Home |
Search |
Today's Posts |
#1
Posted to rec.audio.tubes
|
|||
|
|||
amplifier input sensitivity
Hello,
where is the point in making the input sensitivity of a modern amplifier as low as 250mv ? Lets assume the amplifier needs 0,775Vrms (=0dBu) for full output and most digital sources tend to produce about 2Vrms leaving a 8dB margin for quieter recordings. I found a Marantz Fmtuner with 1Vrms Output and i think other tuners will bei similar. This will give only about 2dB margin, but since most fm broadcasts are modulated noise anyway there should be no problem in achieving full volume. High quality reel-reel tapedecks will have balanced XLR connectors giving +4dBu which equals to 1,2Vrms. So again no reason for low input sensitivity. Riaastages might be different, but i think that a high quality preamplifier should be able to deliver 0dBu even with strangely mastered records. Can anybody here think of any (high quality) source that will actually need such high sensivity ? sincerly, Alexander Dyszewski |
#2
Posted to rec.audio.tubes
|
|||
|
|||
amplifier input sensitivity
"Alexander Dyszewski" where is the point in making the input sensitivity of a modern amplifier as low as 250mv ? ** Plenty of audio signal sources deliver only that much voltage. ( snip **** ) I found a Marantz Fm tuner with 1Vrms Output ** So ****ing what ? You are cherry picking examples to suit yourself. and i think other tuners will bei similar. ** Fraid YOU do not think anything at all. High quality reel-reel tapedecks ** Huh ?? Utterly obsolete in a domestic sound system. Can anybody here think of any (high quality) source that will actually need such high sensivity ? ** Nearly all AM /FM tuners, nearly all cassette decks, most TV sets with AV outputs, and nearly all phono stages. **** Off. ..... Phil |
#3
Posted to rec.audio.tubes
|
|||
|
|||
amplifier input sensitivity
Alexander Dyszewski wrote:
where is the point in making the input sensitivity of a modern amplifier as low as 250mv ? Lets assume the amplifier needs 0,775Vrms (=0dBu) for full output and most digital sources tend to produce about 2Vrms leaving a 8dB margin for quieter recordings. I found a Marantz Fmtuner with 1Vrms Output and i think other tuners will bei similar. This will give only about 2dB margin, but since most fm broadcasts are modulated noise anyway there should be no problem in achieving full volume. High quality reel-reel tapedecks will have balanced XLR connectors giving +4dBu which equals to 1,2Vrms. So again no reason for low input sensitivity. Riaastages might be different, but i think that a high quality preamplifier should be able to deliver 0dBu even with strangely mastered records. Can anybody here think of any (high quality) source that will actually need such high sensivity ? If you're building for yourself, make whatever fits best in your system. Commercial amps are different. Partly for the obvious reason, and partly because the average consumer expects full power to be about 1 o'clock on the volume control. So I read somewhere, AFAIR...possibly Menno van der Veen or Morgan Jones. Ian |
#4
Posted to rec.audio.tubes
|
|||
|
|||
amplifier input sensitivity
flipper schrieb:
where is the point in making the input sensitivity of a modern amplifier as low as 250mv ? I don't know that I'd pick 250mV but the problem is things are not as precise as you're making out. Lets assume the amplifier needs 0,775Vrms (=0dBu) for full output and most digital sources tend to produce about 2Vrms I think you're already in trouble making that assumption as 1Vrms, which you mention below, is also rather common. Is the goal to work with something specific you already have or cover the range of what's available? If the latter then you don't have the luxury of 'most sources' even if you're right about what's common. I want to cover a reasonable range of possible sources like "normal" japanese hifi components. leaving a 8dB margin for quieter recordings. Here you have a *wide* range, depending on what you listen to. First, the '2Vrms' (1Vrms) is digital speak for full scale output, not nominal. There is nothing, not so much as one pico volt, more than FS and that means someone trying to do 'hi quality' is not going to approach full scale because of clipping. If I remember correctly, peaks (true peak meter) 6dB under FS is the recommended 'safe' setting and if it's a 1Vfso device that's already down to 500mV peaks. 'Modern' rock will compress the bejesus out of things to compete in the volume wars, so nominal will not be low. Classical, however, to keep dynamic range, will have nominal depressed along with staying away from full scale. Early rock recordings, from back in the days when they thought CD was supposed to be 'high quality', will be closer to the 'classical' level because it isn't compressed flat as a screaming cow pattie up against FS. I disregarded classical music, because most of my cds are metal, rock and electronic. But you do have a point there, some cds do have quite a lot of dynamic range. On top of that you have listener preference and you seem to be assuming it's to never drive into clipping regardless of what nominal volume level that dictates but some (many?) will turn it up till it hurts, which is clipping, so 'max power' is not necessarily your 'peak' point. This, btw, is especially true of tube gear because they (usually) clip more gracefully than SS. Then, silly as it may seem, there's the volume knob itself and people don't like it pegged on max because that feels like it's 'run out of volume', never mind whether it's usable. I found a Marantz Fmtuner with 1Vrms Output and i think other tuners will bei similar. This will give only about 2dB margin, but since most fm broadcasts are modulated noise anyway there should be no problem in achieving full volume. I have no idea what your reasoning there is. 'Noise' isn't going to help it reach max power. My idea was that since fm-broadcasts tend to be extremely compressed and enchanced for kitchen radios there is no need for additional gain in the preamplifier. High quality reel-reel tapedecks will have balanced XLR connectors giving +4dBu which equals to 1,2Vrms. So again no reason for low input sensitivity. Riaastages might be different, but i think that a high quality preamplifier should be able to deliver 0dBu even with strangely mastered records. Consumer line level is 316mVrms nominal. Can anybody here think of any (high quality) source that will actually need such high sensivity ? Thank you for the very informative post. |
#5
Posted to rec.audio.tubes
|
|||
|
|||
amplifier input sensitivity
On Apr 20, 10:20*am, "Phil Allison" wrote:
"Alexander Dyszewski" Can anybody here think of any (high quality) source that will actually need such high sensivity ? ** Nearly all AM /FM tuners, nearly all cassette decks, most TV sets with AV outputs, and nearly all phono stages. * ***** Off. .... *Phil You'll have to excuse Phil. He's an autistic ****. 775 mV sensitivity at any actual impedance between 10 K and 1 M is "about right" though a little higher or lower is fine. High input impedance is more important if driving directly from a CD player, for instance, since the manufacturers of these refuse to put in adequate output sections and certainly adequate volume controls. The original MC275, which is the canonical stereo amplifier IMO, sensibly has a knob but there are separate ones for each input, which is awkward. |
#6
Posted to rec.audio.tubes
|
|||
|
|||
amplifier input sensitivity
On Apr 20, 9:01*pm, Bret L wrote:
On Apr 20, 10:20*am, "Phil Allison" wrote: "Alexander Dyszewski" Can anybody here think of any (high quality) source that will actually need such high sensivity ? ** Nearly all AM /FM tuners, nearly all cassette decks, most TV sets with AV outputs, and nearly all phono stages. * ***** Off. .... *Phil *You'll have to excuse Phil. He's an autistic ****. 775 mV sensitivity at any actual impedance between 10 K and 1 M is "about right" though a little higher or lower is fine. *High input impedance is more important if driving directly from a CD player, for instance, since the manufacturers of these refuse to put in adequate output sections and certainly adequate volume controls. The original MC275, which is the canonical stereo amplifier IMO, sensibly has a knob but there are separate ones for each input, which is awkward. Line level is not standard among consumer grade audio. It is true, 316mV is very common,but the input (as well output ) voltage is not clear (meaningless) unless impedance is specified. 0dB is actually ~775mV at 600 Ohm in telecom and idustrial /professional audio environment. If the impedance (Z) is not 600 Ohm any voltage level stated becomes rather ambiguous and quite misleading. 0DB(standard) is 1mW = V^2 / Z. |
#7
Posted to rec.audio.tubes
|
|||
|
|||
amplifier input sensitivity
Most consumer audio amplifiers, and in fact most in general (EXCEPT
the "booster" amps of yore such as large Altecs which were nothing more than a single pair of medium mu transmitting triodes with an array of transformers for power supply, input and output) are bridging input loads. They are not designed to accept input power, just voltage. If one needs to drive them from a source which wants a defined load then one simply puts a load resistor across the input. There is no particular reason for the input sensitivity of these to be what they are, it's simple approximate consensus commonality. 775 mV is "about right" because most any source will drive it to full output except a raw phono cartridge or a guitar. If needed you can pad it down, which is easier than the other way. Manufacturers should be URGED to provide convenient internal tie points to add a voltage divider resistor pair if needed in a sanitary fashion. |
#8
Posted to rec.audio.tubes
|
|||
|
|||
amplifier input sensitivity
Consumer grade audio manufacturers build whatever they
think is right and just leave it to the public to find out (by trial and error) what the true signal levels in their gadjets are... Yes. Most of the better test sets have multiple output impedances today: standalone audio gens aren't that common in current use. The pro world has been utterly dominated by AP for about a decade, such that you can get HP 339s, 8903s, and even the grossly overpriced Potomac Instruments boxes cheap. |
#9
Posted to rec.audio.tubes
|
|||
|
|||
amplifier input sensitivity
