Well, I do, and it's correct (both theoretical and practical). A 96dB speaker sounds twice as loud as a 90dB speaker, at the same power. And so is hiss. This is obvious even for those with tin ears like myself
You'd be very happy with BF861 cooled down via a Peltier device. My preliminary results are very good, I can tell that. Evacuating the condensation is not trivial, though.
Logically I would have thought independently of the loudspeaker sensitivity we have to tune our little volume knob to get the same listening level, which would equalize out any difference in phono stage noise or anything else before the volume knob, right?
The only thing we can't equalize out is the noise produced after the volume knob, the noise in the power amplifier for instance.
btw the human ear perceives a sound pressure doubling closer to around 9 dB, don't remember right now the exact figures but I am sure Joachim knows that too.
Cheers Michael
Syn and I talk about how much it can be perceived near a speaker at full open system gain, so the speaker's sensitivity matters. Sort of a tell tale crude test of overall hiss. The head amp would be the most critical chain ring in such a case given all else is done right.
You talk hiss at a given arbitrary SPL as controlled by the volume pot for a range of speakers drivable. In that case, of course you are right, only rough HF peaks in some horn speakers would inflict a disadvantage and not their mean average sensitivity. I would add the defined directivity control on top of that, focusing more HF energy around the listening position instead of a more ''spread out'' power response by a dome tweeter in usual listening environments.
I have 95dB direct radiating speakers for instance. With only one 2SK170 in the head amp, still the hiss is only audible 10cm from the tweeter when I set the volume for about 90dB average at listening position.
Beyond the nuisance factor which is controllable and relative as in the above speak, I am sure that noise climbing up in the lows with Fets and triodes when collaborating with under 1Kohm source resistance (i.e. most MC), and especially under 100 Ohm, softens the bass slam. I got to use 4 in parallel soon that will help by sqrt n. BJTs are better with MC low but I don't mind much beyond 4 2SKs since the rest of my system is valve and the contributed hiss after the phono is not best, will rather dominate, but still adequate for my priorities.
ok, salas i understand now. but it is a party trick. we have a lot of customers like that in germany. they crank the sytem full and then complain that they hear noise then they put their ear near the tweeter. i just wanted to make shure that we understand each other and you made it clear. yes, syn, sell your stuff here. our bürocratic customers will love it. Accuphase is very popular here and i think that is one reason they took so much pain in eliminating the noise. well actually i have not heard or measured it and over 90dB THD/Noise seems a little odd to me when i see the specification. compared to what level? -60dB(1mV) or what.
to michael: yes it is that book
to john: that 0.8 Ohm and 0,35mV is a great combination for very high dynamics. they must use enormous magnets or very innovative solutions. is is a field coil design with DC feed?
to Bas: tomorrow in the pane i will study that circuit, i hope i understand it. never seen something like that and how will it sound ?
To all: have you heard about a therory that a little noise, distortion and DC offset helps a speaker to sound better because it is quite stiff and unlinear in the rest position. the dc will push it a bit out, the second harmonic will mask higher ones and the noise contibutes a dithering action that softens the suspension even further, makes it more "responsive" and masks higher harmonics too ?
to michael: yes it is that book
to john: that 0.8 Ohm and 0,35mV is a great combination for very high dynamics. they must use enormous magnets or very innovative solutions. is is a field coil design with DC feed?
to Bas: tomorrow in the pane i will study that circuit, i hope i understand it. never seen something like that and how will it sound ?
To all: have you heard about a therory that a little noise, distortion and DC offset helps a speaker to sound better because it is quite stiff and unlinear in the rest position. the dc will push it a bit out, the second harmonic will mask higher ones and the noise contibutes a dithering action that softens the suspension even further, makes it more "responsive" and masks higher harmonics too ?
The incremental transfer function can easily be measured especially very near field with a sensitive mic. A dead zone would give crossover-like distortion something rarely(never) seen in any measurements that I have made. Due to reciprocity you could probably do just electrical measurements to see any effect.
Some more explanation:
BL is the magnet strength and very often does not have a maximum in the rest position.
Le is inductance. it varies over the movement of the coil.
Cms is the stiffness of the suspension and should be minimal in rest position. that is often not the case and that yellow "spider" you see is a big problem.
what happens is the following: the cone like to be where BL is maximum, Le is maximum and stiffness is minimum. so what happens in a real speaker that a DC shift is happening that is frequency dependent around the resonace Fs. Asymmetric problems produce 2nd harmonic and symmetric problems produce 3rd harmonic. that are the main sources of distortion in loudspeakers not considering rubb and buzz that create much higher harmonics. i whould not call that crossover distortion but it is a dynamic problem with hysteresis that creates all kinds of frequency and volume dependent intermodulation products. you can learn how to measure that on the Klippel GmbH - Home Page website. we tried to get a handle on that problems with DSP for 20 years have succeded so far more with low tech mechanical improvements. a good modern speaker
at medium volume ( 85-95dB) has about -50dB distortion over a 60dB dynamic range to put that in perspective.
