Gainclone heaven ?

[stefanobilliani]

Is it really DC ,something that let your woofer excurs for a centimeter when music Stops ? What can capacitors really do about that?

Yes, true DC which will misalign the cone in one direction and even burning the coils (depending on the output power and DC Voltage) if its on there too long (getting overheated while not moving!) but that is the least of the problem...

A capacitor in series with the DC/AC signal blocks DC and passes AC (this AC is making your diagramps vibrate and accepted as making music).

Bert

And what if the cap lets your amps response down to say 5Hz? Why do you think the only danger is pure DC ?
Again we must see capacitors as DC coupling ( and filters ) for some reasons of schematic and topologies , not only to block dc and let music play .


Stefano

[BertD]

And what if the cap lets your amps response down to say 5Hz? Why do you think the only danger is pure DC ?

What has this to do with protecting the woofer from DC?

5Hz is still AC (5 times per second) and not just leaning in one direction (DC) making no movement at all...

If your speakers can't handle 5Hz is something for a complete other topic and therefore not on topic in this topic...or even at this forum.



Bert

[stefanobilliani]

well probably this is off topic , and I pass this . But , what have to do dc coupling or blocking or the woofer being capable of the 5Hz response with the PA feature ?


Stefano

[Scroobius]

Bert - servo and Duelunds nice

Stefano -

and maybe the enourmous gain of it looks toward wild instability

There is no instability in these GC's they have been rock solid. I have been using them dc coupled for a long time with 34db gain - no problems whatsoever and sounding wonderful. No instability at all and there is no reason for them to be unstable (well there could be but not here). The point is simply that the GC's amplify DC so put dc in and you get big dc out. That is all.

What is good though is that my prejudice against caps in the signal path seems to have been misplaced. But if caps are needed they have to be a bit special maybe.

P

[stefanobilliani]

Sounds good to me 

[BertD]

well probably this is off topic , and I pass this . But , what have to do dc coupling or blocking or the woofer being capable of the 5Hz response with the PA feature ?

With all respect, I did not bring up the 5Hz response, I guess you have to ask this to yourself? I only tried to explain that your 5Hz is still AC and has nothing to do with DC.

Bert

[stefanobilliani]

No prob , I am not taking any offense .

;-)

[PeterSt]

I only tried to explain that your 5Hz is still AC and has nothing to do with DC.

Outside of the main discussion, I sure am able to concur with Stefano. My take :

If we look at the current side of things, the average current at 5Hz we may see will be zero. If only the meter is slow enough. Otherwise it will jump up and down (eh, 10 times per second). Is that bad ? IT CAN"T BE. Why ? because we just "use" our speakers like that. Okay, maybe not at 5Hz but let's say at 30Hz. And as far as they go low enough of course (notice that the current may still be visible at 5Hz, never mind nothing moves anymore - not sure here). So, 5Hz or 30Hz or 100Hz, no difference. The difference should be in the cooling though. So, the higher the frequency, the more cooling.

Now suppose you feed the driver with DC (offset !). Cooling will be the same, but the average current will not. That will be higher. But :
Absolute current will be higher as well. That is, for that part the offset goes to.

Keep in mind : we are playing music. So this little story will never be about DC to the driver in rest (no AC is fed). This is related to the cooling again, which will be zero when we wouldn't play music.

What, net, remains is this :
In order to play withg offset as loud as we were used to without an offset applied, we will need to apply more current. For the peaks to the one side the offset goes to that is. Twice as much. So, all it comes down to is how much headroom the driver has with you playing on your comfortable max anyway. If that is on the (excursion) limits, no drop of DC is allowed. But if that is half of the (allowed !) excursion only, there is no problem. But the cooling problem remains. And you do NOT hear me say that things only get twice as hot when twice the current applies - this can well be much more.

But what it is about - and it will be here where I concur with Stefano, is that when a driver would be allowed to play 5Hz (no DC offset applied), all the time the diaphragm is at one side it will distort as much (but divided by 1.414 or so) as when the DC is applied and it plays "behind" all the time. For this we'd have to understand that no matter what lower frequency is playing, the higher ride on that so this is speaker driver behavior anyway when lower frequencies apply.

Nobody needs to agree with this, but this is what I think. And say.

Summarized, when a driver is capable of receiving say 20 Volts for allowed excursion, it can receive 10 volts for the excursion to one side only and nothing will distort more now if you only think in terms of 500Hz which needs to play nicely on a 30Hz "base" (wave). Peak current will be the same, but to one side only, BUT THE HEAT WILL BE MORE.

and now suddenly and with reason it is ok to predict that the dc offset of the amps should have same values to every channels , for those concerned about the maybe not neglectable 300mV .

Well Stefano, you said "with reason" so this reads a bit contradictionary to me. But anyway, Phasure is called Phasure for a reason and actually all what I do is about phase as such. You know, ALL we hear is about phase only. Localisation, recognition of voices and instruments ... I try to turn this the other way around so to speak; *Create* localisation in air from the source we call speakers. This is in the software and this is in the NOS1. It takes a few years thoug, but slowly I'm getting there.

I will announce it more explicitly on the forum, but next month Bert and I are going to a show, and there we will demonstrate what can be achieved when these aspects are taken to that other level. Half of that is software + DAC; the other half is the speaker. It will really be something different ...