Actually, the decibel, originated from loss of power over 1 mile of standard telephone cable defined as "a cable having uniformly distributed resistances of 88 ohms per loop mile and uniformly distributed shunt capacitance of .054 microfarad per mile," or 1 MSC.- Hide quoted text - - Show quoted text - Very good info! However, if the impedances are different the source voltage will not remain the same as soon as it is hooked up to the load ! In fact, it will drop significantly. If we take an audio signal generator (and most of them have nominally 600 Ohm output impedance) and set it at 0.775V level (open circuit). Now let us connect a load (resistor of some other value,say 75) across and measure the output voltage. It will NOT be 0.775V any longer: in an average piece of test equipment this level WILL drop if the load is anything DIFFERENT than source impedance. If the load is higher than 600 it may not be as pronounced. Even oscilloscope input Z is rated (typically 1 MOhm or higher) to be aware of possible misleading measurements. This is real life , not pure theory from textbooks.To keep the output level constant at different loads will require a constant voltage source and is normally not found in a typical audio generator. Consumer grade audio manufacturers build whatever they think is right and just leave it to the public to find out (by trial and error) what the true signal levels in their gadjets are... ***For several related reasons, it's usual to ensure that input impedance is at least 10 times source impedance. With that ratio, the signal is reduced by about 9% from its open circuit voltage. Just a rule of thumb, but useful guide to the limit of what can be considered a voltage source for the purpose of audio. ***The ratio used for efficient power transfer would be 1:1, presumably. Flipper is quite right: you're hung up on rules that apply to different circumstances, and for different purposes. ***What would be an average output impedance for a modern source, and how much variation is there, actually rather than in a suspicious mind? Similarly for amplifier input impedance? I don't think you'll find much cause for concern. For valve amps, much less so. For valve sources into SS amps, you might need to be more careful. ***The decibel is a ratio of two quantities. Since the units of the two quantities must be the same, the decibel ends up with no unit of its own. Some argue that a unit remains implicit even when it's cancelled out, but that makes no sense to my maths. It does make sense to claim that, by convention of language, it has a few particular applications, and therefore one of a few units is implied, depending on context. If it is not absolutely clear from the context, then the unit should be stated, as in "dBV". "dBm" is smart-arse jargon that stuck. ***Ian |
#10
Posted to rec.audio.tubes
|
|||
|
|||
amplifier input sensitivity
In article
, Bret L wrote: Most consumer audio amplifiers, and in fact most in general (EXCEPT the "booster" amps of yore such as large Altecs which were nothing more than a single pair of medium mu transmitting triodes with an array of transformers for power supply, input and output) are bridging input loads. They are not designed to accept input power, just voltage. If one needs to drive them from a source which wants a defined load then one simply puts a load resistor across the input. I don't agree, I remember some of those PA "booster" amplifiers were designed to accept input power, this was sometimes necessary because some of them used transmitting tubes that were driven into grid current, i.e. class B2. -- Regards, John Byrns Surf my web pages at, http://fmamradios.com/ |
#11
Posted to rec.audio.tubes
|
|||
|
|||
amplifier input sensitivity
"flipper" If the impedance (Z) is not 600 Ohm any voltage level stated becomes rather ambiguous and quite misleading. No it isn't. dBV is specified relative to 1V. I.E. 0dBV is 1V, 10dBV is 10V, etc ** Sorry - 10 dBV is 3.16 volts. dBu is specified relative to .775V (rounded). I.E. 0dBV is .775V (rounded), 10dBu is 7.746V, etc. ** Same error again - 10 dBu is 2.449 volts ..... Phil |
#12
Posted to rec.audio.tubes
|
|||
|
|||
amplifier input sensitivity
However, if the impedances are different the source voltage will not remain the same as soon as it is hooked up to the load ! In fact, it will drop significantly. ** Wot absurd drivel. If we take an audio signal generator (and most of them have nominally 600 Ohm output impedance) and set it at 0.775V level (open circuit). Now let us connect a load (resistor of some other value,say 75) across and measure the output voltage. It will NOT be 0.775V any longer: in an average piece of test equipment this level WILL drop if the load is anything DIFFERENT than source impedance. ** Totally off with the fairies. The voltage level will always DROP NO MATTER WHAT THE LOAD IMPEDANCE IS !! If the load is higher than 600 it may not be as pronounced. ** The bats in this loon's belfry are having a party. Consumer grade audio manufacturers build whatever they think is right and just leave it to the public to find out (by trial and error) what the true signal levels in their gadjets are... ** Total crap. The line voltage levels quoted by makers of consumer audio gear are based on a simple assumption - ie that the load impedance as specified or is WITHIN the range of values found in consumer audio gear, ie at least 10kohm. The source impedance values used are low enough that the output level is not significantly affected by such a load. This has been ** standard practice ** for over 60 years and it still amazes me how many fools like this posturing ASS have not caught up with it. ..... Phil |
#13
Posted to rec.audio.tubes
|
|||
|
|||
amplifier input sensitivity
On Apr 21, 11:46*pm, flipper wrote:
On Thu, 22 Apr 2010 13:18:40 +1000, "Phil Allison" wrote: "flipper" If the impedance (Z) *is not 600 Ohm any *voltage level stated becomes rather ambiguous and quite misleading. No it isn't. dBV is specified relative to 1V. I.E. 0dBV is 1V, 10dBV is 10V, etc ** Sorry *- * *10 dBV * is 3.16 volts. dBu is specified relative to .775V (rounded). I.E. 0dBV is .775V (rounded), 10dBu is 7.746V, etc. ** Same error again *- * 10 dBu is 2.449 volts Quite right. My bad. I swapped power ratios for voltage. .... * Phil- Hide quoted text - - Show quoted text -- Hide quoted text - - Show quoted text - Everybody must be happy at this stage! |
#14
Posted to rec.audio.tubes
|
|||
|
|||
amplifier input sensitivity
Alex wrote:
No it isn't. dBV is specified relative to 1V. I.E. 0dBV is 1V, 10dBV is 10V, etc. There is no 'ambiguity' nor is it 'misleading'. This is a mistake, by the way. +10dBV is approximately 3.1V. +20dBV is 10V. dBu is specified relative to .775V (rounded). I.E. 0dBV is .775V (rounded), 10dBu is 7.746V, etc. The same mistake. +20dBu corresponds to 7.75V. The dB has become a dog's dinner of a non unit that causes confusion wherever it goes. As a convention, it's perfectly clear once you've learned it, but it isn't obvious. In particular, if you wanted a confusing way of expressing the voltage necessary to increase the power by a certain proportion, given constant resistance, dBV would be a good candidate. Then having established that form of expression you could add the totally contrary dBm, for good measure. Isn't this the kind of thing that Europe was supposed to sort out? A more consistent convention for dealing with its subscripts...the u, V, m etc....would help maybe. On a positive note, I've noticed that many of those here who appear versed in audio technology are quite disciplined in their usage of the decibel, especially with respect to voltage. In circumstances where a voltage is significant on its own account, I generally see volts or percentages. Where voltage is significant by dint of its relation to power, I generally see dB. Not a habit I've fully adopted, I must admit. It would be nice, I think, if the same kind of distinction were made when talking about frequency. We don't use octaves as much as we ought, I fear. Ian |
#15
Posted to rec.audio.tubes
|
|||
|
|||
amplifier input sensitivity
Flipper wrote:
dBu is specified relative to .775V (rounded). I.E. 0dBV is .775V (rounded), 10dBu is 7.746V, etc. The same mistake. +20dBu corresponds to 7.75V. The dB has become a dog's dinner of a non unit that causes confusion wherever it goes. As a convention, it's perfectly clear once you've learned it, but it isn't obvious. In particular, if you wanted a confusing way of expressing the voltage necessary to increase the power by a certain proportion, given constant resistance, dBV would be a good candidate. For that purpose it doesn't matter what flavor of dB you use because a doubling of power is, by definition, 3dB, is 3dB, is 3dB. Well yes. But the simple ratio of voltages is not the same as the simple ratio of powers. Half the voltage is -6dB, half the power is -3dB. Sticking the V on the end of the dB doesn't give much of a clue that this will be the case, especially as there is no explicit indication that the dBV has anything to do with power. Surely you agree that many ppl find the dB at least a bit hard to grasp? Why do you want to demonstrate that it is straightforward? For example, when double checking FB on a power amp in spice I disconnect FB, measure output power, connect FB, measure output power, and verify it's 100 to 1 for 20db or 10 to 1 for 10dB. Which, btw, is how I made the stupid mistake. In spice I can measure power 'directly' and reflexively know 10dB is 10x power. A stupid mistake from not thinking about what I was typing. I don't quite follow what mistake you mean, I'm afraid. One thing that confused me for some time arises from the fact that my spice automatically labels the y-axis in terms of current or voltage, because there's no such thing as a power "probe". So if I multiply a current by a voltage, and graph the result, it's not obvious that the result is power, although, almost by fluke, it still gets the dB scale right in the sense that the numerical values are correct. Similarly with resistance, and even stuff like flux density, as long as I'm careful to stick with units from the same family. What 'nominal' or 'reference' power do I start with or 'calibrate' to? Nothing. It doesn't matter. 