Ap2,
I don't think that if we had loudspeakers that actually produced usable output to even 20hz flat with the power output of the rest of the bandwidth of our speakers we would have any happy neighbors. I know that even at higher frequencies it is hard for our homes to contain those low frequencies. I am not talking about using sub-woofers, those are typically such narrow band devices that I don't consider that a normal part of a regular loudspeaker.
I don't think that if we had loudspeakers that actually produced usable output to even 20hz flat with the power output of the rest of the bandwidth of our speakers we would have any happy neighbors. I know that even at higher frequencies it is hard for our homes to contain those low frequencies. I am not talking about using sub-woofers, those are typically such narrow band devices that I don't consider that a normal part of a regular loudspeaker.
Andrew I agree but I see the lower limit ~2-4Hz as set to minimize any phase shift introduced at the lower speaker range of 20-40Hz.
So at least for the amplifiers PSU I wouldnt expect one to have to size ripple at this lower 2-4Hz.
Thanks
-Antonio
yes, I agree with your post. I was referring to the amplifier and consequently to the psu. 4Hz (if true) requires a lot of work on the amplifier, according to the psu dimensioned and calculated for this frequency. I mean, that too easily declare 4Hz flat on amp's.
I DID do some square wave testing at the same peak amplitude as the sines, but only for the middle three cases in the most-recent table.
I was able to find the minimum capacitance for those cases, too, within 1000 uF I think. I am at work and the numbers are at home but I believe that they were 10000uF, 11000uF, and 6000uF. I would need to verify those, if anyone wanted to rely on them. And I think I did those with a positive-only square wave, for some reason, and then started messing with the .meas commands and never got back to continue with the square waves.
But whatever the minimum C numbers are for squarewaves at the same peak amplitude as the sines, using those values (or more) for reservoir capacitance should give very high confidence that the C value is at least not too low for the power supply to handle the peak output signal, regardless of waveform type.
Regards,
Tom
Last edited:
The signal itself would/shouldn't have to go as low as 4Hz. That marker is just to reduce phase shift of the roll-on filter at the "usable" start of the audio band, around 15Hz and up signal wise. I have the roll-on filter on MF-80 set at 1.5Hz. Try it, it will be a whole new bass experience when you're 'used' to higher frequency roll-on filters.
MagicBox,
I think that we are all in agreement that a wide bandwidth in the amplifier is important and that we don't want the band pass filters to get into the audio range. How much difference there is between 1.5 hz and 4 hz I don't know without doing the math and seeing how far that shifts the high pass frequency where it is still flat in the pass band. I don't think that anyone is going to argue this point, at least I hope not.
I think that we are all in agreement that a wide bandwidth in the amplifier is important and that we don't want the band pass filters to get into the audio range. How much difference there is between 1.5 hz and 4 hz I don't know without doing the math and seeing how far that shifts the high pass frequency where it is still flat in the pass band. I don't think that anyone is going to argue this point, at least I hope not.
LF rolloff is not really relevant to this thread, but there may be good reasons to limit the bandwidth. There is little point in forcing a transducer to attempt to reproduce a frequency which it cannot handle. There will already be significant LF phase shift from the loudspeaker, and this would be a second or third-order rolloff. A first-order rolloff from the amplifier could be harmless in comparison.
DF96,
I agree and often wonder about some who insist on needing a system that actually produces bass down to 20hz and over 25khz. I have good hearing and I know that 25Khz is out of my range, though I still get up to about 20khz. And what besides a pipe organ or a steam train or synthesizer has anything musical at 20hz? Not a stand up bass or anything like a normal drum head or any orchestral instrument that I can think of. And yes I do still have a turntable and albums in my collection. Nothing like the rumble of a low note getting into feedback with the turntable and arm. I have been to many a live show behind the mixing console and I always see the low frequency cut set higher than 20hz. Bass feedback through the mics can be worse than a high frequency feedback if it just barely is perceptable. I can't even stand the sun roof open in some cars, the low frequency rumble will drive me nuts.
I agree and often wonder about some who insist on needing a system that actually produces bass down to 20hz and over 25khz. I have good hearing and I know that 25Khz is out of my range, though I still get up to about 20khz. And what besides a pipe organ or a steam train or synthesizer has anything musical at 20hz? Not a stand up bass or anything like a normal drum head or any orchestral instrument that I can think of. And yes I do still have a turntable and albums in my collection. Nothing like the rumble of a low note getting into feedback with the turntable and arm. I have been to many a live show behind the mixing console and I always see the low frequency cut set higher than 20hz. Bass feedback through the mics can be worse than a high frequency feedback if it just barely is perceptable. I can't even stand the sun roof open in some cars, the low frequency rumble will drive me nuts.
