No, "heard of them but never tried them; actually I never owned headphones, could stand more than two songs, drove me crazy. Getting more used to it since I got an Ipod for my birthday; ordered a pair of HD380 from Juno rcords for 118 including shipping see that goes..
Also Cross feed delay lines also seem to have been around since the late 50's, but I guess everyone's head size is different though & hard to tune. Also stuff a couple of microphones in a dummy and evidently you come close to realism.
I brought this up because headphones are super efficient ( easy to"amplify") but still "realism" seems even more remote. I am pretty good at DIY speakers
and definitely prefer that format for real enjoyment
Hi,
In the case where I encountered it, (big round of amp swapping out of quite collection) it stuck out enough (and in a positive sense) for me to ask what it was. Turned out it was Goldmund... Didn't know what it was and it was by far the puniest amp in the collection.
As you suggest it is as best average sonically, I take it you have heard this Amplifier and compared it to other high end amplifiers?
Ciao T
In the case where I encountered it, (big round of amp swapping out of quite collection) it stuck out enough (and in a positive sense) for me to ask what it was. Turned out it was Goldmund... Didn't know what it was and it was by far the puniest amp in the collection.
As you suggest it is as best average sonically, I take it you have heard this Amplifier and compared it to other high end amplifiers?
Ciao T
Smyth SVS Realizer is an interesting work around for headphone listening
you "calibrate" the system setting in the sweet spot of a real room, loudspeaker setup with mics in your ears - the DSP measures your hrtf and the loudspeaker+room frequency response - has head angle tracking
then corrects your headphones
the result in a very convincing "out of the head" reproduction of the loudspeaker&room "sound" over the headphones
Tvr, the output inductor can be looked at another way. First, there is a series resistor between 1 and 10 ohms. It is this resistor that protects the amp. The inductor is added to the degree necessary that you get unconditional stability with EVERY cap load within the realm of reason OR 1000pf-10uF. Usually 2uH is enough. It was enough for the Crown DC300, for example. In recent years, designers have tried to reduce the amount of inductance necessary, sometimes down to 1/2 uH. OR some manufacturers have eliminated it completely, and still have relatively stable designs with virtually any practical cap load.
Or, maybe establish first which topologies do actually sound diferent from others. You can have a well-executed SE versus a sloppy NFB SS give one result, and a sloppy SE versus a well executed NFB SS another result.
First things first: do different, well-executed topologies reliably and repeatably sound different?
jan
Hi,
In order to answer this question we need to define "well-executed".
This means we need to define objective criteria by which to measure the execution to determine if it is well executed or ill executed.
I think we all agree that the selection of such criteria should be evidence driven and not reflect various marketing campaigns over time.
The promotion of low measured distortion, of high damping factors and many other features in amplifier design have been marketing driven, since the day's when Harold J. Leak decided to outdo his competitors by rating his equipment at 0.1% Distortion and to heavily promote his gear as low distortion using the "point one" moniker.
In fact his equipment used what was at the time totally generic industry standard circuitry and parts and did not offer actually lower distortion then the competition using the same bog standard circuitry and parts, when the measurements where normalised.
At this point I find that there is little if anything evidence driven in the various measurements that are used to characterise Amplifiers (and other audio electronics).
In fact the situation mirrors to mind often the use of the ducking stool by Captain Tobias Slater in the "Scream satan scream" segment of "Dr. Terrible's House of Horrible" - namely when questioned on the utility of using the ducking stool to establish if a certain comely lass was in a fact a witch or not he replied:
"Charlotte: But I'm damned if I do and I'm damned if I don't. That doesn't make any sense.
Captain Tobias Slater: Yes... But we've got this ducking stool."
Ciao T
In order to answer this question we need to define "well-executed".
This means we need to define objective criteria by which to measure the execution to determine if it is well executed or ill executed.
I think we all agree that the selection of such criteria should be evidence driven and not reflect various marketing campaigns over time.
