This was very helpful, Thanks
It took me awhile to find the right recording to get a truthful answer. I ended up using the VU meter in foobar to help me find a recording that would use the dynamic range available in my PC-based pre-amp.
My favorite recordings, although recorded relatively dynamic, peaked around -6dB, resulting in RMS readings of 18 volts, the full RMS voltage of my chip amps! DVD movies fall into this catagory also.
If I played compressed internet radio music, I got RMS readings at 6 volts.
Once I found a recording that used the full gain of the pre-amp without clipping, I got a reading of 9.66 volts. So I voted for the 5-10 group.
I am sharing the above information, to hopefully help others realize the importance of playing the right recording to get the most helpful answer out of this test. It would be a shame to build a 200 watt amp and not "need" it!
My system is as follows:
Alpair 12s in Avebury cabinet - Full-range 89 dB efficent, 8 ohm
PC based playback system - Realtek audio (not the best)
Peter Daniel's standard LM3875 chip amps (18 volt RMS)
My room is 10ft X 14ft with openings on the sides.
Conclusions:
I need a bigger amp, at least for Avebury. Based on the math given above, i need a 94.74 watt amp, lets say 100 watt amp to run this system in this room, with dynamic material, esp movies. The chip amps are my first amp build, so... But, at least I know why they were chipping and what to shoot for next, atleast in requards to Avebury.
The Chip amps would be suited for a more sensitive speaker system, say 94dB or above?!? The future build perhaps...
It is safe to admit at this point that I like it Loud, when it is clear and fatigue free.
Allen
It took me awhile to find the right recording to get a truthful answer. I ended up using the VU meter in foobar to help me find a recording that would use the dynamic range available in my PC-based pre-amp.
My favorite recordings, although recorded relatively dynamic, peaked around -6dB, resulting in RMS readings of 18 volts, the full RMS voltage of my chip amps! DVD movies fall into this catagory also.
If I played compressed internet radio music, I got RMS readings at 6 volts.
Once I found a recording that used the full gain of the pre-amp without clipping, I got a reading of 9.66 volts. So I voted for the 5-10 group.
I am sharing the above information, to hopefully help others realize the importance of playing the right recording to get the most helpful answer out of this test. It would be a shame to build a 200 watt amp and not "need" it!
My system is as follows:
Alpair 12s in Avebury cabinet - Full-range 89 dB efficent, 8 ohm
PC based playback system - Realtek audio (not the best)
Peter Daniel's standard LM3875 chip amps (18 volt RMS)
My room is 10ft X 14ft with openings on the sides.
Conclusions:
I need a bigger amp, at least for Avebury. Based on the math given above, i need a 94.74 watt amp, lets say 100 watt amp to run this system in this room, with dynamic material, esp movies. The chip amps are my first amp build, so... But, at least I know why they were chipping and what to shoot for next, atleast in requards to Avebury.
The Chip amps would be suited for a more sensitive speaker system, say 94dB or above?!? The future build perhaps...
It is safe to admit at this point that I like it Loud, when it is clear and fatigue free.
Allen
Allen. Thanks for your results.
Yes, if you got 9.66V from the test tones at your volume setting then I'd say you need at least an amp rated at 95 WPC 8 ohms to keep from clipping. Of course most of the musical signal is far below that, but you still need a 95 watt amp not to crush your >38V peaks.
Yes, if you got 9.66V from the test tones at your volume setting then I'd say you need at least an amp rated at 95 WPC 8 ohms to keep from clipping. Of course most of the musical signal is far below that, but you still need a 95 watt amp not to crush your >38V peaks.
Thanks for doing this pano.
My system is:
Acoustat 2+2 speakers
Revel B12a subwoofer
dcB1/ldr as a preamp
Modified Behringer DEQ2496 as a dac
mpd on ubuntu on spdif output
everything but your test files are flacs
gmpc client
I used folk/rock cuts to get to the "max" level. Since the Acoustats are full range I measured both frequencies, with the sub in and out:
120Hz (sub turned off) - 11.34v
220Hz (sub on) - 11.5v
My system is:
Acoustat 2+2 speakers
Revel B12a subwoofer
dcB1/ldr as a preamp
Modified Behringer DEQ2496 as a dac
mpd on ubuntu on spdif output
everything but your test files are flacs
gmpc client
I used folk/rock cuts to get to the "max" level. Since the Acoustats are full range I measured both frequencies, with the sub in and out:
120Hz (sub turned off) - 11.34v
220Hz (sub on) - 11.5v
Yes, the PS Audio is rated at 400W @4 ohms... Not the best sounding, but it doesn't complain when driving the Acoustats to loud levels. I also have them in a large room and sit about 13' from them.
