I would like to try tube amplification, but I am having a hard time to figure out how much power I really need. Look at my speakers description:
- 2 way using 3/4" tweeter Hiquphon OW1 and 6 1/2" woofer Seas Excel W18 with magnesium cone
- 83db sensitivity
- impedance: 8 ohms, but can go as low as 6.2 in the bass region (Attention - NOT LOWER)
- rear ported
- SPL
- Frequency Response: F3 is about 40hz in a 22 liter enclosure
My room: 18'x16' or (about 6x5 m). 2m tall.
The music that I listening at: almost ANYTHING
How loud I need to sound?
- ussually I am listening at 80db SPL (measured at 8' = 3m)
- when I liked it I am listening at 85db
- when I really want to cranck it: 90db is MAXIMUM that I would listed to.
Now, having ALL this technical considerations, I was not able to find someone or a formula that would allow me to compute how much power my speakers need to produce 90db at the listening position.
Is anyone out there that can help?
Thank you a LOT!
- 2 way using 3/4" tweeter Hiquphon OW1 and 6 1/2" woofer Seas Excel W18 with magnesium cone
- 83db sensitivity
- impedance: 8 ohms, but can go as low as 6.2 in the bass region (Attention - NOT LOWER)
- rear ported
- SPL
- Frequency Response: F3 is about 40hz in a 22 liter enclosure
My room: 18'x16' or (about 6x5 m). 2m tall.
The music that I listening at: almost ANYTHING
How loud I need to sound?
- ussually I am listening at 80db SPL (measured at 8' = 3m)
- when I liked it I am listening at 85db
- when I really want to cranck it: 90db is MAXIMUM that I would listed to.
Now, having ALL this technical considerations, I was not able to find someone or a formula that would allow me to compute how much power my speakers need to produce 90db at the listening position.
Is anyone out there that can help?
Thank you a LOT!
As far as I remember you get 3dB extra SPL when you double the power (someone correct me if I'm wrong...). So that means you would need approx. 4W for 90dB SPL.
BUT... The SPLs you have measured are probably average and not peaks, so you will need some more headroom for peaks. You would probably need 15-20W at least, but less may work. 3W is enough for me, and I don't have horn speakers...
Best regards,
Mikkel C. Simonsen
BUT... The SPLs you have measured are probably average and not peaks, so you will need some more headroom for peaks. You would probably need 15-20W at least, but less may work. 3W is enough for me, and I don't have horn speakers...
Best regards,
Mikkel C. Simonsen
There is a lot of personal taste involved in this, some would suggest you need to allow for head room, and that your 90dB target is too low. However, I think 90 dB is plenty, and so taking your figure of 90dB at 3 metres as the target, you need 90- 87 = 7 dB over 1 watt of power to get to 90dB at 1 metre, and then a certain amount of power over that to get to 90dB at 3 metres.
That "certain amount of power" is the tricky bit, due to the room characteristics. Some people would calculate that power at say 6 dB, but in my view with a room your size the only way to really know is to measure it. In my room, which is smaller than yours at 11 x 16 feet, the sound pressure at some spots 2 metres away is 3 dB Higher than 1 metre away from the speakers. In other words, forget about a formula. Add 3 dB "just in case", giving you a total of 10dB over 1 watt, or in other words, 10 watts power required.
As a sanity check for these calculations, I use a tubed receiver in the room, using 6BM8 output tubes good for perhaps 7 Watts, and I never have to turn up the volume control past 10 or 11 oclock. Your room is bigger, maybe you would feel better with 10 -15 watts
That "certain amount of power" is the tricky bit, due to the room characteristics. Some people would calculate that power at say 6 dB, but in my view with a room your size the only way to really know is to measure it. In my room, which is smaller than yours at 11 x 16 feet, the sound pressure at some spots 2 metres away is 3 dB Higher than 1 metre away from the speakers. In other words, forget about a formula. Add 3 dB "just in case", giving you a total of 10dB over 1 watt, or in other words, 10 watts power required.