The promotion of low measured distortion, of high damping factors and many other features in amplifier design have been marketing driven, since the day's when Harold J. Leak decided to outdo his competitors by rating his equipment at 0.1% Distortion and to heavily promote his gear as low distortion using the "point one" moniker.
In fact his equipment used what was at the time totally generic industry standard circuitry and parts and did not offer actually lower distortion then the competition using the same bog standard circuitry and parts, when the measurements where normalised.
At this point I find that there is little if anything evidence driven in the various measurements that are used to characterise Amplifiers (and other audio electronics).
In fact the situation mirrors to mind often the use of the ducking stool by Captain Tobias Slater in the "Scream satan scream" segment of "Dr. Terrible's House of Horrible" - namely when questioned on the utility of using the ducking stool to establish if a certain comely lass was in a fact a witch or not he replied:
"Charlotte: But I'm damned if I do and I'm damned if I don't. That doesn't make any sense.
Captain Tobias Slater: Yes... But we've got this ducking stool."
Ciao T
Or just get two amps (for example) that sound different in a reliable and repeatable way, this should not be a great problem.
THEN do listening tests to get an agreed position in what way they sound different, which might be more difficult.
Then try to find a technical reason for that, change it and repeat the listening tests.
It should all start and end with listening of course.
But it's a bit like asking Yves Saint Laurent to do a controlled catwalk with different dresses to find out what it is exactly what the customers like in his designs....
jan
Last edited:
well, my cochlea, basilar membrane and all neurons and nerves involved have no idea whatsoever what class A, AB or D is, nor do they care about transistors or tubes. they seem to only care about the way some oxygen and nitrogen molecules move.
certain people keep saying that the classical distortion/freq response/etc measurements tell us nothing or close to nothing but I haven't seen anyone coming up with a different set of measurements.
I can't help wondering how come it's 2012 and we have no idea.
I'm perfectly willing to accept for instance that SS could never compete with tubes (just a random example) as long as I see some serious research documenting specific features of the generated sound wave that are related to human perception. but mostly I see speculation.
also, some keep saying that formal analysis of what makes good sound has no point because tastes vary. that doesn't make too much sense, especially when the same people assert with great authority what is bad gear and what is good gear.
I believe that even subjectivity can be measured to a point. (it's one thing to say "I like this amp more" while "I like this amp more because..." is a whole different story)
what I find kind of sad is that there always has to be this dichotomy: subjectivists vs objectivists. it's always divided.
subjectivists says: these measurements are useless.
objectivist says: I'll just keep playing with these numbers, they must lead somewhere because it's all math.
both are wrong, obviously.
certain people keep saying that the classical distortion/freq response/etc measurements tell us nothing or close to nothing but I haven't seen anyone coming up with a different set of measurements.
I can't help wondering how come it's 2012 and we have no idea.
I'm perfectly willing to accept for instance that SS could never compete with tubes (just a random example) as long as I see some serious research documenting specific features of the generated sound wave that are related to human perception. but mostly I see speculation.
also, some keep saying that formal analysis of what makes good sound has no point because tastes vary. that doesn't make too much sense, especially when the same people assert with great authority what is bad gear and what is good gear.
I believe that even subjectivity can be measured to a point. (it's one thing to say "I like this amp more" while "I like this amp more because..." is a whole different story)
what I find kind of sad is that there always has to be this dichotomy: subjectivists vs objectivists. it's always divided.
subjectivists says: these measurements are useless.
objectivist says: I'll just keep playing with these numbers, they must lead somewhere because it's all math.
both are wrong, obviously.
Just think of it as a rough approximation to a diplexer, known as a crossover in the audio world. The idea is that at higher frequencies the amplifier does not see whatever impedance the speaker+cable presents, but instead sees the resistor in the Zobel network. We are not trying to "cancel driver inductance", just hide whatever reactance it has from the amp at high frequencies where the NFB loop has enough to do already to maintain stability.tvrgeek said:
Ah, so I might assume the 6mH is a result of the 120 having no source resistors?