10dB is 10dB. The flavor only matters if you care about an interchange reference level like, say, building a repeater that you need to work in a 'standard' telecom system that expects 0dB to be 1mW into 600 Ohms. Proper units don't need all these ifs and buts. A metre or a second can be simply defined in a jiffy. Then having established that form of expression you could add the totally contrary dBm, for good measure. It isn't 'contrary' and I'm afraid you have it backwards. dBm (as simply dB, the "Bel" in honor of Alexander Graham Bell since it was telephone transmission the unit was created for) came first and it's arguable that dBV was to 'simplify' things like you opine about. I didn't mean to imply a historical sequence, but rather the order one might introduce the dB such as to highlight the confusion. It is contrary. In dBV, the V is a unit. In dBm, the m is not a unit. How can that possibly be consistent? In plain dB, there's no subsript at all. The rules of subscripting are different in each case. When people don't know about something they tend to think it's 'arbitrary' or 'stupid' but that's rarely the case. Like, who's the idiot that decided there should be so many '12s' in things? Well, I don't know who the 'idiot' was but the reason is that, back when most people we're mathematically challenged, 12 is divisible by 2, 3, and 4. It made life simpler. You invented the idiot part. Dozens are clearly useful, as are 60s and 360s. 10 is the unfortunate choice, but even then I'm sure everyone knows how that came about. I'm not saying the dB is arbitrary or stupid. On the contrary, it serves a purpose very well. It's the inconsistency of subsripting that's a dog's dinner, and that in combination with the absence of explicit reference to power (unavoidable I guess) that can make it hard to follow. Transforms are a bit tricky for folk to get their heads around, and a cluttering of historical debris doesn't make it easier for the novice. I tried to give a brief background on dBm for the same reason. Waaaay back when you have people building the first phone lines, and we all know the signal degrades over distance, we obviously need some way to express that and the first 'unit' was the MSC (Mile [of] Standard Cable). I.E. How much the signal degrades over a mile of the stuff we use. Makes sense if you think about it, at least for a starter. Now, as things progress it's not really wise to tie a 'standard' to a particular type of cable (we use multiple types now) so in the 1920's the "TU" (transmission unit) was defined (as ten times the base10 logarithm of the ratio of measured power to reference power) and, to reduce confusion, intentionally made it close to the MSC (=1.056 TU). The TU then became the decibel. Now, I don't know why 1mW was the preferred reference but I assure you there was some practical reason and, at any rate, back when the MSC was done there were no radars, hi-fi stereos, AM radio, or any other 'electronics' as we know it to bother with and no 'confusion' about what's being worked with. It was a telephone system, period. That was the problem being dealt with, you care about impedance matching, power transfer, and it works. As I previously mentioned, when you start adding electronic repeaters it makes perfectly good sense to use the dB(m) reference level because that's the system you're putting them into. And it makes perfect sense that, if one is going to make a general purpose amplifier, people used to making repeaters are likely to use the same terminology, not to mention the first ones are likely to be nothing more than the same repeaters used for a different purpose, and we know they expect 0dBm as reference, so that's likely to be the 'reference level' for whatever else you stick them in because why redesign the damn things? Of course, eventually you do because not all applications have the same requirements. And, again of course, after time and things have multiplied a million fold, so people are separated from the 'original reason', there's likely to be someone who figures things would be 'simpler' if we used '1' as a reference, instead of the 'arbitrary' and 'stupid' .775V, so we get dBV. And when we get to radar reflection we get another use for dB, and another letter. And then we have speakers, and another use for dB, and so on, because, as it turns out, power is power and the fundamental ratio applies but 1mW into 600 Ohms is simply not a useful reference for radar and speakers. Again, I can't tell you how 'everything' came about but I assure you there were practical reasons. Or, at the very least, someone thought there was a good practical reason and if it persisted it probably was. I haven't argued that the dB is not useful, or not valid. I'm trying to identify why ppl find it hard to follow. It is hard to follow. Ppl have big arguments about it and come from all sorts of directions. What's your explanation for that? Ohm's Law, for example, doesn't attract any debate at all. Neither do simple units like seconds or metres. Even logs don't attract the same level of confusion. Even you have just made a couple of errors, which you wouldn't with seconds, metres or logs. Isn't this the kind of thing that Europe was supposed to sort out? A more consistent convention for dealing with its subscripts...the u, V, m etc....would help maybe. What inconsistency are you claming? See above. In terms of self-consistency, V is an SI unit whereas m is a misplaced SI prefix and actually I don't know what the u stands for. Is it a corruption of a mu symbol? In that case it would be a misplaced corruption of an SI prefix. In terms of consistency with SI units, dBV relates to power but makes no mention of a unit of power, the m isn't a properly applicable prefix even if it wasn't misplaced, and the u isn't either, and it's corrupt to boot. On a positive note, I've noticed that many of those here who appear versed in audio technology are quite disciplined in their usage of the decibel, especially with respect to voltage. In circumstances where a voltage is significant on its own account, I generally see volts or percentages. Where voltage is significant by dint of its relation to power, I generally see dB. Not a habit I've fully adopted, I must admit. It would be nice, I think, if the same kind of distinction were made when talking about frequency. We don't use octaves as much as we ought, I fear. Like most things it depends on what is of interest. Talk to musicians and you'll get an earful about octaves (plus a lot more). And yet, when we are concerned about audible bandwidth, octaves would be linear with respect to perception, so there is just as much reason to use the octave in audio as the decibel. They both linearise an axis that would otherwise require a log scale to fit with perception, and therefore with significance. Ian |
#16
Posted to rec.audio.tubes
|
|||
|
|||
amplifier input sensitivity
On Apr 21, 9:43*pm, John Byrns wrote:
In article , *Bret L wrote: *Most consumer audio amplifiers, and in fact most in general (EXCEPT the "booster" amps of yore such as large Altecs which were nothing more than a single pair of medium mu transmitting triodes with an array of transformers for power supply, input and output) are bridging input loads. They are not designed to accept input power, just voltage. If one needs to drive them from a source which wants a defined load then one simply puts a load resistor across the input. I don't agree, I remember some of those PA "booster" amplifiers were designed to accept input power, this was sometimes necessary because some of them used transmitting tubes that were driven into grid current, i.e. class B2. Isn't that just what I said? The booster amps were indeed desirous of input power and for precisely that reason, BUT they were uncommon then and very rare now. It added very little weight and power to provide them with a high impedance input driver section but by the time that was figured out the era of big tube audio amplifiers was over. Refer to the "Audio Anthologies" which I believe you said you possess, the "Musician's Amplifier Senior" by Sarser and Sprinkle (Sarser is still alive and I have acquaintences who correspond with him occasionally) was the prototype for several of the big Altecs such as the 260B. Most ball parks had a couple of these in that era. You also saw them used in the occasional electronic church organ install. |
#17
Posted to rec.audio.tubes
|
|||
|
|||
amplifier input sensitivity
"flipper" wrote in message ... No it isn't. dBV is specified relative to 1V. I.E. 0dBV is 1V, 10dBV is 10V, etc. There is no 'ambiguity' nor is it 'misleading'. This is a mistake, by the way. +10dBV is approximately 3.1V. +20dBV is 10V. dBu is specified relative to .775V (rounded). I.E. 0dBV is .775V (rounded), 10dBu is 7.746V, etc. The same mistake. +20dBu corresponds to 7.75V. |
#18
Posted to rec.audio.tubes
|
|||
|
|||
amplifier input sensitivity
In article
, Bret L wrote: On Apr 21, 9:43*pm, John Byrns wrote: In article , *Bret L wrote: *Most consumer audio amplifiers, and in fact most in general (EXCEPT the "booster" amps of yore such as large Altecs which were nothing more than a single pair of medium mu transmitting triodes with an array of transformers for power supply, input and output) are bridging input loads. They are not designed to accept input power, just voltage. If one needs to drive them from a source which wants a defined load then one simply puts a load resistor across the input. I don't agree, I remember some of those PA "booster" amplifiers were designed to accept input power, this was sometimes necessary because some of them used transmitting tubes that were driven into grid current, i.e. class B2. Isn't that just what I said? No, you said "They are not designed to accept input power, just voltage." You said they were "NOT designed to accept input power", emphasis of the NOT is mine. I said just the opposite that they "were designed to accept input power", note the absence of the "not" in what I said. The booster amps were indeed desirous of input power and for precisely that reason, BUT they were uncommon then and very rare now. It added very little weight and power to provide them with a high impedance input driver section but by the time that was figured out the era of big tube audio amplifiers was over. Refer to the "Audio Anthologies" which I believe you said you possess, Yes, you are correct, I do possess the "Audio Anthologies", however when did I admit to possessing them? the "Musician's Amplifier Senior" by Sarser and Sprinkle (Sarser is still alive and I have acquaintences who correspond with him occasionally) was the prototype for several of the big Altecs such as the 260B. Most ball parks had a couple of these in that era. You also saw them used in the occasional electronic church organ install. Regards, John Byrns -- Regards, John Byrns Surf my web pages at, http://fmamradios.com/ |
#19
Posted to rec.audio.tubes
|
|||
|
|||
amplifier input sensitivity
I think your being a tad contrary.
