Trying to eliminate noise totally

[AlainGr]

It's been a long time !

Alain

@ Alain You are right.... Sorry... 

No problem at all, just happy to see you again

Alain

[Gerard]

It's been a long time !

Alain

@ Alain You are right.... Sorry... 

No problem at all, just happy to see you again

Alain

Well that is nice from you to say.... 

[Flecko]

Hi Mani,

the first thing you should look at is an amplifier with low gain, because the gain defines how strong your amp will amplifiy the noise of your dac. Especially with the ultra high efficiency of your horns I would exclude anything above ~25db. 15-20db could be a good value to look for. The noise of the amp itself is normally no problem.
My personal way to get noise down is to have speakers with not so high efficiency (90db) an amp with low gain (21db) and a volume poti. With my poti fully opened I can hear a noise up to 30cm distance "ear to speaker". If I set the poti to listening levels the noise is so low I just can hear it with my ear directly at the membrane. "Hearing" is in a way exaggerated because I have to try very hard to hear it at all, it is more I imagine there must be a noise and maybe "yes" there semes to be something. (I have not made any earthing tweaks)

Greetings
Adrian

[manisandher]

Hi Flecko, thanks for your thoughts. But see next post...

Mani

[manisandher]

Arghhh!

I'm an idiot, and I admit it.

When I've been comparing amps in the past, one of the things I haven't been doing is adjusting the gain of the bass amp in the speakers. What I did today was to try all the amps (minus the Soul) but adjust the gain of the speakers' bass amp to get a nice balance. Without doing this, the SS amps can sound too anemic, and it's no wonder that I've preferred the low-powered SE valve amps, with their larger bass 'bloom'.


But what's perhaps more striking is the fact that the amps have different levels of noise in this order (from most to least):

1. Soul (15wpc DC-coupled valve SE)
2. Berning (5wpc DC-coupled valve SE)
3. Sauermann monos (40wpc class A PP)
4. Sanders (~500wpc class AB)
5. BD-Design monos (~100wpc gainclones - currently with input caps)

Stupidly, I've never tried the BD amps with the Swing speakers for any prolonged period of time until today. Why? Well firstly because of the aforementioned anemic sound (not their fault as I wasn't adjusting the bass gain). But also because one of the PSU transformers has always hummed from the day I received the amps. Bert immediately sent me a couple of replacement trannies to try, but to this day I've not gone ahead and taken the soldering iron out to do this.  I've now got the BD amps in the system and one of the trannies still hums a little, but nowhere near at the level of hum from the single Soul trannie.

I really am an idiot. With the bass gain much higher than 'normal', I really like what I'm hearing from the BD amps. I've said before that the BD amps are simply the cleanest sounding amps I've ever heard. But with the bass gain increased to a nice balance, they're very sweet and musical too.

As mentioned, they're also easily the quietest amps I've tried with the Swings. What's interesting is that the channel with the quiet PSU has virtually no noise coming out of the horn (or the powered bass driver for that matter). I really mean virtually nothing. With my ear in the horn, I can hear just a little hiss. That's all.

However, the channel with the humming trannie has humming noise coming out of the horn and bass. At a pretty low level, but it's still audible from a couple of meters away.

So I'm now wondering whether I even have a 'global' noise issue. Maybe I don't! Perhaps the issue is simply noisy PSUs? This would also explain why the BD amps are quieter than the Sauermann amp. The former are class AB and draw virtually nothing at idle, whilst the latter are class A, so the trannies are working overtime even at idle. (Edit: I think the Sauermann monos have a higher gain also.)

Of course, the next step will be to eliminate the noisy trannie from the single BD amp. I should have some time this weekend to try this and will report back.

As for the Soul, well I will only take it back now if it is as quiet as the BD amps. They are now my new reference.

Mani.

[CoenP]

No simple answer available.

Thanks for your thoughts Coen.

It presents a bit of a dilemma I think. Single-ended valve amps tend to be low-powered, so need high efficiency speakers. But these are ruthless with noise, as I'm finding. So the only way to go is to switch to push-pull... which kind of defeats the purpose IMO.

Mani.