Pano - I just found this fascinating thread as a result of reading one of your posts in the Class D thread and seeing the footnote about it.
I did the test yesterday. Our system (the part used for the test) is a pair of upgraded 1958 Klipschorns driven by an old Harman Kardon 430 SS amp. (For some reason the Khorns love that amp!) All of our sources are on our HTPC that contains an Asus Essence ST soundcard that feeds the power amps directly. We use JRiver MC17 as a player and use its internal 64 bit volume control.
I used Steely Dan Two Against Nature for the reference music. I downloaded your test tones and used the 220 HZ tone.
Played the Steeley Dan about as loud as we would care to listen to it in our 1755 cubic foot room, which BTW is quite small for KHorns. Got 2.51 volts on a DMM.
The Khorn is rated as an 8 ohm speaker. Based on that our amp was putting out 3/4 watt and x 4 = 3 watt peaks. I believe the Khorn bass bin is closer to 6 ohms at that frequency so that would be 1 watt from the amp and 4 watt peaks. No wonder Paul Klipsch liked to say "It's the first watt that counts!" The Khorns are 104 DB sensitive speakers.
I also used a Radio Shack analog Sound Level Meter at the LP which is 10 feet from the Khorns. Got 94 DB on the reference music and 90 DB on the Sine Wave.
In reviewing the test I did as I write this I think I made some mistakes. I had both stereo speakers going and our Danley DTS-10 sub on. Don't think the sub was doing anything @220, but it seems a proper test to know how much power (voltage, SPL whatever you want to call it) would be to only have the one channel driven that is being used for the test. I could certainly have played it at a higher voltage if only the one speaker was on.
So my figures may be (probably are) flawed. I haven't heard this issue discussed, but it may be that I'm the only one here that ran both speakers when only testing one! I'll run the test again with only the one side on, but it may be awhile.
Also I have in storage a very old Tektronix 453 scope. If it still works, and is still in any kind of calibration, it will be very interesting to see at what point clipping occurs on the old HK amp that is rated for 30 WPC. It is though a very honestly rated amp and has dual power supplies. Of course with the Khorns in our small room I might need ear protection when I run enough power to drive it to clipping.
Pano I want to thank you very much for devising this thoroughly ingenious test. I think it makes a lot of sense and is certainly something that most anyone can do without any fancy test equipment.
Rod
I did the test yesterday. Our system (the part used for the test) is a pair of upgraded 1958 Klipschorns driven by an old Harman Kardon 430 SS amp. (For some reason the Khorns love that amp!) All of our sources are on our HTPC that contains an Asus Essence ST soundcard that feeds the power amps directly. We use JRiver MC17 as a player and use its internal 64 bit volume control.
I used Steely Dan Two Against Nature for the reference music. I downloaded your test tones and used the 220 HZ tone.
Played the Steeley Dan about as loud as we would care to listen to it in our 1755 cubic foot room, which BTW is quite small for KHorns. Got 2.51 volts on a DMM.
The Khorn is rated as an 8 ohm speaker. Based on that our amp was putting out 3/4 watt and x 4 = 3 watt peaks. I believe the Khorn bass bin is closer to 6 ohms at that frequency so that would be 1 watt from the amp and 4 watt peaks. No wonder Paul Klipsch liked to say "It's the first watt that counts!" The Khorns are 104 DB sensitive speakers.
I also used a Radio Shack analog Sound Level Meter at the LP which is 10 feet from the Khorns. Got 94 DB on the reference music and 90 DB on the Sine Wave.
In reviewing the test I did as I write this I think I made some mistakes. I had both stereo speakers going and our Danley DTS-10 sub on. Don't think the sub was doing anything @220, but it seems a proper test to know how much power (voltage, SPL whatever you want to call it) would be to only have the one channel driven that is being used for the test. I could certainly have played it at a higher voltage if only the one speaker was on.
So my figures may be (probably are) flawed. I haven't heard this issue discussed, but it may be that I'm the only one here that ran both speakers when only testing one! I'll run the test again with only the one side on, but it may be awhile.
Also I have in storage a very old Tektronix 453 scope. If it still works, and is still in any kind of calibration, it will be very interesting to see at what point clipping occurs on the old HK amp that is rated for 30 WPC. It is though a very honestly rated amp and has dual power supplies. Of course with the Khorns in our small room I might need ear protection when I run enough power to drive it to clipping.