So: it's not inconsistent that "m" rather than "mW" stands for "milliwatt", and "m" isn't in this case a prefix, and "V" is not a unit, and no-one finds this any more confusing than anything else they are too stupid to know about. Thanks, Flips. The glorious light of a new dawn. Everyone knows what an ovtave sounds like, and if they don't you can tell them its like somewhere, in somwhere over the rainbow. A reduction from 20kHz to 10kHz is the loss of an octave. Similarly, a cut at the low end from 20Hz to 40Hz would lose an octave. Makes sense to me. Make an effort. Ian "flipper" wrote in message news On Fri, 23 Apr 2010 01:05:59 +0100, "Ian Iveson" wrote: Flipper wrote: dBu is specified relative to .775V (rounded). I.E. 0dBV is .775V (rounded), 10dBu is 7.746V, etc. The same mistake. +20dBu corresponds to 7.75V. The dB has become a dog's dinner of a non unit that causes confusion wherever it goes. As a convention, it's perfectly clear once you've learned it, but it isn't obvious. In particular, if you wanted a confusing way of expressing the voltage necessary to increase the power by a certain proportion, given constant resistance, dBV would be a good candidate. For that purpose it doesn't matter what flavor of dB you use because a doubling of power is, by definition, 3dB, is 3dB, is 3dB. Well yes. But the simple ratio of voltages is not the same as the simple ratio of powers. Half the voltage is -6dB, half the power is -3dB. Sticking the V on the end of the dB doesn't give much of a clue that this will be the case, especially as there is no explicit indication that the dBV has anything to do with power. A symbol is not a self contained tutorial. You have to learn what they stand for and that's true of any symbol. The definition for power decibels is 10log10(ratio) and for amplitudes 20log10(ratio) Surely you agree that many ppl find the dB at least a bit hard to grasp? Why do you want to demonstrate that it is straightforward? People are usually confused by the jargon of any technical field they're unfamiliar with and it's usually straightforward to those who've learned and routinely use it. For example, when double checking FB on a power amp in spice I disconnect FB, measure output power, connect FB, measure output power, and verify it's 100 to 1 for 20db or 10 to 1 for 10dB. Which, btw, is how I made the stupid mistake. In spice I can measure power 'directly' and reflexively know 10dB is 10x power. A stupid mistake from not thinking about what I was typing. I don't quite follow what mistake you mean, I'm afraid. At the top where I used power ratio for voltage ratio. One thing that confused me for some time arises from the fact that my spice automatically labels the y-axis in terms of current or voltage, because there's no such thing as a power "probe". So if I multiply a current by a voltage, and graph the result, it's not obvious that the result is power, although, almost by fluke, it still gets the dB scale right in the sense that the numerical values are correct. Similarly with resistance, and even stuff like flux density, as long as I'm careful to stick with units from the same family. I can't speak to the user friendliness, or lack thereof, in your spice program. In mine I can put a probe on the 'body' of a resistor, and most active devices, to get a direct, instantaneous, power reading or a plot of it vs time. What 'nominal' or 'reference' power do I start with or 'calibrate' to? Nothing. It doesn't matter. 10dB is 10dB. The flavor only matters if you care about an interchange reference level like, say, building a repeater that you need to work in a 'standard' telecom system that expects 0dB to be 1mW into 600 Ohms. Proper units don't need all these ifs and buts. A metre or a second can be simply defined in a jiffy. Try explaining 'why' there is a 'meter', a 'yard', and a 'furlong' and the history of when they came about for what reason. You'll find yourself employing just as many "ifs and buts." dBm, itself, is just as simple as a meter. A meter is --- this long --- and 0 dBm is 1 mW into 600 Ohms. Then having established that form of expression you could add the totally contrary dBm, for good measure. It isn't 'contrary' and I'm afraid you have it backwards. dBm (as simply dB, the "Bel" in honor of Alexander Graham Bell since it was telephone transmission the unit was created for) came first and it's arguable that dBV was to 'simplify' things like you opine about. I didn't mean to imply a historical sequence, but rather the order one might introduce the dB such as to highlight the confusion. My object was to reduce confusion. Why do you want to increase it? It is contrary. In dBV, the V is a unit. In dBm, the m is not a unit. How can that possibly be consistent? In plain dB, there's no subsript at all. The rules of subscripting are different in each case. No, it isn't. Decibel is dimensionless and neither the V or m is a 'unit'. They are one letter abbreviations for a 'word' that makes it simple to remember when combined into the 'symbol'. dB - decibel, V - Volt, m - milliwatt. dBV - decibels-Volt. That the word or letter may serve as a 'unit' in another context is how English works. There's only 26 letters and a few appropriate words for any one thing so they're bound to have multiple uses. You are trying to find ways to confuse yourself. When people don't know about something they tend to think it's 'arbitrary' or 'stupid' but that's rarely the case. Like, who's the idiot that decided there should be so many '12s' in things? Well, I don't know who the 'idiot' was but the reason is that, back when most people we're mathematically challenged, 12 is divisible by 2, 3, and 4. It made life simpler. You invented the idiot part. I simply recount what I've heard people 'complain' about. Dozens are clearly useful, as are 60s and 360s. 10 is the unfortunate choice, but even then I'm sure everyone knows how that came about. That's fine. In your case 10 is 'stupid'. Most people I've heard rant about it think '12' is stupid. The point remains that people like to call things they don't understand the reasons for 'stupid' or 'needlessly complex' or some other complaint. I'm not saying the dB is arbitrary or stupid. On the contrary, it serves a purpose very well. It's the inconsistency of subsripting that's a dog's dinner, and that in combination with the absence of explicit reference to power (unavoidable I guess) that can make it hard to follow. It is you who are inventing the 'contrariness' and supposed 'inconsistency'. Transforms are a bit tricky for folk to get their heads around, and a cluttering of historical debris doesn't make it easier for the novice. It would if you didn't insist on creating your own 'meanings'. I tried to give a brief background on dBm for the same reason. Waaaay back when you have people building the first phone lines, and we all know the signal degrades over distance, we obviously need some way to express that and the first 'unit' was the MSC (Mile [of] Standard Cable). I.E. How much the signal degrades over a mile of the stuff we use. Makes sense if you think about it, at least for a starter. Now, as things progress it's not really wise to tie a 'standard' to a particular type of cable (we use multiple types now) so in the 1920's the "TU" (transmission unit) was defined (as ten times the base10 logarithm of the ratio of measured power to reference power) and, to reduce confusion, intentionally made it close to the MSC (=1.056 TU). The TU then became the decibel. Now, I don't know why 1mW was the preferred reference but I assure you there was some practical reason and, at any rate, back when the MSC was done there were no radars, hi-fi stereos, AM radio, or any other 'electronics' as we know it to bother with and no 'confusion' about what's being worked with. It was a telephone system, period. That was the problem being dealt with, you care about impedance matching, power transfer, and it works. As I previously mentioned, when you start adding electronic repeaters it makes perfectly good sense to use the dB(m) reference level because that's the system you're putting them into. And it makes perfect sense that, if one is going to make a general purpose amplifier, people used to making repeaters are likely to use the same terminology, not to mention the first ones are likely to be nothing more than the same repeaters used for a different purpose, and we know they expect 0dBm as reference, so that's likely to be the 'reference level' for whatever else you stick them in because why redesign the damn things? Of course, eventually you do because not all applications have the same requirements. And, again of course, after time and things have multiplied a million fold, so people are separated from the 'original reason', there's likely to be someone who figures things would be 'simpler' if we used '1' as a reference, instead of the 'arbitrary' and 'stupid' .775V, so we get dBV. And when we get to radar reflection we get another use for dB, and another letter. And then we have speakers, and another use for dB, and so on, because, as it turns out, power is power and the fundamental ratio applies but 1mW into 600 Ohms is simply not a useful reference for radar and speakers. Again, I can't tell you how 'everything' came about but I assure you there were practical reasons. Or, at the very least, someone thought there was a good practical reason and if it persisted it probably was. I haven't argued that the dB is not useful, or not valid. I'm trying to identify why ppl find it hard to follow. It is hard to follow. Ppl have big arguments about it and come from all sorts of directions. What's your explanation for that? Already answered. They're unfamiliar with it or know just enough to think they know when they don't. You've probably noticed that a lack of knowledge seldom stops people from arguing about things, regardless of the topic. Ohm's Law, for example, doesn't attract any debate at all. Neither do simple units like seconds or metres. Even logs don't attract the same level of confusion. Even you have just made a couple of errors, which you wouldn't with seconds, metres or logs. I appreciate the confidence in