Btw Stefano, my GainClones add nothing to the "distortion level" (THD) of the DAC but the gain (for noise floor). Officially though (measured (eh, by me) including noise) it is 0.004%. Not the 0.1% you read in the datasheet ...

Ok, this was my take. All in secrecy of course and as said, outside the main discussion.

Regards,
Peter

[BertD]

Simply because it will displace the coil out of its center alignment and that will give extra non-linear distortion on its own and it will also limit its max linear excursion.

Think about a 6mm high coil in a 9mm thick magnet gap...move it 1,5mm with DC (not that much DC is needed to do that!) and the driver will not be linear as soon as it starts to move when playing music.

Just my 2 cts.

Bert

[PeterSt]

Not to debate this Bert (of course not). But all what follows from it is that speaker drivers are not allowed to have any lower frequency applied to them. So all what I say is that since we allow for that, applying real DC is the same (but with all the side notes from my previous post).

The better way would be to use a driver with spades of allowed excursion which just is not used (like say your subwoofer appliance).

So this little part summarized : we agree, but then we shouldn't allow those low frequencies to pass. That is just temporal DC ...

Peter

[Scroobius]

So Peter "Phasure" did not come from your secret Government work after all then? he he

P

[BertD]

So this little part summarized : we agree, but then we shouldn't allow those low frequencies to pass. That is just temporal DC ...

Agreed but when is there 5Hz in the signal or even 10Hz. Perhaps rumble from a wobbly LP gives these ultra-low frequencies on a turntable but with digital sources?

Besides that, the level compared to higher frequencies is always a lot less in dB's and also highly reduced by most  speakers themselves (mechanically restricted by suspension or air load).

I do not see a problem with that theoretical 5Hz signal simply because it is not there or very low in gain; it will hurt less than using a capacitor trying to reduce these ultra-low frequencies (phase shifts already kicking in 4 octaves higher and ringing with very steep filters which should then be used anyway to be somewhat effective...).

Bert

[PeterSt]

Let's not forget that this is still outside the official discussion (eh, which was ? ).

So I too didn't make up the 5Hz, but I do say that what counts for 5 Hz counts as well for 30Hz. It is only "less DC" but the theoretical problem is the same. Only less severe.

Perhaps rumble from a wobbly LP gives these ultra-low frequencies

1.81Hz ? haha

phase shifts already kicking in 4 octaves higher

Yea, something like that. But when high passed at 0.5Hz that leaves 4Hz and above to be okay I think.


Now, what was this all about ?
When there is this source device (take the NOS1 for the example) there's very minor DC. Let's say it can't get much lower for any DAC. Practice denotes that 5mV is about most. But common knowlegde always tells that 20mV is allowed.

In the mean time we audio geeks think that DC blocking is not the best. So we don't when we have the chance.
And THIS is the wrong idea and it was nicely written down by Paul. So with his gain of 34 - some call that very high, so we take 20 now, this is DC of 20 x 5mV = 100mV. Is this bad ? I say no. Notice that this is static DC, so it is there without music playing;
Since this is not about the NOS1 at all, and 20mV is well accepted, we now talk about 20x20 = 400mV. This already starts to look less nice.

The new thing is (I think) that nobody says that this 20mV should be there on both sides or otherwise it is way bad. So, this is what I now say and claim. For fun now we use +20mV and -20mV which will be as accepted. But now we talk about +400mV vs -400mV. Well, unless you know what you are doing to this regard, this will be connected like that to the L/R speakers. One plays behind with a tension which meets nature somewat (a slam goes forward) and the other works against nature (the slam is breaked by the tension of the diaphragm which already is in the wrong direction).

Do we expect good sound from this ?
I don't.

The story from practice is next that something like GC chips "produce" their own gain (but theoretically every amplifier can exhibit this) and this too isn't necessarily equal for L/R. This is to be added to the DC already there incurred by the input. And do it gets worse again.

If DC is blocked at the input then at least that part is eliminated, and it is the part which can be gained. Of course, 0 remains 0.

And so I say that DC coupled amplifiers are not allowed, unless you know that the source blocks DC.

It is not said that putting in some nice capacitor of the right value (to begin with) has a net better result. But think :
Suppose the phase anomalies can't be prevented (say it exhibits up to 200Hz), what do you think is better ... the same anomaly Left and Right or that other anomaly which will cancel out frequencies in unpredictable fashion because it is out of control ? I am sure I choose for the former.

It is way more complicated than this alone because most solutions (to random audio problems) compare like apples and oranges. The above is an example of that. One problem is eliminated, but another comes in place of that, and that other exhibits totally different. When this is carefully worked out, we can still choose for the best solution. And this is not a subjective thing. So, no way you'd want different phase behavior from left and right, or worse from bass vs. mid. Only if you subjectively (!) like spatious "music" you'd choose for the L/R difference. But this is because you know listen, and THAT is subject to ... subjectivity.

Once the solution is chosen to block DC, we can always work on the solution that exhibits the best phase behavior. And when this is not my 0.5Hz high pass (effectively 4Hz or so), then work as long as necessary to get it done anyway.
or
Well, you will see. Later (with regards to Paul ).

Peter

[Scroobius]

Hey Peter

The story from practice is next that something like GC chips "produce" their own gain (but theoretically every amplifier can exhibit this)

Did you mean "dc offset" rather than "gain"?

P

[PeterSt]

Yes I did !

Apologies ...