BL is the magnet strength and very often does not have a maximum in the rest position.
Le is inductance. it varies over the movement of the coil.
Cms is the stiffness of the suspension and should be minimal in rest position. that is often not the case and that yellow "spider" you see is a big problem.
what happens is the following: the cone like to be where BL is maximum, Le is maximum and stiffness is minimum. so what happens in a real speaker that a DC shift is happening that is frequency dependent around the resonace Fs. Asymmetric problems produce 2nd harmonic and symmetric problems produce 3rd harmonic. that are the main sources of distortion in loudspeakers not considering rubb and buzz that create much higher harmonics. i whould not call that crossover distortion but it is a dynamic problem with hysteresis that creates all kinds of frequency and volume dependent intermodulation products. you can learn how to measure that on the Klippel GmbH - Home Page website. we tried to get a handle on that problems with DSP for 20 years have succeded so far more with low tech mechanical improvements. a good modern speaker
at medium volume ( 85-95dB) has about -50dB distortion over a 60dB dynamic range to put that in perspective.
I just confirmed with a simple sim that the circuit in #217 or one very similar to it yields a null of both offset and second harmonic with two trims for very unmatched JFET's.
On speakers I was only stating that in my experience the DC transfer function of displacement vs current is continuous and fairly smooth. Systems with second order nonlinearities will typically exhibit DC shift when excited.
On speakers I was only stating that in my experience the DC transfer function of displacement vs current is continuous and fairly smooth. Systems with second order nonlinearities will typically exhibit DC shift when excited.
hi scott!
what circuit are yot refering too? in post 217 i made a link to another guys stage that has no fets. i asume it was one of my circuits and i whould like to know which. what was the dissimilarity to mine you mention ?
Anyway, thank you very much for simulating the circuit. i just start to learn that and have not much experience. of cause i do a lot of simulation with loudspeakers.
what circuit are yot refering too? in post 217 i made a link to another guys stage that has no fets. i asume it was one of my circuits and i whould like to know which. what was the dissimilarity to mine you mention ?
Anyway, thank you very much for simulating the circuit. i just start to learn that and have not much experience. of cause i do a lot of simulation with loudspeakers.
Sorry meant #201. I just used resistive loads to the rails not current sources and only two FET's with 50 Ohm source resistors to ground through 10 Ohms so there are only 4 active devices. -100db @ 2v p-p out
A few Ohms unbalance in the 50 Ohm resistors cancels most of the second and a few 10's of Ohms in the other resistors balances the offset.
Matching j74's to k170's is a thankless task anyway.
A few Ohms unbalance in the 50 Ohm resistors cancels most of the second and a few 10's of Ohms in the other resistors balances the offset.
Matching j74's to k170's is a thankless task anyway.
I saw an interesting RIAA stage in the december 2009 issue of Audioxpress. because we have discussed the possibility to use coils for equalization in this thread and for compleetness
i publish the circuit here with default values (eccept the coil) because i do not know if i am allowed to publish the circuit with the correct values. the author makes a good point for using a coil in the first stage and it has a resonable value of 25mH and some ohms so it should be possible to find that coil or make one yourself. the design does the second breakpoint (500Hz) passive and the third breakpoint (2122Hz) active in the output stage. i whould try it also the other way around with 75usec passive and the 500Hz active but one would end up with a higher value capacitor in the feedback loop of the output opamp. i can recommend that Issue because it also has an interesting interview with Professor Malkolm Hawsford in it. I worked with Malkolm a lot in the 90th and he is one of my most important influences. i do not know if the digital version of Audioxpress is yet available but i know from the editorial that they are working on it.
i publish the circuit here with default values (eccept the coil) because i do not know if i am allowed to publish the circuit with the correct values. the author makes a good point for using a coil in the first stage and it has a resonable value of 25mH and some ohms so it should be possible to find that coil or make one yourself. the design does the second breakpoint (500Hz) passive and the third breakpoint (2122Hz) active in the output stage. i whould try it also the other way around with 75usec passive and the 500Hz active but one would end up with a higher value capacitor in the feedback loop of the output opamp. i can recommend that Issue because it also has an interesting interview with Professor Malkolm Hawsford in it. I worked with Malkolm a lot in the 90th and he is one of my most important influences. i do not know if the digital version of Audioxpress is yet available but i know from the editorial that they are working on it.