my otherwise infallibility. Isn't this the kind of thing that Europe was supposed to sort out? A more consistent convention for dealing with its subscripts...the u, V, m etc....would help maybe. What inconsistency are you claming? See above. In terms of self-consistency, V is an SI unit whereas m is a misplaced SI prefix and actually I don't know what the u stands for. Is it a corruption of a mu symbol? In that case it would be a misplaced corruption of an SI prefix. In terms of consistency with SI units, dBV relates to power but makes no mention of a unit of power, the m isn't a properly applicable prefix even if it wasn't misplaced, and the u isn't either, and it's corrupt to boot. As I explained above, you are inventing meanings that do not exist. Not only that, you go out of your way to invent problems, like claiming a prefix that isn't even in the prefix position, so obviously not a prefix, has TWO errors with one being it isn't in the prefix position which, to most people, would be proof enough it isn't a prefix.. yet you decide it is. I couldn't change it even if I were inclined to try so you can either learn the convention that's been around for a hundred years or not. Your choice. On a positive note, I've noticed that many of those here who appear versed in audio technology are quite disciplined in their usage of the decibel, especially with respect to voltage. In circumstances where a voltage is significant on its own account, I generally see volts or percentages. Where voltage is significant by dint of its relation to power, I generally see dB. Not a habit I've fully adopted, I must admit. It would be nice, I think, if the same kind of distinction were made when talking about frequency. We don't use octaves as much as we ought, I fear. Like most things it depends on what is of interest. Talk to musicians and you'll get an earful about octaves (plus a lot more). And yet, when we are concerned about audible bandwidth, octaves would be linear with respect to perception, so there is just as much reason to use the octave in audio as the decibel. 'Linear' to what perception? And if I'm building a 'hi-fi' amplifier that is supposed to cover all audible frequencies why do I care how one 'perceives' 6kHz relative to 1kHz? They're both there so... what? Ok, so an amp with a 20-10kHz bandwidth is what? octave? what? than one with a 20-20kHz bandwidth? Twice, or thrice, or half, or what of what? What additional usefulness do I get by measuring bandwidth in octaves? They both linearise an axis that would otherwise require a log scale to fit with perception, and therefore with significance. I do plot it on a log scale, log10. Ian |
#20
Posted to rec.audio.tubes
|
|||
|
|||
amplifier input sensitivity
On Apr 25, 5:45*am, "Ian Iveson"
wrote: I think your being a tad contrary. So: it's not inconsistent that "m" rather than "mW" stands for "milliwatt", and "m" isn't in this case a prefix, and "V" is not a unit, and no-one finds this any more confusing than anything else they are too stupid to know about. Thanks, Flips. The glorious light of a new dawn. Everyone knows what an ovtave sounds like, and if they don't you can tell them its like somewhere, in somwhere over the rainbow. A reduction from 20kHz to 10kHz is the loss of an octave. Similarly, a cut at the low end from 20Hz to 40Hz would lose an octave. Makes sense to me. Make an effort. Ian "flipper" wrote in message news On Fri, 23 Apr 2010 01:05:59 +0100, "Ian Iveson" wrote: Flipper wrote: dBu is specified relative to .775V (rounded). I.E. 0dBV is .775V (rounded), 10dBu is 7.746V, etc. The same mistake. +20dBu corresponds to 7.75V. The dB has become a dog's dinner of a non unit that causes confusion wherever it goes. As a convention, it's perfectly clear once you've learned it, but it isn't obvious. In particular, if you wanted a confusing way of expressing the voltage necessary to increase the power by a certain proportion, given constant resistance, dBV would be a good candidate. For that purpose it doesn't matter what flavor of dB you use because a doubling of power is, by definition, 3dB, is 3dB, is 3dB. Well yes. But the simple ratio of voltages is not the same as the simple ratio of powers. Half the voltage is -6dB, half the power is -3dB. Sticking the V on the end of the dB doesn't give much of a clue that this will be the case, especially as there is no explicit indication that the dBV has anything to do with power. A symbol is not a self contained tutorial. You have to learn what they stand for and that's true of any symbol. The definition for power decibels is 10log10(ratio) and for amplitudes 20log10(ratio) Surely you agree that many ppl find the dB at least a bit hard to grasp? Why do you want to demonstrate that it is straightforward? People are usually confused by the jargon of any technical field they're unfamiliar with and it's usually straightforward to those who've learned and routinely use it. For example, when double checking FB on a power amp in spice I disconnect FB, measure output power, connect FB, measure output power, and verify it's 100 to 1 for 20db or 10 to 1 for 10dB. Which, btw, is how I made the stupid mistake. In spice I can measure power 'directly' and reflexively know 10dB is 10x power. A stupid mistake from not thinking about what I was typing. I don't quite follow what mistake you mean, I'm afraid. At the top where I used power ratio for voltage ratio. One thing that confused me for some time arises from the fact that my spice automatically labels the y-axis in terms of current or voltage, because there's no such thing as a power "probe". So if I multiply a current by a voltage, and graph the result, it's not obvious that the result is power, although, almost by fluke, it still gets the dB scale right in the sense that the numerical values are correct. Similarly with resistance, and even stuff like flux density, as long as I'm careful to stick with units from the same family. I can't speak to the user friendliness, or lack thereof, in your spice program. In mine I can put a probe on the 'body' of a resistor, and most active devices, to get a direct, instantaneous, power reading or a plot of it vs time. What 'nominal' or 'reference' power do I start with or 'calibrate' to? Nothing. It doesn't matter. 10dB is 10dB. The flavor only matters if you care about an interchange reference level like, say, building a repeater that you need to work in a 'standard' telecom system that expects 0dB to be 1mW into 600 Ohms. Proper units don't need all these ifs and buts. A metre or a second can be simply defined in a jiffy. Try explaining 'why' there is a 'meter', a 'yard', and a 'furlong' and the history of when they came about for what reason. You'll find yourself employing just as many "ifs and buts." dBm, itself, is just as simple as a meter. A meter is --- this long --- and 0 dBm is 1 mW into 600 Ohms. Then having established that form of expression you could add the totally contrary dBm, for good measure. It isn't 'contrary' and I'm afraid you have it backwards. dBm (as simply dB, the "Bel" in honor of Alexander Graham Bell since it was telephone transmission the unit was created for) came first and it's arguable that dBV was to 'simplify' things like you opine about. I didn't mean to imply a historical sequence, but rather the order one might introduce the dB such as to highlight the confusion. My object was to reduce confusion. Why do you want to increase it? It is contrary. In dBV, the V is a unit. In dBm, the m is not a unit. How can that possibly be consistent? In plain dB, there's no subsript at all. The rules of subscripting are different in each case. No, it isn't. Decibel is dimensionless and neither the V or m is a 'unit'. They are one letter abbreviations for a 'word' that makes it simple to remember when combined into the 'symbol'. dB - decibel, V - Volt, m - milliwatt. dBV - decibels-Volt. That the word or letter may serve as a 'unit' in another context is how English works. There's only 26 letters and a few appropriate words for any one thing so they're bound to have multiple uses. You are trying to find ways to confuse yourself. When people don't know about something they tend to think it's 'arbitrary' or 'stupid' but that's rarely the case. Like, who's the idiot that decided there should be so many '12s' in things? Well, I don't know who the 'idiot' was but the reason is that, back when most people we're mathematically challenged, 12 is divisible by 2, 3, and 4. It made life simpler. You invented the idiot part. I simply recount what I've heard people 'complain' about. Dozens are clearly useful, as are 60s and 360s. 10 is the unfortunate choice, but even then I'm sure everyone knows how that came about. That's fine. In your case 10 is 'stupid'. Most people I've heard rant about it think '12' is stupid. The point remains that people like to call things they don't understand the reasons for 'stupid' or 'needlessly complex' or some other complaint. I'm not saying the dB is arbitrary or stupid. On the contrary, it serves a purpose very well. It's the inconsistency of subsripting that's a dog's dinner, and that in combination with the absence of explicit reference to power (unavoidable I guess) that can make it hard to follow. It is you who are inventing the 'contrariness' and supposed 'inconsistency'. Transforms are a bit tricky for folk to get their heads around, and a cluttering of historical debris doesn't make it easier for the novice. It would if you didn't insist on creating your own 'meanings'. I tried to give a brief background on dBm for the same reason. Waaaay back when you have people building the first phone lines, and we all know the signal degrades over distance, we obviously need some way to express that and the first 'unit' was the MSC (Mile [of] Standard Cable). I.E. How much the signal degrades over a mile of the stuff we use. Makes sense if you think about it, at least for a starter. Now, as things progress it's not really wise to tie a 'standard' to a particular type of cable (we use multiple types now) so in the 1920's the "TU" (transmission unit) was defined (as ten times the base10 logarithm of the ratio of measured power to reference power) and, to reduce confusion, intentionally made it close to the MSC (=1.056 TU). The TU then became the decibel. Now, I don't know why 1mW was the preferred reference but I assure you there was some practical reason and, at any rate, back when the MSC was done there were no radars, hi-fi stereos, AM radio, or any other 'electronics' as we know it to bother with and no 'confusion' about what's being worked with. It was a telephone system, period. That was the problem being dealt with, you care about impedance matching, power transfer, and it works. As I previously mentioned, when you start adding electronic repeaters it makes perfectly good sense to use the dB(m) reference level because that's the system you're putting them into. And it makes perfect sense that, if one is going to make a general purpose amplifier, people used to making repeaters are likely to use the same terminology, not to mention the first ones are likely to be nothing more than the same repeaters used for a different purpose, and we know they expect 0dBm as reference, so that's likely to be the 'reference level' for whatever else you stick them in because why redesign the damn things? Of course, eventually you do because not all applications have the same requirements. And, again of course, after time and things have multiplied a million fold, so people are separated from the 'original reason', there's likely to be someone who figures ... read more »- Hide quoted text - - Show quoted text - This thread is becoming an exercise in futility. A decibel and an octave can mean quite different things in different cases. In consumer grade audio environment (and even in pro audio) 1 volt signal levels have become usual and bridged input is more common due to an improved frequency response . O dBm is a historical reference level and we all have to respect it and understand why and how it was established. Voltage bridging is very approximate in cosumer grade audio and the impedances involved may range from about 1KOhm to 10K or even higher for inputs... |
#21
Posted to rec.audio.tubes
|
|||
|
|||
amplifier input sensitivity
This thread is becoming an exercise in futility. A decibel
and an octave can mean quite different things in different cases. In consumer grade audio environment (and even in pro audio) 1 volt signal levels have become usual and bridged input is more common due to an improved frequency response . O dBm is a historical reference level and we all have to respect it and understand why and how it was established. Voltage bridging is very approximate in cosumer grade audio and the impedances involved may range from about 1KOhm to 10K or even higher for inputs... This is recreational audio tubes. Surely futility is what we come here for? At a guess, the three of us know what a dB is, and why it's useful. Considering how immediately useful it is, it needs no historical explanation IMO, although it's a bit interesting. My point was about intelligibility of expression...about consistency of language. dBm, in particular, is a mess in that respect. In addition to the points I've made already, even the "d" prefix is odd in the context of standard units. The dB itself is accepted by, but not a part of, the SI system. In this respect it is one of a few odd units. The dBm, AFAIK, is outlawed from SI standard units, because it is considered to be confusing and ambiguous. There is rarely a need to know what it means...it's part of the jargon that's used in the business, that's all. Similarly dBu. Dunno what it means, don't care, and can't be arsed to look it up. How would you abbreviate the decibel-metre, incidentally? dB.m? And how would you write one thousandth of a decibel? mdB? The octave is similar in that, if you wanted to express an absolute frequency in octaves, you would need to attach a reference frequency. It is also similar in that it is a ratio that is mathematically unitless, and yet is inextricably linked to frequency, just like the dB is linked to power. If I found myself needing to teach the meaning of the dB, the octave might be a useful comparison, because everyone knows what an octave sounds like, and everyone knows it is related to frequency, and they also know that it sounds essentially the same whatever the absolute frequency span. I can sing "somewhere" in any key and the octave sounds like the same span, even though the frequency difference would not be the same in each case. All futile to those who don't want to see the point, obviously. They are both useful in audio insofar as they relate more directly to perception than their "host" units, Watt and Hz. Significance in audio is very often about perception. If we can linearise what is significant, it becomes more intelligible. Should anyone wish to know more, I suggest they google the SI system of units. Ian |
#22
Posted to rec.audio.tubes
|
|||
|
|||
amplifier input sensitivity
On Apr 26, 11:50*pm, "Ian Iveson"
wrote: *This thread is becoming an exercise in futility. A decibel and an octave can mean quite different things in different cases. In consumer grade audio environment (and even in pro audio) 1 volt signal levels have become usual and bridged input is more common due to an improved frequency response . *O dBm is a historical reference level and we all have to respect it and understand why and how it was established. Voltage bridging is very approximate in cosumer grade audio and the impedances involved may range from about 1KOhm to 10K or even higher for inputs... This is recreational audio tubes. Surely futility is what we come here for? At a guess, the three of us know what a dB is, and why it's useful. Considering how immediately useful it is, it needs no historical explanation IMO, although it's a bit interesting. My point was about intelligibility of expression...about consistency of language. dBm, in particular, is a mess in that respect. In addition to the points I've made already, even the "d" prefix is odd in the context of standard units. The dB itself is accepted by, but not a part of, the SI system. In this respect it is one of a few odd units. The dBm, AFAIK, is outlawed from SI standard units, because it is considered to be confusing and ambiguous. There is rarely a need to know what it means...it's part of the jargon that's used in the business, that's all. Similarly dBu. Dunno what it means, don't care, and can't be arsed to look it up. How would you abbreviate the decibel-metre, incidentally? dB.m? And how would you write one thousandth of a decibel? mdB? The octave is similar in that, if you wanted to express an absolute frequency in octaves, you would need to attach a reference frequency. It is also similar in that it is a ratio that is mathematically unitless, and yet is inextricably linked to frequency, just like the dB is linked to power. If I found myself needing to teach the meaning of the dB, the octave might be a useful comparison, because everyone knows what an octave sounds like, and everyone knows it is related to frequency, and they also know that it sounds essentially the same whatever the absolute frequency span. I can sing "somewhere" in any key and the octave sounds like the same span, even though the frequency difference would not be the same in each case. All futile to those who don't want to see the point, obviously. They are both useful in audio insofar as they relate more directly to perception than their "host" units, Watt and Hz. Significance in audio is very often about perception. If we can linearise what is significant, it becomes more intelligible. Should anyone wish to know more, I suggest they google the SI system of units. Ian Amen. |
#23
Posted to rec.audio.tubes
|
|||
|
|||
amplifier input sensitivity
flipper wrote:
[below] I've said that the strange notation of the dB and its derivative expressions is part of the reason why it causes confusion to the novice. Perhaps it's different in the US, but here the standard school curriculum strictly and explicitly adheres to the SI system. In this context, the SI and I consider the dB to be strange but unavoidable, whereas its derivatives are particularly inconsistent with the standard, and therefore confusing. That's enough for me. Let's just agree to differ, eh? With respect to the octave, I suggested we might perhaps use it more as a measure of frequency, and I have said most of why, in that it linearises a log scale in an intelligible manner. This occurred to me because I wondered why, in the standard form I have seen of plots of gain v bandwidth, a decimal log scale is used for frequency, marked in Hz, whereas the gain axis is linearised using dB. Why linearise the scale for one axis and not the other? I wondered. There is a very basic reason to use the octave, it seems to me, but for the moment let's look at the simple intelligibility side of things. A filter might be said to roll of at 3db per octave. I've seen this use of the octave lots of times. Then I look at the graph and it's shown in Hz rather than octaves and that's not very convenient is it? If the graph were calibrated in octaves I could look at it and think ah yes, I see, a one octave drop in frequency results in a 3dB fall in gain. Us Europeans like this kind of thing. For more advanced intelligibility, I could say that, in order for an amp to be stable, the frequency difference between the dominant pole and the next must be a certain number of octaves per dB of open-loop gain. That makes a simple picture out of a problem that lots of people seem to have advanced difficulty with. Now, so far you might say that the octave has no advantage over the decade, which would also make the significance of frequency-doubling explicit. But consider that the construction of our ears, and that of musical instruments, and therefore our expectations of what is musical and what is not, is all built round the same rules of how filters work everywhere, in electronics and the rest of the physical world. The harmonic structure of music and speech conform to the same rules that apply to the transient responses of our hi-fi systems. It might be useful, for example, to show harmonic distortion on a spectrogram with the frequency axis marked in octaves, using the input frequency as the reference. That would be convenient because there would be no need for negative octaves. It would be particularly revealing if the octaves were subdivided into tones, especially coz I know you like dozens and the dozen and the eight combine to make music! Or not, if it's the 7th harmonic, for example, that falls close to an unfortunate tone IIRC. It could be even better if, instead of using a 1kHz test tone, some standard note were used instead. Then the tones on the frequency axis could be marked in notes, and it would be even easier to imagine what the distortion would sound like. A graph showing bandwidth might be clearer if it were calibrated in octaves and notes above and below middle C, or A, or whatever. The span of instruments, ears, amps, speakers and music could all be accommodated on the same scale, making comparison a doddle. Worth a thought, it seemed to me. But have it your own way, I don't care. We may well be on different planets. Ian "flipper" wrote in message ... On Tue, 27 Apr 2010 04:50:46 +0100, "Ian Iveson" wrote: This thread is becoming an exercise in futility. A decibel and an octave can mean quite different things in different cases. In consumer grade audio environment (and even in pro audio) 1 volt signal levels have become usual and bridged input is more common due to an improved frequency response . O dBm is a historical reference level and we all have to respect it and understand why and how it was established. Voltage bridging is very approximate in cosumer grade audio and the impedances involved may range from about 1KOhm to 10K or even higher for inputs... This is recreational audio tubes. Surely futility is what we come here for? At a guess, the three of us know what a dB is, and why it's useful. Considering how immediately useful it is, it needs no historical explanation IMO, although it's a bit interesting. The historical explanation was specifically for how dBm and dBu came about.. I think it is also useful for those who imagine decibel is some arbitrary invention that it's based on empirical studies of human perception with the minimum observed detectable volume change being, on average, .1 Bel. People tend to like 'units' that correspond to the 'common size' of whatever they're measuring, which is why meter is popular despite the fact we could use 6.684587122671E-12 au. And, conversely, why astronomers like au instead of meters. I.E. 1 decibel (dB) is 1 'loudness perception unit'. My point was about intelligibility of expression...about consistency of language. dBm, in particular, is a mess in that respect. In addition to the points I've made already, even the "d" prefix is odd in the context of standard units. Never heard of deci? How about centi? http://www.asknumbers.com/Centimeter...onversion.aspx Centimeters (cm) To Decimeters (dm) Conversion The dB itself is accepted by, but not a part of, the SI system. In this respect it is one of a few odd units. The dBm, AFAIK, is outlawed from SI standard units, because it is considered to be confusing and ambiguous. The standards committee told you that, did they? SI unit rules and style conventions have a built in paradox with respect to logarithmic 'ratio units'. They requires knowing the reference but SI objects to both 'creating'/'describing' units with suffixes and textual 'explanation' of units. There's no 'SI standard' way of saying 'what it is'. That's understandable, given the SI view that all things derive from the 7 basic units, and gets to the more philosophical problem that decibel (and neper) does not represent a physical property or process, like ampere or the derived unit ohm do, so you have the conundrum that it doesn't 'exist', as indicated by being dimensionless. It's not one of the 7 base units nor can it be definitively derived from them (because it can ratio 'anything'). Of course, it has dimension when you know the reference but see SI problem number 1. SI conundrums aside, the dB is clearly useful and the only thing 'ambiguous' about dBm is your steadfast refusal to accept the definition of it, not to mention it's rather obtuse to complain it doesn't follow current SI convention when there isn't one. dBmW is just as 'non SI' as dBm but if it makes you happy then write dBmW. Some people do but the vast majority do not so if you want to know what the heck they're talking about you need to know what dBm is whether you 'approve' of it or not. There is rarely a need to know what it means...it's part of the jargon that's used in the business, that's all. Similarly dBu. Dunno what it means, don't care, and can't be arsed to look it up. No need as long as one doesn't care about working with telecom, consumer, or professional audio equipment, all of which reference to one or the other of dBm, dBV or dBu (not to mention the scores of other dB[something] used in all sorts of technical fields) http://www.behringer.com/EN/download...P0573_S_EN.pdf It might be nice to know what they mean when spec'ing, for example, max input as +22 dBu @ 0 dB gain. Standard telecom fiber optic transmitters output 0 to +10 dBm. Might not matter if your a cable plugger but the thread is "amplifier input sensitivity" so just how do you suggest one determine it without knowing what the dern expected levels are? How would you abbreviate the decibel-metre, incidentally? dB.m? The first question would be why you're even thinking of it but I imagine it would be just as confusing as the equally mysterious metre-Watt (m.W). And how would you write one thousandth of a decibel? mdB? I've seen it done. http://cires.colorado.edu/jimenez-gr.../NIDAQ_Ref.pdf Attenuation (in mdB) = -[20 log10 (Vo/Vi)]*1000 The octave is similar in that, if you wanted to express an absolute frequency in octaves, you would need to attach a reference frequency. It is also similar in that it is a ratio that is mathematically unitless, and yet is inextricably linked to frequency, just like the dB is linked to power. If I found myself needing to teach the meaning of the dB, the octave might be a useful comparison, because everyone knows what an octave sounds like, and everyone knows it is related to frequency, and they also know that it sounds essentially the same whatever the absolute frequency span. I can sing "somewhere" in any key and the octave sounds like the same span, even though the frequency difference would not be the same in each case. All futile to those who don't want to see the point, obviously. You never presented any 'point' of using octave as some sort of 'teaching tool'. You argued I should plot things in octaves and when I asked 'what for' gave nothing. They are both useful in audio insofar as they relate more directly to perception than their "host" units, Watt and Hz. Significance in audio is very often about perception. If we can linearise what is significant, it becomes more intelligible. Octave is, as I already said, certainly useful to 'music'. What use it is to 'audio' you have yet to explain. Should anyone wish to know more, I suggest they google the SI system of units. There isn't an SI unit at all for measuring 'perception units', be it loudness or your octaves. Let me ask, besides satisfying your urge to bitch what use does deliberately trying to confuse things serve? Ian |
#24
Posted to rec.audio.tubes
|
|||
|
|||
amplifier input sensitivity
On Apr 28, 9:37*pm, "Ian Iveson"
wrote: flipper wrote: [below] I've said that the strange notation of the dB and its derivative expressions is part of the reason why it causes confusion to the novice. Perhaps it's different in the US, but here the standard school curriculum strictly and explicitly adheres to the SI system. In this context, the SI and I consider the dB to be strange but unavoidable, whereas its derivatives are particularly inconsistent with the standard, and therefore confusing. That's enough for me. Let's just agree to differ, eh? With respect to the octave, I suggested we might perhaps use it more as a measure of frequency, and I have said most of why, in that it linearises a log scale in an intelligible manner. This occurred to me because I wondered why, in the standard form I have seen of plots of gain v bandwidth, a decimal log scale is used for frequency, marked in Hz, whereas the gain axis is linearised using dB. Why linearise the scale for one axis and not the other? I wondered. There is a very basic reason to use the octave, it seems to me, but for the moment let's look at the simple intelligibility side of things. A filter might be said to roll of at 3db per octave. I've seen this use of the octave lots of times. Then I look at the graph and it's shown in Hz rather than octaves and that's not very convenient is it? If the graph were calibrated in octaves I could look at it and think ah yes, I see, a one octave drop in frequency results in a 3dB fall in gain. Us Europeans like this kind of thing. For more advanced intelligibility, I could say that, in order for an amp to be stable, the frequency difference between the dominant pole and the next must be a certain number of octaves per dB of open-loop gain. That makes a simple picture out of a problem that lots of people seem to have advanced difficulty with. Now, so far you might say that the octave has no advantage over the decade, which would also make the significance of frequency-doubling explicit. But consider that the construction of our ears, and that of musical instruments, and therefore our expectations of what is musical and what is not, is all built round the same rules of how filters work everywhere, in electronics and the rest of the physical world. The harmonic structure of music and speech conform to the same rules that apply to the transient responses of our hi-fi systems. It might be useful, for example, to show harmonic distortion on a spectrogram with the frequency axis marked in octaves, using the input frequency as the reference. That would be convenient because there would be no need for negative octaves. It would be particularly revealing if the octaves were subdivided into tones, especially coz I know you like dozens and the dozen and the eight combine to make music! Or not, if it's the 7th harmonic, for example, that falls close to an unfortunate tone IIRC. It could be even better if, instead of using a 1kHz test tone, some standard note were used instead. Then the tones on the frequency axis could be marked in notes, and it would be even easier to imagine what the distortion would sound like. A graph showing bandwidth might be clearer if it were calibrated in octaves and notes above and below middle C, or A, or whatever. The span of instruments, ears, amps, speakers and music could all be accommodated on the same scale, making comparison a doddle. Worth a thought, it seemed to me. But have it your own way, I don't care. We may well be on different planets. Ian "flipper" wrote in message ... On Tue, 27 Apr 2010 04:50:46 +0100, "Ian Iveson" wrote: This thread is becoming an exercise in futility. A decibel and an octave can mean quite different things in different cases. In consumer grade audio environment (and even in pro audio) 1 volt signal levels have become usual and bridged input is more common due to an improved frequency response . *O dBm is a historical reference level and we all have to respect it and understand why and how it was established. Voltage bridging is very approximate in cosumer grade audio and the impedances involved may range from about 1KOhm to 10K or even higher for inputs... This is recreational audio tubes. Surely futility is what we come here for? At a guess, the three of us know what a dB is, and why it's useful. Considering how immediately useful it is, it needs no historical explanation IMO, although it's a bit interesting. The historical explanation was specifically for how dBm and dBu came about.. I think it is also useful for those who imagine decibel is some arbitrary invention that it's based on empirical studies of human perception with the minimum observed detectable volume change being, on average, .1 Bel. People tend to like 'units' that correspond to the 'common size' of whatever they're measuring, which is why meter is popular despite the fact we could use 6.684587122671E-12 *au. And, conversely, why astronomers like au instead of meters. I.E. 1 decibel (dB) is 1 'loudness perception unit'. My point was about intelligibility of expression...about consistency of language. dBm, in particular, is a mess in that respect. In addition to the points I've made already, even the "d" prefix is odd in the context of standard units. Never heard of deci? How about centi? http://www.asknumbers.com/Centimeter...onversion.aspx Centimeters (cm) To Decimeters (dm) Conversion The dB itself is accepted by, but not a part of, the SI system. In this respect it is one of a few odd units. The dBm, AFAIK, is outlawed from SI standard units, because it is considered to be confusing and ambiguous. The standards committee told you that, did they? SI unit rules and style conventions have a built in paradox with respect to logarithmic 'ratio units'. They requires knowing the reference but SI objects to both 'creating'/'describing' units with suffixes and textual 'explanation' of units. There's no 'SI standard' way of saying 'what it is'. That's understandable, given the SI view that all things derive from the 7 basic units, and gets to the more philosophical problem that decibel (and neper) does not represent a physical property or process, like ampere or the derived unit ohm do, so you have the conundrum that it doesn't 'exist', as indicated by being dimensionless. It's not one of the 7 base units nor can it be definitively derived from them (because it can ratio 'anything'). Of course, it has dimension when you know the reference but see SI problem number 1. SI conundrums aside, the dB is clearly useful and the only thing 'ambiguous' about dBm is your steadfast refusal to accept the definition of it, not to mention it's rather obtuse to complain it doesn't follow current SI convention when there isn't one. dBmW is just as 'non SI' as dBm but if it makes you happy then write dBmW. Some people do but the vast majority do not so if you want to know what the heck they're talking about you need to know what dBm is whether you 'approve' of it or not. There is rarely a need to know what it means...it's part of the jargon that's used in the business, that's all. Similarly dBu. Dunno what it means, don't care, and can't be arsed to look it up. No need as long as one doesn't care about working with telecom, consumer, or professional audio equipment, all of which reference to one or the other of dBm, dBV or dBu (not to mention the scores of other dB[something] used in all sorts of technical fields) http://www.behringer.com/EN/download...P0573_S_EN.pdf It might be nice to know what they mean when spec'ing, for example, max input as +22 dBu @ 0 dB gain. Standard telecom fiber optic transmitters output 0 to +10 dBm. Might not matter if your a cable plugger but the thread is "amplifier input sensitivity" so just how do you suggest one determine it without knowing what the dern expected levels are? How would you abbreviate the decibel-metre, incidentally? dB.m? The first question would be why you're even thinking of it but I imagine it would be just as confusing as the equally mysterious metre-Watt (m.W). And how would you write one thousandth of a decibel? mdB? I've seen it done. http://cires.colorado.edu/jimenez-gr.../NIDAQ_Ref.pdf Attenuation (in mdB) = -[20 log10 (Vo/Vi)]*1000 The octave is similar in that, if you wanted to express an absolute frequency in octaves, you would need to attach a reference frequency. It is also similar in that it is a ratio that is mathematically unitless, and yet is inextricably linked to frequency, just like the dB is linked to power. If I found myself needing to teach the meaning of the dB, the octave might be a useful comparison, because everyone knows what an octave sounds like, and everyone knows it is related to frequency, and they also know that it sounds essentially the same whatever the absolute frequency span. I can sing "somewhere" in any key and the octave sounds like the same span, even though the frequency difference would not be the same in each case. All futile to those who don't want to see the point, obviously. You never presented any 'point' of using octave as some sort of 'teaching tool'. You argued I should plot things in octaves and when I asked 'what for' gave nothing. They are both useful in audio insofar as they relate more directly to perception than their "host" units, Watt and Hz. Significance in audio is very often about perception. If we can linearise what is significant, it becomes more intelligible. Octave is, as I already said, certainly useful to 'music'. What use it is to 'audio' you have yet to explain. Should anyone wish to know more, I suggest they google the SI system of units. There isn't an SI unit at all for measuring 'perception units', be it loudness or your octaves. Let me ask, besides satisfying your urge to bitch what use does deliberately trying to confuse things serve? Ian- Hide quoted text - - Show quoted text - Obviously,Erich von Daniken had something in mind in his first best- seller book. |
#25
Posted to rec.audio.tubes
|
|||
|
|||
amplifier input sensitivity
** LEARN TO TRIM - imbecile !! |
#26
Posted to rec.audio.tubes
|
|||
|
|||
amplifier input sensitivity
On 15/06/10 23:46, flipper wrote:
On Thu, 29 Apr 2010 02:37:37 +0100, "Ian Iveson" wrote: flipper wrote: [below] I've said that the strange notation of the dB and its derivative expressions is part of the reason why it causes confusion to the novice. Any notation is 'confusing to the novice', till they learn it. Perhaps it's different in the US, but here the standard school curriculum strictly and explicitly adheres to the SI system. In this context, the SI and I consider the dB to be strange but unavoidable, whereas its derivatives are particularly inconsistent with the standard, and therefore confusing. That's enough for me. You sound like a sales brochure for SI but SI has no mechanism whatsoever for dealing with referenced logarithmic ratios. There is no 'labeling' that *could* be 'consistent' with SI, as it stands, so it's a red herring to complain the clearly useful dB is 'inconsistent' with something that has no means to deal with it, or anything like it, or any substitute. Not to mention that the dB is the ratio of two identically dimensioned quantities and is therefore dimensionless so a system of standardising dimensions is meaningless in this context. Cheers Ian |
#27
Posted to rec.audio.tubes
|
|||
|
|||
amplifier input sensitivity
On Jun 16, 7:42*am, Ian Bell wrote:
On 15/06/10 23:46, flipper wrote: On Thu, 29 Apr 2010 02:37:37 +0100, "Ian Iveson" *wrote: flipper wrote: [below] I've said that the strange notation of the dB and its derivative expressions is part of the reason why it causes confusion to the novice. Any notation is 'confusing to the novice', till they learn it. Perhaps it's different in the US, but here the standard school curriculum strictly and explicitly adheres to the SI system. In this context, the SI and I consider the dB to be strange but unavoidable, whereas its derivatives are particularly inconsistent with the standard, and therefore confusing. That's enough for me. You sound like a sales brochure for SI but SI has no mechanism whatsoever for dealing with referenced logarithmic ratios. There is no 'labeling' that *could* be 'consistent' with SI, as it stands, so it's a red herring to complain the clearly useful dB is 'inconsistent' with something that has no means to deal with it, or anything like it, or any substitute. Not to mention that the dB is the ratio of two identically dimensioned quantities and is therefore dimensionless so a system of standardising dimensions is meaningless in this context. Cheers Ian- Hide quoted text - - Show quoted text - Amen... |
Reply |
Thread Tools | |
Display Modes | |
|
|
Similar Threads | ||||
Thread | Forum | |||
Amplifier Input Impedance | Vacuum Tubes | |||
Input Sensitivity Settings on amps. | Car Audio | |||
Input Sensitivity and Low Pass Filter Settings | Car Audio | |||
Eclipse 5444 and Amplifier input sensitivity? | Car Audio | |||
Ev P3000 - input sensitivity | Pro Audio |