As a sanity check for these calculations, I use a tubed receiver in the room, using 6BM8 output tubes good for perhaps 7 Watts, and I never have to turn up the volume control past 10 or 11 oclock. Your room is bigger, maybe you would feel better with 10 -15 watts
Another consideration is musical taste. Due to the insane amount of compression and limiting common in popular music 6 watts of Slipknot will be subjectively much louder than 6 of Haydyn's 'Surprise'. Recorded peak to average ratio is the name of the game. 92 dB average level with 20 dB 'classical' instantaneous peaks requires a lot of power.
The reallity is that I am a classical music lover. And indeed, I observed what you are saying.
When I was saying that I don't listen over 90db, I was refering to an average not to peaks. As you are saying, I measured peaks of as much as 115db!!!
This doesn't meand that I need 115db. I will be crazy to listen that loud.
When I was saying that I don't listen over 90db, I was refering to an average not to peaks. As you are saying, I measured peaks of as much as 115db!!!
This doesn't meand that I need 115db. I will be crazy to listen that loud.
> How much power do I really need?
How much power do you have NOW? Is it enough? Is it too much? Looking at an existing system is often a better bet than trying to calculate numbers out of thin air.
> having ALL this technical considerations
You are missing an important number: the amount of acoustic absorption in the room. That is much more important than room L*W*D.
However, there is a shortcut. For most "living rooms", the Critical Distance is about 4 feet. The SPL over most of the room is similar to the speaker's 4-foot rating, which is close to its 1 Meter rating.
> I am listening at 85db
Averaged or Peak? Actually, that "must" be an averaged reading: 85dB SPL peak is fairly soft, and 85dB SPL averaged has been the typical "good" listening level for half a century, now recognized in the Dolby references.
Peaks will be 10dB to 20dB higher. 10dB peaks are very common, 20dB peaks maybe once a week. 16dB is a good round number. If power is precious and its clipping is inoffensive, 15dB peak allowance is reasonable and gives a nice round "100dB SPL" for peaks.
> 83db sensitivity
Back when tubes ruled the world, we did NOT use 83dB speakers. 93 was common. AR was very bold to try to sell 88dB speakers. And a 6" speaker was a "good radio", not a High-Fidelity. What you got there is a speaker that leverages the very low cost/Watt of sand-state amplifiers to give modest size (about 1/8th the volume we used to use to reach 40Hz at reasonable efficiency).
That's a clue. You may want an amp MUCH bigger than we used in the Golden Classic Era of tube home hi-fi. While beasts like the MC-50 existed, there were many more 10W and 20W amps used.
83dB/Watt???? That must be 0.3% efficiency, or 299/300 IN-efficiency. We often ran 2% efficiency speakers with tube amps in living rooms.
OK then. 83dB/Watt at 3.3 feet is 82dB/Watt at 4 feet, or all over a typical living room. You need to cover peaks of 100dB SPL. 100dB-82dB= 18dB increase of power. You need 63 Watts. Taking your "cranked" number and full margin, you need 90dB+16dB-82dB= 24dB above 1 Watt, or 251 Watts. Given 2-channels, this could be 30W+30W to 125W+125W.
A 250W or 125+125W tube amp is actually bigger and heavier than one of your speakers. Inefficient speakers are smaller, except amp size can cancel the benefit. Maybe you have the space; there is also the huge cost of 250W of tube-amp compared to the modest cost of a sheet of plywood and a 15" woofer that will do 90dB SPL/Watt, making your acoustic goals with 10W-40W total tube power.
Cross-checks: a home hi-fi should deliver about a half-watt of acoustic power. At 0.3% efficiency, 158 Watts of electrical power, similar to the 60-250W calculated from SPL/W and typical living-room approximations.
AR suggested 10W-40W for the 88dB/W/4' AR-3. 82dB/W/4' needs 6dB more power: 40W-160W, again in the ballpark.
You can increase the acoustic level by pulling all the couches, carpets, drapes, bookshelves, etc out of the room and hardening the walls. An empty concrete cellar with one human body in it will have a critical distance near 1 foot, acoustic levels 10dB higher, 1/10th the amplifier power needed. However it will also have reverberation time so long that it will mask most music. OTOH, if you line all surfaces with thick fluffy pillows you can stretch the critical distance to the far wall, approximating the "outdoors" condition, but few people want that much fluff in the room.
How much power do you have NOW? Is it enough? Is it too much? Looking at an existing system is often a better bet than trying to calculate numbers out of thin air.
> having ALL this technical considerations
You are missing an important number: the amount of acoustic absorption in the room. That is much more important than room L*W*D.
However, there is a shortcut. For most "living rooms", the Critical Distance is about 4 feet. The SPL over most of the room is similar to the speaker's 4-foot rating, which is close to its 1 Meter rating.
> I am listening at 85db
Averaged or Peak? Actually, that "must" be an averaged reading: 85dB SPL peak is fairly soft, and 85dB SPL averaged has been the typical "good" listening level for half a century, now recognized in the Dolby references.
Peaks will be 10dB to 20dB higher. 10dB peaks are very common, 20dB peaks maybe once a week. 16dB is a good round number. If power is precious and its clipping is inoffensive, 15dB peak allowance is reasonable and gives a nice round "100dB SPL" for peaks.
> 83db sensitivity
Back when tubes ruled the world, we did NOT use 83dB speakers. 93 was common. AR was very bold to try to sell 88dB speakers. And a 6" speaker was a "good radio", not a High-Fidelity. What you got there is a speaker that leverages the very low cost/Watt of sand-state amplifiers to give modest size (about 1/8th the volume we used to use to reach 40Hz at reasonable efficiency).
That's a clue. You may want an amp MUCH bigger than we used in the Golden Classic Era of tube home hi-fi. While beasts like the MC-50 existed, there were many more 10W and 20W amps used.
83dB/Watt???? That must be 0.3% efficiency, or 299/300 IN-efficiency. We often ran 2% efficiency speakers with tube amps in living rooms.
OK then. 83dB/Watt at 3.3 feet is 82dB/Watt at 4 feet, or all over a typical living room. You need to cover peaks of 100dB SPL. 100dB-82dB= 18dB increase of power. You need 63 Watts. Taking your "cranked" number and full margin, you need 90dB+16dB-82dB= 24dB above 1 Watt, or 251 Watts. Given 2-channels, this could be 30W+30W to 125W+125W.
A 250W or 125+125W tube amp is actually bigger and heavier than one of your speakers. Inefficient speakers are smaller, except amp size can cancel the benefit. Maybe you have the space; there is also the huge cost of 250W of tube-amp compared to the modest cost of a sheet of plywood and a 15" woofer that will do 90dB SPL/Watt, making your acoustic goals with 10W-40W total tube power.
Cross-checks: a home hi-fi should deliver about a half-watt of acoustic power. At 0.3% efficiency, 158 Watts of electrical power, similar to the 60-250W calculated from SPL/W and typical living-room approximations.
AR suggested 10W-40W for the 88dB/W/4' AR-3. 82dB/W/4' needs 6dB more power: 40W-160W, again in the ballpark.
You can increase the acoustic level by pulling all the couches, carpets, drapes, bookshelves, etc out of the room and hardening the walls. An empty concrete cellar with one human body in it will have a critical distance near 1 foot, acoustic levels 10dB higher, 1/10th the amplifier power needed. However it will also have reverberation time so long that it will mask most music. OTOH, if you line all surfaces with thick fluffy pillows you can stretch the critical distance to the far wall, approximating the "outdoors" condition, but few people want that much fluff in the room.
practical tests
I used to run a 5 watt stereo valve amp with efficient speakers then moved to 10/12 watts and Transmission Line speakers in a room considerably bigger than yours with good results on classical.
On shifting to Solid State first was a 25watts amp, then a much better 100 watt amp.
However volume and distortion on peaks was never an issue with any of these setups.
Currently I am experimenting with an Amp3 T amp rated at 12watts/channel and 88db speakers in a room 10m x 6m x 2.5m, but with mostly hard surfaces. Maximum clean volume is in way in excess of my normal listening levels. What is of more interest is that the PSU is fused at 250ma, indicating an average power drain of less than 1.8 watts per channel at normal listening levels. If we assume 50% efficiency at low power outputs, (its probably less) that equates to less than 1 watt.
The additional power needed for headroom is a matter of opinion, but there is general agreement that it can be less for valves than SS. For a while I worked for a firm who manufactured 12w and 25w valve amps of very similar design. These were virtually indistinguishable on listening tests. Certainly 25 watts valve is more than enough to drive even the most inefficient speakers in a large room, and 10 watts more than adequate for most people.