Pano I want to thank you very much for devising this thoroughly ingenious test. I think it makes a lot of sense and is certainly something that most anyone can do without any fancy test equipment.
Rod
Hi NW boater. Sorry I missed your reply back on the 1st.
You should be running both speakers to set your loud listening level, unless you are in the habit of listening to only one at a time.
The test tells you how much voltage you need per channel. It would not be right to set your loudest level using only one speaker if you normally listen to two. The voltage on the other channel should be the same, but of course power will be doubled because you are running twice the number of channels.
You should be running both speakers to set your loud listening level, unless you are in the habit of listening to only one at a time.
The test tells you how much voltage you need per channel. It would not be right to set your loudest level using only one speaker if you normally listen to two. The voltage on the other channel should be the same, but of course power will be doubled because you are running twice the number of channels.
Siberia - your numbers look about right. If you got 3.2V on the test tone, then you should be hitting just over 20W peak into a nominal 8 ohm load. An amp that is honestly rated at about 10WPC into 8 ohms should just be clipping. My guess is you'd want an amp of about 20W RMS to keep well clear of clipping.
An amp that low? Couldn't believe this, so, I thought, my triamping binding posts are playing games on me. But I've measured all 3 of them and got same results.
Now this is the math I'm doing:
Peak voltage: 3.2 x 2.83=9.06V
W = V2 / R = 82.08 / 5.7 R (minimum impedance vs. nominal 8 ohms) = 14.40W
That is per channel, right? I would need 30W amp then. And that's all gain I need with my DCB1 buffer! Note the size of my room. It's rather huge.
Thank you for the nice test, BTW.
Now this is the math I'm doing:
Peak voltage: 3.2 x 2.83=9.06V
W = V2 / R = 82.08 / 5.7 R (minimum impedance vs. nominal 8 ohms) = 14.40W
That is per channel, right? I would need 30W amp then. And that's all gain I need with my DCB1 buffer! Note the size of my room. It's rather huge.
Thank you for the nice test, BTW.
effective maximum power is V^2 / Rload.
Your Rload is the nominal impedance of your speaker.
The V is the AC voltage presented to the load.
But this "test" is arranged to predict your maximum power by cancelling the nominal Rload with the -9dB signal level.
Pmax = 3.2^2 = 10.2W
The amplifier should be designed to deliver the extra current to meet the speaker demand at any combination of frequencies and signal shape that come along.
There is no need to add a further 3dB of overhead. This "test" is designed to indicate the maximum signal that can ever be sent to the speaker, for that combination of digital source and speaker and room and the loudest you are ever likely to need.
So just buy yourself an 11W amplifier and be happy, that you can never clip any music signal.
Yeah, tell me another !
Your Rload is the nominal impedance of your speaker.
The V is the AC voltage presented to the load.
But this "test" is arranged to predict your maximum power by cancelling the nominal Rload with the -9dB signal level.
Pmax = 3.2^2 = 10.2W
The amplifier should be designed to deliver the extra current to meet the speaker demand at any combination of frequencies and signal shape that come along.
There is no need to add a further 3dB of overhead. This "test" is designed to indicate the maximum signal that can ever be sent to the speaker, for that combination of digital source and speaker and room and the loudest you are ever likely to need.
So just buy yourself an 11W amplifier and be happy, that you can never clip any music signal.
Yeah, tell me another !
are you sure?Andrew it has been previously acknowledge that you should times the final Power Numbers by 4...
Andrew and Wayne, I'm sorry if you don't understand the test, Ohms Law or both. It's been explained in this thread over and over. Not much left to be said about that.
Now; It is possible that folks are measuring wrong, or doing that math wrong, but that's a different matter.
Siberia. The test is super simple. Use dynamic piece of music and turn it up as loud as you ever do. Keep that volume setting when playing the test tone(s). Make you voltage measurement. Your highest peak voltage will be 4X what you measured on the tone for that digital source and volume setting. Highest RMS voltage (sine wave) will be 2.82X what you measured.
The confusion comes because amps are rated in Watts, not Volts. I think it would be much easier for all of us if they were rated in Volts. You can figure backward using Ohm's Law to find an amp power rating that will give you the peak or RMS voltage you need.
Now; It is possible that folks are measuring wrong, or doing that math wrong, but that's a different matter.