Push pull will "eliminate" power supply noise, which is the most common source of noise in tube amps (so not originating in the tube itself). SE has hardly any power supply noise rejection so your Soul amp addresses this firmly with a tightly regulated supply according to the rap on the site. Nothing wrong here.

Good tubes used within their ratings generate very little noise. You can make a practcally noise free MM amp with tubes and a little care. Power tubes add very very little noise themselves, and the driver and voltage amp are allready dealing with a firm signal.

A well designed SE amp can be completely silent.

Regards, Coen

[manisandher]

Thanks Coen. I will remind Anthony of this if he comes back and says his amp is as quiet as it can be.

Cheers, Mani.

[manisandher]

I've got a 1KHz track recorded at 0dB FS. With the Sauermann monos I can still hear this track over the amp's (white) noise with XX set to -120dB! OK, that's with my ears right next to the 115dB horn.

Haha. I didn't try any lower with the Sauermanns, but with the quiet channel of the BD amp, I can hear the tone with XX's vol set to -144dB!!!

Weirdly, as the attenuation gets very high, the tone starts coming and going. At > -132dB attenuation, it seems to be riding on a 1-2Hz modulation.

Now, I have no idea if XX's vol control is at fault here. Or maybe there's some sort of 'Nick effect' going on - the house is 130 years old, so ghosts can't be ruled out.

Mani.

[Nick]

I've got a 1KHz track recorded at 0dB FS. With the Sauermann monos I can still hear this track over the amp's (white) noise with XX set to -120dB! OK, that's with my ears right next to the 115dB horn.

Haha. I didn't try any lower with the Sauermanns, but with the quiet channel of the BD amp, I can hear the tone with XX's vol set to -144dB!!!

Weirdly, as the attenuation gets very high, the tone starts coming and going. At > -132dB attenuation, it seems to be riding on a 1-2Hz modulation.

Now, I have no idea if XX's vol control is at fault here. Or maybe there's some sort of 'Nick effect' going on - the house is 130 years old, so ghosts can't be ruled out.

Mani.

Mani hi,

That is the breakthrough suspected I2S de modulation noise I get as well. I can hear it with music playing at -144db vol particularly music with a strong beat. I also use a 0db 1khz tone teach for testing. I don't get the modulation that is interesting.

Do you know what the gain is of you Sauremann amps are my bridged gainclones are about 36db IIRC. I wonder if amp gain it important.

Regards,

Nick.

Ps hears another test try playing the music / test signal into the NOS with the NOS turned ON but with switch 2 turned OFF ( eg the DAC board not powered on). Can you hear anything ?

[manisandher]

Nick, I'll try this over the weekend.

Mani.

Edit: Not sure of the gain of the Sauermann or BD amps.

[manisandher]

Slightly off topic...

I'm interested in understanding the mechanism that causes the bass to change its quality with different amplifiers. Generally, the more powerful amps have a tighter and more controlled bass. This is exactly as one would expect. HOWEVER... the amps are not driving the 15" bass driver directly. They are driving the in-built gainclone amp, which increases the LF gain by I guess 15dB or so to match the MF/HF horn sensitivity. Presumably, this in-built amp has a very high damping factor.

So how is it that the bass quality can change so much when none of the amps are even 'seeing' the bass driver? Any thoughts anyone?

Mani.

[PeterSt]

Mani,

Because the first amp (like the Soul) acts like a pre-amp for the second (the BD30 in your case). And the pre-amp determines the sound ...
(largely)

Peter

[PeterSt]

And Mani, there is one aspect of huge importance (but neglected by about everyone - ehm, including you because I already told it two weeks or so ago) :

When you change the volume of the bass section like it can be done with the Swings, you will be molesting the crossover (and not a little bit). Or the other way around : don't do that but connect a "first amp" with different gain than the one previously connected - same story.

Sad point to make : you won't know the "reference" hence what the exact (and I mean EXACT) gain must be, because the Swing comes with the bass amp only. So I said it : this is neglected by most ... if you know what I mean. But not anymore (if you know what I mean) because it is clear now.

