When I was developing my current amp I had a pair of Celestion SL6 speakers (the first with a copper dome tweeter) that I used for all my testing, and which had an efficiency of just 82db/watt. I used these for many years on a Pioneer A80 amp (150w rms/8 ohm). The all metal (aerolam) version was the SL600 at more than 2.5 times the cost.
Celestion SL6
Anyhow, the real point is that I wanted to know for sure "how much power I really needed" when designing the new amp. What voltage rails to use and so on. The only way was to scope the speaker drive looking at transients and see for sure. And I was playing this LOUD to really push it and I do remember seeing peaks of 40 volts plus although this was at levels way beyond those I would ever normally listen at. The speaker would have handled more but I decided that -/+48 volt rails was certainly sufficient.
I bought the speakers from a local store (where an ex colleague worked) and I remember us hanging the SL6's the end of a "Doctor Thomas" amp and really cranking it up. What was interesting was that the SL6's despite their low efficiency were also more than adequate for all normal listening on what would be considered quite low powered gear. My parents had the SL6's on the end of an original Rotel RA820 and it was great for normal listening.
There is no "magic" the power required depends in the speakers. every 3dB difference in speaker sensitivity means your need to double or half the amp's power.
So a 4W amp plying through a 100dB speaker is the same as an 8W amp and a 98db
So I can make a table of systems that _should_ sound alike
5W, 95db
10W, 92dB
20W, 89db
40W, 86db
80W, 83db
Does this seem right be most of you? Would they sound alike?
In theory it should but many people argue that the 5W, 95db system sounds best even if they all have the same loudness. My theory is that either the small amps have better dynamic range or the 95db speakers have lower distortion and sound cleaner. Or likely both.
So a 4W amp plying through a 100dB speaker is the same as an 8W amp and a 98db
So I can make a table of systems that _should_ sound alike
5W, 95db
10W, 92dB
20W, 89db
40W, 86db
80W, 83db
Does this seem right be most of you? Would they sound alike?
In theory it should but many people argue that the 5W, 95db system sounds best even if they all have the same loudness. My theory is that either the small amps have better dynamic range or the 95db speakers have lower distortion and sound cleaner. Or likely both.
Yes, that's a nice chart.
Power compression is an important factor and affects the sound. Power compression is very small for high efficiency speakers in a domestic setting. That can sure help dynamics.
It's been discussed a lot in other threads, but is worth mentioning here.
I think it's also easier to make and "overbuilt" 10W amp than an overbuilt 200W amp.
Power compression is an important factor and affects the sound. Power compression is very small for high efficiency speakers in a domestic setting. That can sure help dynamics.
It's been discussed a lot in other threads, but is worth mentioning here.
I think it's also easier to make and "overbuilt" 10W amp than an overbuilt 200W amp.
It varies with frequency. Horns and lines initially fall off much slower, but eventually adhere to the 6 dB rule. Bottom line though is that to ensure enough power handling is to use the rule.
WRT to dynamic headroom, the article's +15 dB isn't near enough for HT's +20 dB and if you listen to some symphonic music CDs, especially the better pipe organ ones you need up to +30 dB and even then the CD's signal is clipped, so can't go 100% 'live' no matter how big/powerful/eff. your system is.
GM
WRT to dynamic headroom, the article's +15 dB isn't near enough for HT's +20 dB and if you listen to some symphonic music CDs, especially the better pipe organ ones you need up to +30 dB and even then the CD's signal is clipped, so can't go 100% 'live' no matter how big/powerful/eff. your system is.
GM
How do you know? Do you drive them to continuous clipping? At what average SPL at the listening position do you find music, etc. most enjoyable? 'Loud' is controlled by whoever has access to the volume control and how much distortion they are comfortable with.
Regardless, as already noted, most of the amp's power is used up properly tracking fast transients which isn't even perceived as 'loud', just more 'air' added around the instruments, so depending on the source material and distance from the speakers, you may be needing more than 100 W just to listen at Dolby's modest spec'd 78 dB/listening position average at low distortion for 'made for TV' movies.
GM
And then there is this other thing: stereo uses two speakers. So, a pair of 82 dB at 1 W at 1 meter would produce ... not 85 dB, but quite a bit more. The reason is that adding two incoincedent sound sources of equal strength adds 3dB SPL, but two coincedent sound sources add up to 6dB extra. Music is a mix of coincident and incoincedent sound over both speakers. Since most of the coincident sound is in the bass area, which requires by far the most part of the power to produce, in real life that pair of 82dB sensitivity might produce a total SPL of 86 or 88 dB when both speakers are fed with one watt.
In order to prevent being caught on an apparent logical inconsistency the following. After all, when you feed one speaker with 1 watt, you need to double this for 3dB extra. How can it be, that if you feed this additional 1 watt into an additional speaker, you could get 6dB more? It is somewhat counterintuitive, but the reason is that with two speakers you double Vas. Feed that into the formula to calculate loudspeaker efficiency, and presto. A doubling of Vas leads to 3 dB additional sensitivity.
vac
In order to prevent being caught on an apparent logical inconsistency the following. After all, when you feed one speaker with 1 watt, you need to double this for 3dB extra. How can it be, that if you feed this additional 1 watt into an additional speaker, you could get 6dB more? It is somewhat counterintuitive, but the reason is that with two speakers you double Vas. Feed that into the formula to calculate loudspeaker efficiency, and presto. A doubling of Vas leads to 3 dB additional sensitivity.
vac
A long story ruined by the omission of stating the restricted bandwidth over which your +6dB would hold true.
If the speakers are 2m apart, what is the bandwidth that +6dB is available?
At what frequency will the stereo SPL gain have fallen to +4dB, or +3dB?
Well, ruined is not the word. The beauty is that in the long wavelengths, the L+R signals usually contain the same coincident signal. That is were your amplification runs out of juice soonest. It is also the bass notes that influence the dB measurements the most, because they contain the most energy. Therefore, the tables linked to in this thread, and which just double power for every 3dB increase, are incorrect, even in an anechoic chamber. These tables do not reflect how sound from two speakers combine in the real world when driven by a music signal.
At higher frequencies you are right in the sense that with speakers some distance apart, you will get an increasing number of lobes with rising frequency. In the central lobe, however, they would still add up to + 6dB with coincident sound sources. That is also the central idea behind the LR xover. Much of the stereo separation is in the frequencies above the low end, so the two speakers shift to becoming incoincident sound sources anyways with rising frequency.
vac
At higher frequencies you are right in the sense that with speakers some distance apart, you will get an increasing number of lobes with rising frequency. In the central lobe, however, they would still add up to + 6dB with coincident sound sources. That is also the central idea behind the LR xover. Much of the stereo separation is in the frequencies above the low end, so the two speakers shift to becoming incoincident sound sources anyways with rising frequency.
vac
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.