Siberia. The test is super simple. Use dynamic piece of music and turn it up as loud as you ever do. Keep that volume setting when playing the test tone(s). Make you voltage measurement. Your highest peak voltage will be 4X what you measured on the tone for that digital source and volume setting. Highest RMS voltage (sine wave) will be 2.82X what you measured.
The confusion comes because amps are rated in Watts, not Volts. I think it would be much easier for all of us if they were rated in Volts. You can figure backward using Ohm's Law to find an amp power rating that will give you the peak or RMS voltage you need.
Pano,
why are you apologising to me on the grounds that I don't understand the test you have proposed?
It is a very clever and very easy to carry out the test to your method (if one can create the CD required for the test signal).
It is also very easy to determine the peak voltages that will be sent to the speaker and further it is very easy to determine the maximum power required of the amplifier into the nominal speaker impedance.
The fact that I don't agree that the test delivers what is promised has nothing to do with my understanding of the test method.
I would much rather be looking at a range of amplifiers stated and compared by voltage output into a specified load.
For example, is this 20V amplifier similar, or different from the other 25V amplifier when both have the same specified loading?
why are you apologising to me on the grounds that I don't understand the test you have proposed?
It is a very clever and very easy to carry out the test to your method (if one can create the CD required for the test signal).
It is also very easy to determine the peak voltages that will be sent to the speaker and further it is very easy to determine the maximum power required of the amplifier into the nominal speaker impedance.
The fact that I don't agree that the test delivers what is promised has nothing to do with my understanding of the test method.
I agree absolutely on this.
I would much rather be looking at a range of amplifiers stated and compared by voltage output into a specified load.
For example, is this 20V amplifier similar, or different from the other 25V amplifier when both have the same specified loading?
Last edited:
If you have set your volume correctly (loud as you ever do on a dynamic recording) and you measured 3.2V RMS on the test tone then 9.1V is the highest RMS voltage you would see out that amp at that volume setting. That is 9dB RMS higher than the test tone. Peak would be just under 13V - or 12dB higher peak.
Since amps are normally rated for RMS power (not peak) you can figure backward for amp power rating. You need 9.1V RMS. Into 8 ohms that would be 10.35 watts. An amplifier that is clean (not clipping, not distorted) at 12 Watts would be enough. That's 12 watts per channel.
That said, you may like the sound of a 25 WPC amp better, just because it is better built, has a bigger power supply or other reasons. The difference in watts is double, but the RMS output voltage is just 14.14V. Not a big difference when you look at the voltage, is it?
You have to decide if an amp that claims 12WPC can really deliver that power cleanly. You'd want to see test results, for sure. Or just get a more powerful amp and hope it can hit the 10-12 watts you need without straining.
OK, I see your point, thanks.
Agree, not all amps with the same power rating are created equal! Power supply makes a big difference, as well as topology. I like amps with over-sized power supplies, they have a lot of sonic advantages to my ears.
But let us not forget - this test is done at what we hope is the maximum volume level the listener will ever use. Everything else will be below that. It simply meant to give you an idea of how much power you use when playing your system at its loudest. From that you can get a general idea if you have enough amp or not.
I'll take that voltage example a little further.
three different amplifiers, specified by voltage and loading
Amp 1 . . . . Amp 2 . . . . Amp 3
20V 8r0 . . . 21V 8r0 . . . 25V 8r0
19V 4r0 . . . 19V 4r0 . . . 21V 4r0
17V 2r0 . . . 16V 2r0 . . . 15V 2r0
minimum speaker load
4ohms . . . . 6ohms . . . . 4ohms
I can see at a glance that Amp 3 cannot properly drive a 4ohms speaker.
I can also see at a glance that Amp 1 manages 4r0 better than Amp 2 and that the recommended 6ohms for Amp 2 makes sense, i.e. both manufacturers are being honest.
three different amplifiers, specified by voltage and loading
Amp 1 . . . . Amp 2 . . . . Amp 3
20V 8r0 . . . 21V 8r0 . . . 25V 8r0
19V 4r0 . . . 19V 4r0 . . . 21V 4r0
17V 2r0 . . . 16V 2r0 . . . 15V 2r0
minimum speaker load
4ohms . . . . 6ohms . . . . 4ohms
I can see at a glance that Amp 3 cannot properly drive a 4ohms speaker.
I can also see at a glance that Amp 1 manages 4r0 better than Amp 2 and that the recommended 6ohms for Amp 2 makes sense, i.e. both manufacturers are being honest.
Last edited: