards, Coen

[manisandher]

Hi Coen, thanks for your thoughts. I'll liaise with Peter and see what we come up.

Cheers, Mani.

PS. Are you familiar with Tube Distinctions? The designer, Anthony Matthews, worked with Tom Evans for a while, and apparently is the real mastermind behind the Linear A/B amps (along with another guy called Peter Chappell).

[CoenP]

Before I forget: a loop needs to be closed so there must be a connection between left and right channels in the (new) amp too.

So this can't be solved without the complete picture including the connections in your amp.


Regards, Coen

P.s. i am not familiar with the brand. I'll look into it.
[ edit i removed a useless sentence]

[Scroobius]

Hi Mani - just to say that my experience was that I had to separate the left and right channel grounds in my GC amplifier to get low noise levels. With them connected there was a lot of noise. Not sure if you said that yours were connected or not inside your amp.

Paul

[PeterSt]

Mani, ignorant me here :

Your GC should not produce noise, and your Sauermann shouldn't either. This has been talked through before.
Careful about the "shouldn't" because I obviously did not have those both physical same amp sets here. But the same for their design, yes.

To me this tells that you have a general other problem *first*, no matter you can't find/solve it.
Of course this does not prove that this other problem is playing a role here, but it can.

Peter

[PeterSt]

This is about the internals of the NOS1, I guess Peter will jump in soon.

Ok, not that soon, but now I'll give it a try.

Dit you try to redo the offset of the NOS1?

So, what about this ? Anything changed ?

Apart from that, yea, blame the poor NOS1. But I don't see how that could happen with the NOS1 as the cause. I think it will or can be related though ...

Signal ground is "obviously" connected in the NOS1 for both channels. This is because ground is created from the differential setup. Call that floating, but with an absolute 0V point.
Now, nothing tells that the same happens at the other end, especially not with the design of that amplifier. Well, if I judge correctly of course;

So, floating ground. Ok, but what is the ground reference then, and how is it created ? It's not a differential design as far as I can tell from this distance (I didn't look up specs or anything), so, floating yes, but absolute 0V how ?
Besides this, anything with a gnd and plus will lift the gnd somewhat. So, even when intended as 0V it will be somewhat higher. At least that is what I think.

Side note : I ever saw it happening anywhere that the DC Offset as measuered in the NOS1 changes. So, as you know yourself, we measure it here and I ask every customer to "get close to that", and all come up with the exact same (but give it a few 0.1mV difference which already happens because of longer warming up; we usually don't give it that much time.).

So, midpoint creation within the NOS1 is stronger than what the other side implies inherently. This is prone to groundloops maybe. But "in" the other side (I think).
And thus my cautious conclusion is that the groundloop emerges inside you amp.
(hey, I must blame something, right ? )

What I would try first, and now assuming that both channels inside the amp are separated for their floating ground (hard to imagine, but I think it can exist) is connecting the input gnds of the amp by means of a short wire (like 10cm between the both input terminals). If normal RCA you can just use the outside of the connector, assumed they are metal. But check with a multimeter whether both ends are connected through this means. Anyway, you'll find your way.

Alternatively you can do this on the NOS1's side (but I don't think that will help).
Or both sides.

When this helps, it is again my cautious conclusion that both generated grounds in the amp are not equal;
When this can be proven (or admitted by the manufacturer), "we" can consider whether the now noise free solution is the solution at all. Thus, better make those channels equal ?

What you can also try (when the above does not help) is applying a separate ground wire (per channel) of thicker gauge and of shorter length than the interlinks. Can't make a difference ? suprise yourself. Well, hopefully.
So, apply these in parallel over the interlinks. Again connecting them to the outside of the interlink connectors would be the most easy, but they must be connected to the gnd wire in the cable again.

Let us know what happens ...
Peter

PS: Advice of another kind : stop buying amps and keep your wife.

[Leo]

Well up to now he has been buying some great amps....

[CoenP]

I think you amp has only one powersupply. Signal grounds of both channels should come together there.

Next to the experiments Peter suggested, I propose to look into screening of the amp and cable. If the amp is floating wrt the chassis, you can ground the chassis and strongly reduce the influence of outside EM inside the box. Same thing for the interlinks (did you experiment with different ones?).

If this doesn't bring relief the noise source is likely inside one of the boxes. Probably some oscillation as result of parasytics (unintentional inductance and capacitance) in the loop.

Regards, Coen

[manisandher]

Thanks for all the ideas guys.

To start off with, I've just checked the NOS1's meters and they read -5.0mV and -0.5mV respectively. Is a relative offset of 4.5mV acceptable? I guess it is. (I can't remember what the respective values were when I received the upgraded NOS1 back from Peter.)

The other thing that I've noticed is that the power transformer of the amp is humming. According to the designer, this simply should not happen - it's dead quiet at his place. And furthermore, the noise coming out of the speakers is identical in quality to the sound of the transformer!

I'm going to proceed with the hypothesis that the humming transformer is cause of the noise through the speakers and will try a few things to see if I can reduce the humming. The main suspect right now is my dedicated mains rings with its two 5KVA balanced transformers. I'll try feeding everything off the house mains and see what happens.

Incidentally, the power supplies of Bert's amps hum loudly also. But the Sauermann and Sanders amp PSs don't. The Berning has a switch-mode PS which makes no audible noise, even with your ear close up. Yet the noise from the speakers even with the Berning is uncannily similar to that from the Soul amp that I'm currently using.

Thanks again for all your thoughts. If this post triggers some more, I'd love to hear.

Mani.

[CoenP]

Haven't we heard this all before?

I think your PC PSU is prime suspect of creating DC on the mains.

Did you allready pull the plug of the PC out of the powerbar? Is there still noise?

With a couple of HV diodes and some caps you can create a "DC blocker" (google for that). Some brands using big toroids (which are very susceptible to DC) have those on board.

regards, Coen

[manisandher]

Coen, I think things are a bit more complicated...

With everything running off the house mains, the noise is still exactly the same. And here I'm using my the other music PC from my office (totally different PSU, and absolutely no noise in my office system, even with preamp at full gain and ear next to Genelecs).

I've tried a Kemp DC blocker too and it makes absolutely zero difference.

Next on my list is to bring in my Sanders amp in and see what happens. Noise is quoted as 'More than 110dB below rated output'. The Sanders is the only amp I have with both RCA and XLR inputs, so might prove interesting.

Mani.

[CoenP]

Ok,

Having more than one amp makes perfect sense now!

So: All transformers humming on an isolated mains supply and not on the normal mains..... How does the isolation xformer sound, does it ham/rattle too?

Regards, Coen

[Nick]

Mani hi,

I remember reading superb reviews of the Soul amplifier in HiFi+ a few years ago. Have to say I really liked the ideas behind the amp design, IIRC it has number of very high performance power supplies for vairious roles in the amp.

Definatly on my list of amps to try after a lottery win 

Nick.

[CoenP]

Coen, I think things are a bit more complicated...

With everything running off the house mains, the noise is still exactly the same. And here I'm using my the other music PC from my office (totally different PSU, and absolutely no noise in my office system, even with preamp at full gain and ear next to Genelecs).

I've tried a Kemp DC blocker too and it makes absolutely zero difference.

I may have misread your post.

So your regular office system (at a different location?) is dead silent. If you move that pc to the "gand room" and put everything on the mains, you do have noise. Probably all amps are dead silent with shorted inputs (nothing but the speakers connected) or do the powertransformers hum?

This PSU humming could very well be the cause of the groundloop noise since the left and right channels are likely to be connected together in the vicinity of the magnetic fields (either from the x-former or psu caps).

Is your main rigs ring circuit braker connection ok?

Regards, Coen