OK, I am looking at the PN2222A Fairchild spec sheet. I am trying to determine what eureka I should be having. It would seem that the designer intended this part to be used with about 10V EC and at about 10 to 20 mA. The bias spreader at idle has about 2.53 volts EC and if I can believe the charts with a BE of .61V, about .06mA going through it. Not exactly where it looks they had in mind it being used. What else am I to take away from this?
OK, I am looking at the PN2222A Fairchild spec sheet. I am trying to determine what eureka I should be having. It would seem that the designer intended this part to be used with about 10V EC and at about 10 to 20 mA. The bias spreader at idle has about 2.53 volts EC and if I can believe the charts with a BE of .61V, about .06mA going through it. Not exactly where it looks they had in mind it being used. What else am I to take away from this?
Unless i am totally missing the point, you are not using the curves / parameters correctly.
A 2222 is actually generally used as a switch, not in linear mode, but it is such a good workhorse it does well in both.
The point about the unusual number of gain tests is to show the wide range of Hfe over Ic. The only guaranteed spread is 150 to 300 at 150mA. But this is a digression.
Nowhere does one FORCE a Vbe to get an Ic! Exponential relationship. OK in a SIM!
"The bias spreader at idle has about 2.53 volts EC and if I can believe the charts with a BE of .61V, about .06mA going through it. Not exactly where it looks they had in mind it being used. What else am I to take away from this?"
You need to work on how to understand the circuit? Very little to do with the actual 2N2222 parameters.
No Cliff, you are not missing the point. John is making the point I need to understand the data sheet better. First pass. Clearly I have not had the eureka moment.
This is the sheet specifically.
http://www.fairchildsemi.com/ds/PN/PN2222A.pdf
So, which graph is it that I am not getting the ramifications of?
The circuit as I think I understand it: R8, 9, P1 and Q8 attempt to maintain a constant DC voltage difference between the bases of Q10 and 11. CCS Q9 providing the bias current. The majority of this current passing through R9, 8 and very little through Q8 with the voltages as supplied in the manual for idle. The voltages with reference to the rails would move following the signal, the amplitude of which is the result of Q6-Q7.
R7? Confused being in series with P2, @500 Ohms. I don't actually see the reason it is there unless to limit missadjustment of P2. R7 +P2 with Q9 etc is the CCS. I am missing more than something. ( besides D4-7 and why the current is so high through the entire D1-9) Let me go off and try and write down all the currents in this stage based on the assembly manual. Not having one running is a hinderence.
This is the sheet specifically.
http://www.fairchildsemi.com/ds/PN/PN2222A.pdf
So, which graph is it that I am not getting the ramifications of?
The circuit as I think I understand it: R8, 9, P1 and Q8 attempt to maintain a constant DC voltage difference between the bases of Q10 and 11. CCS Q9 providing the bias current. The majority of this current passing through R9, 8 and very little through Q8 with the voltages as supplied in the manual for idle. The voltages with reference to the rails would move following the signal, the amplitude of which is the result of Q6-Q7.
R7? Confused being in series with P2, @500 Ohms. I don't actually see the reason it is there unless to limit missadjustment of P2. R7 +P2 with Q9 etc is the CCS. I am missing more than something. ( besides D4-7 and why the current is so high through the entire D1-9) Let me go off and try and write down all the currents in this stage based on the assembly manual. Not having one running is a hinderence.
I am confused by this and don't really understand what you are saying.
My understanding of how things work is radically different, so I'll bow out and leave you to it.
Cliff
Real world aside, looking at the Hafler voltage chart, I sure as heck hope there is not 3 tenths of a volt difference between the base of Q3 and Q4. Considering the vintage, I guess they could have used a Simpson 260 to measure them! We take out Flukes for granted now. No more calculating for the meter loading.
- Status
- Not open for further replies.