So for example : That Orelino speaker comes with all the amplification for a reason; it guarantees that the XOver is still as how I tuned it (or we tuned it).
Get it now ? Please do, because it is crucial.

If you tune that volume you shouldn't watch for bass but watch the music in the XOver area. Not that this is really feasable (for 100% good result) but it still is better than "just watch the bass").
What you also can do (or maybe should do) is get yourself a radioshack SPL meter, hold it at 1 meter or so exactly between the top and bottom horn, create a 285Hz test signal, play that and let someone else slowly turn the bass volume. When the output is the highest the XOver should be at its best.
This, assumed that the XOver midpoint for the Orphean horn is 285 Hz indeed, which I don't know. But most probably so.

When you found the highest output, do the same for 400Hz and 110Hz but now hold the SPL meter in front of the respective horns. First of all the level of both should be the same as were you ended with the volume balance. If it is not, than or the XOver just can't be made well at all because the both slopes are not equal and you are in bad luck.
Still you can attempt now to make the levels equal for the both frequencies from the respective horns, and after this look at the 285Hz again (if it is that at all) in the middle of both horns like before. Now you have the levels right (and mangling those for the better means subjective ears) but the XOver will be wrong. To what degree (haha) will be shown by the SPL meter. Don't be shocked when you see 10dB or more less output.

When you can't make any sense out of it at all, switch the plus/minus from the bass section (not sure whether that is easy to do) and start all over. Better now ? then leave it at that.

Notice that you can create a sort of best of worse by lowering (not highering) the bass volume, which may let rise the 285Hz level. Don't let it rise more than the level from the Orphean. So, now envision that you are creating an upwards going graduate level from bass to mid. Not the best, but also no hole in the Xover.

All right, from now on you will never forget how important this is, because you will see the results yourself.

Super large disclaimer :
Although this may help somewhat, it is a method of nothing which starts with not knowing the exact XOver midpoint and the necessary guarantee that the slopes of both filters are equal. What it sure will teach you is how wrong it is for whatever listening situation.

Have fun (or maybe better drop the whole subject ?),
Peter

[PeterSt]

That is the breakthrough suspected I2S de modulation noise I get as well. I can hear it with music playing at -144db vol particularly music with a strong beat.

I don't agree Nick and I'm araid it is too hard to explain it. For me it is proven that this is not Mani's case, just because I saw it coming reading his posts. At least I see it "happening" because of no change in that level occurring anymore from -120dBFS and down. This is where the resolution of the chip stops but the glitching always continues. Remember ?
It nicely mimics the frequency - each frequency (zero crossing).

Your i2s story is true too, but don't tell me why nobody but you is bothered by this. So, everybody should perceive the tone like Mani has it (because of glitching), but no way this 30 or so uV will get profound when music plays through it at 100dB more of level (and 2V or whatever is in order).
The i2s situation is a totaly different one (call it a 2nd issue), but accidentally implies a very similar behavior.

Ok, I just helped with this going (too much) off topic. So let's not do that. There's a topic for that so we can continue it there when needed.

Regards,
Peter

[Nick]

Peter hi,

I didn't post the suggestion that Mani is getting I2S de-modulation breakthrough until until he mentioned the test tone can be heard at -144db.

Haha. I didn't try any lower with the Sauermanns, but with the quiet channel of the BD amp, I can hear the tone with XX's vol set to -144dB!!!

For me it is proven that this is not Mani's case, just because I saw it coming reading his posts. At least I see it "happening" because of no change in that level occurring anymore from -120dBFS and down. This is where the resolution of the chip stops but the glitching always continues.

From -122db to -144db I also get more or less constant volume of breakthrough but the characteristic of sound "improves" moving from -144db to -122 db as the I2S pulse width modulation effect starts to happen above -144db. What Mani is hearing could could still be either DAC glitch or I2S demodulation. The way to establish if it is I2S demodulation is there is to try playing music with the DAC board turned off. If the sound is still there this really points towards I2S demodulation being involved because with no power on the DAC chips so glitching cannot happen. This is why I asked Mani to try the test with his DAC section off.

(Sound withe the DAC board off will be quieter because the gain of the IV board is also switched off)

Regards,

Nick.