DF96
Probably agree with you (not an appreciable effect) but in the interest of turning all stones, a simple RC rolloff at 15Hz would leave roughly 20deg phase shift at 40Hz.
I originally just commented that one doesnt need to use these low F's for estimating ripple.
Thanks
-Antonio
Just to add a tangent to where the conversation is at the moment, because my belief is that the perceived quality of an amp is due to how well it handles the high frequency stuff, and also that I wanted to show a simple technique for comparing different approaches, here's a variation on Tom's latest test jig:
Note that the original circuit still stands, but I've duplicated the circuit loading the power supply, hooked that one up to perfect voltage sources, and am now driving both versions with the same signal. This makes it easy to compare one scenario with another, without having to worry about comparing the output to the input. And here are the results, driven by a full bore 20KHz signal:
The top waveform, which shows the difference between the signals across the two versions' loads indicates where the problem is using class AB with real rather than theoretical supplies: nasty glitches, at very high frequencies, at the time when the current draws through the active devices switch off, ie. crossover artifacts. This of course will be supposedly corrected by feedback, but if it can't quite do the job properly because the supply is glitching at the time, then things might start looking, and sounding, worse ...
Frank
Note that the original circuit still stands, but I've duplicated the circuit loading the power supply, hooked that one up to perfect voltage sources, and am now driving both versions with the same signal. This makes it easy to compare one scenario with another, without having to worry about comparing the output to the input. And here are the results, driven by a full bore 20KHz signal:
The top waveform, which shows the difference between the signals across the two versions' loads indicates where the problem is using class AB with real rather than theoretical supplies: nasty glitches, at very high frequencies, at the time when the current draws through the active devices switch off, ie. crossover artifacts. This of course will be supposedly corrected by feedback, but if it can't quite do the job properly because the supply is glitching at the time, then things might start looking, and sounding, worse ...
Frank
Last edited:
How much more capacitance than basic?
Power supply board has so many similarities to an output cap and the roll off is not harmless. Unfortunately, I fear that speaker support requirement of the power board has inflated and therefore hindered the capacitance per ampere calculation efforts. Problem: Up to a certain point (it is speaker support) the power supply capacitance is the same for 1 watt amplifier as for 1000 watt amplifier.
There's 2 tasks for the power supply reservoir. Can we calculate them individually?
To me, it is nicer to contemplate a fixed basic power supply that accomplishes the speaker support and then afterwards the super interesting part becomes how much more capacitance must we add per higher power amplifier? That question is a bit different.
My apologies if the prospect was unsuitable.
Other way to go is to try calculating for 10,000 watt amplifier since that power supply could so easily swamp the speaker support minimum requirements.
Power supply board has so many similarities to an output cap and the roll off is not harmless. Unfortunately, I fear that speaker support requirement of the power board has inflated and therefore hindered the capacitance per ampere calculation efforts. Problem: Up to a certain point (it is speaker support) the power supply capacitance is the same for 1 watt amplifier as for 1000 watt amplifier.
There's 2 tasks for the power supply reservoir. Can we calculate them individually?
To me, it is nicer to contemplate a fixed basic power supply that accomplishes the speaker support and then afterwards the super interesting part becomes how much more capacitance must we add per higher power amplifier? That question is a bit different.
My apologies if the prospect was unsuitable.
Other way to go is to try calculating for 10,000 watt amplifier since that power supply could so easily swamp the speaker support minimum requirements.
Last edited:
I think that a 10,000 watt amp is not practical. How about designing for the max size amp that can be built using a standard 110 volt 15 amp source of power. That amp should more than satisfy the normal requirements for home listening even if it is a mono amp for a sub woofer. That to me is just what is practical. The previous chart is also very helpful.
All in all this has been a very enlightening thread, I certainly learned a lot.
All in all this has been a very enlightening thread, I certainly learned a lot.
Krisfr,
That statement is the most poignant statement I have seen in this thread so far. What use is designing a power supply that you don't have the current to supply. The upper limit may be a 20 amp circuit though if you use a dedicated electrical outlet if we are talking 110v. You could also look at a 220v limit if you used a single phase system with both legs of the power. But what do we need even a thousand watt amp in our homes for? You could go and get a PA amplifier with a SMPS and get more than you could ever use but the sound of most of those amps are not exactly hi-fi in nature. I would be interested to see what are the upper limits with the limitation of supply current and voltage.
That statement is the most poignant statement I have seen in this thread so far. What use is designing a power supply that you don't have the current to supply. The upper limit may be a 20 amp circuit though if you use a dedicated electrical outlet if we are talking 110v. You could also look at a 220v limit if you used a single phase system with both legs of the power. But what do we need even a thousand watt amp in our homes for? You could go and get a PA amplifier with a SMPS and get more than you could ever use but the sound of most of those amps are not exactly hi-fi in nature. I would be interested to see what are the upper limits with the limitation of supply current and voltage.
an 8KW amplifier is not impossible in a household setting.
You will need to draw power from a separate fuse on the fusebox and dedicated cables.
the other question though is... WHY? WHY would you NEED that sort of power? And where would you find the speakers to handle it? I want one person, (deaf or otherwise, it doesn't matter, at those SPLs, the small veins in your lungs will resonate, along with your cerebral cavity) to claim that they actually listened to 1KW going into one speaker and they were in the same room.
You will need to draw power from a separate fuse on the fusebox and dedicated cables.
the other question though is... WHY? WHY would you NEED that sort of power? And where would you find the speakers to handle it? I want one person, (deaf or otherwise, it doesn't matter, at those SPLs, the small veins in your lungs will resonate, along with your cerebral cavity) to claim that they actually listened to 1KW going into one speaker and they were in the same room.
What's the problem?
If the maximum output power of the installed multichannel system is say 10kW, then the average output power will be around 100W total, for all the channels.
If the amplifiers, taking account of the quiescent currents and the transformers losses, make up for another 100W, then on a 240Vac supply the average current draw from the wall socket is ~ 800mAac.
As I said, what is the problem?
If the maximum output power of the installed multichannel system is say 10kW, then the average output power will be around 100W total, for all the channels.
If the amplifiers, taking account of the quiescent currents and the transformers losses, make up for another 100W, then on a 240Vac supply the average current draw from the wall socket is ~ 800mAac.
As I said, what is the problem?
I don't know what you mean by 'speaker support'. Perhaps LF rolloff at the output?danielwritesbac said:
The PSU caps don't behave like an output cap, because the PSU caps get (approximately) clamped to a fixed voltage 100/120 times a second - this makes them behave like a low value resistor rather than a cap, for LF signal voltages.
I answer: what is the power for listening at home, I mean a lover of good music as audiophile.
well, in my living room 8x4mt with good absorption (carpets, curtains and sofas), I wanted to measure the current absorption, while listening to a song composed by bass, trumpet, drums, female voice. speaker a pair of B & W 802 D2.
apparently the volume looks very clean about 80w, bass enveloping everything very well defined. you want to know how much power was out? repetitive peak 10Amp per channel, with impedance around 3.3R on the load. this was part of some tests to see how much power without thd, must be able to provide an excellent high end amplifier. is very insignificant power to declare the 8R or 4R on an amplifier. the amplifier under test can deliver 750w (2R) with 0.05% thd.
then, if I produce an amplifier for audiophili, it is obvious that I have to declare the power and performance, the lowest impedance obtained.
apparent, this amplifier have only 200w at 8R. so it looks like a small amp.
but in reality, everything is designed for high current output clean.
So, what should be dimensioned power supply?
well, in my living room 8x4mt with good absorption (carpets, curtains and sofas), I wanted to measure the current absorption, while listening to a song composed by bass, trumpet, drums, female voice. speaker a pair of B & W 802 D2.
apparently the volume looks very clean about 80w, bass enveloping everything very well defined. you want to know how much power was out? repetitive peak 10Amp per channel, with impedance around 3.3R on the load. this was part of some tests to see how much power without thd, must be able to provide an excellent high end amplifier. is very insignificant power to declare the 8R or 4R on an amplifier. the amplifier under test can deliver 750w (2R) with 0.05% thd.
then, if I produce an amplifier for audiophili, it is obvious that I have to declare the power and performance, the lowest impedance obtained.
apparent, this amplifier have only 200w at 8R. so it looks like a small amp.
but in reality, everything is designed for high current output clean.
So, what should be dimensioned power supply?
Last edited:
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.