I used to run a 5 watt stereo valve amp with efficient speakers then moved to 10/12 watts and Transmission Line speakers in a room considerably bigger than yours with good results on classical.
On shifting to Solid State first was a 25watts amp, then a much better 100 watt amp.
However volume and distortion on peaks was never an issue with any of these setups.
Currently I am experimenting with an Amp3 T amp rated at 12watts/channel and 88db speakers in a room 10m x 6m x 2.5m, but with mostly hard surfaces. Maximum clean volume is in way in excess of my normal listening levels. What is of more interest is that the PSU is fused at 250ma, indicating an average power drain of less than 1.8 watts per channel at normal listening levels. If we assume 50% efficiency at low power outputs, (its probably less) that equates to less than 1 watt.
The additional power needed for headroom is a matter of opinion, but there is general agreement that it can be less for valves than SS. For a while I worked for a firm who manufactured 12w and 25w valve amps of very similar design. These were virtually indistinguishable on listening tests. Certainly 25 watts valve is more than enough to drive even the most inefficient speakers in a large room, and 10 watts more than adequate for most people.
> that equates to less than 1 watt.
I don't follow your calculation, and low-level efficiency will be far below 50%, but yes: average electrical levels are rarely even 1 Watt on loud orchestral passages.
I used to do PA to crowds of 350 with 1A fuses in the speakers, and never blew them: less than 16 average electric watts for much more acoustic power than you ever need at home.
This affects the economic design of sand-state B-class speech/music amplifiers: you can heat-sink for the average power, not the maximum power. This is only a small benefit: Class-B efficiency sinks rapidly as output drops, so you still need considerable heatsinking. Marginal sinking leads to wide swings in temperature which eventually cracks the seals. And testing and especially US FTC rating require extended high-power operation not seen in normal use.
And input power on tube amps, even AB, tends to be nearly constant as output power varies. Some extreme cases vary current 5:1, but 2:1 is more likely, and many hifi amps draw nearly constant power.
But for clean reproduction of speech/music, we have to cover the peaks. On orchestral, the ratio of peak to average power over a whole work may exceed 24dB: if played at the 1 electric Watt level, we need 250 Watts to cover the few highest peaks. With typical (not low-efficiency) speakers, we would not want to average even 1 Watt over the work, so 100 Watts or less covers the peaks.
> power needed for headroom is a matter of opinion, but there is general agreement that it can be less for valves than SS.
Quite large amounts of clean clipping are inaudible. In classical music production I may find one or two single-cycle peaks 6dB higher than anything else. I routinely apply few-milliSecond fades to bring these in line without altering the timbre, to allow 6dB higher average levels on CD and on playback systems. Tube amps tend to clip gently; transistor amps tend to splatt. The clipping is subtle but the splatt is horrendous. So it isn't surprising that tube amps can run 3dB to 6dB into clipping, and that we can use amps of 1/2 to 1/4 the nominal power that we need in sand-state to avoid splatt.
> it seems that 35w will be just enough.
There will be some fine listening in 35W, even 15W, even with 83dB speakers. If you hang with live orchestra, then try to get the same level at home, you may have some audible peak-mashing in crescendos; however most people don't listen at home as loud as a full orchestra plays.
I don't follow your calculation, and low-level efficiency will be far below 50%, but yes: average electrical levels are rarely even 1 Watt on loud orchestral passages.
I used to do PA to crowds of 350 with 1A fuses in the speakers, and never blew them: less than 16 average electric watts for much more acoustic power than you ever need at home.
This affects the economic design of sand-state B-class speech/music amplifiers: you can heat-sink for the average power, not the maximum power. This is only a small benefit: Class-B efficiency sinks rapidly as output drops, so you still need considerable heatsinking. Marginal sinking leads to wide swings in temperature which eventually cracks the seals. And testing and especially US FTC rating require extended high-power operation not seen in normal use.
And input power on tube amps, even AB, tends to be nearly constant as output power varies. Some extreme cases vary current 5:1, but 2:1 is more likely, and many hifi amps draw nearly constant power.
But for clean reproduction of speech/music, we have to cover the peaks. On orchestral, the ratio of peak to average power over a whole work may exceed 24dB: if played at the 1 electric Watt level, we need 250 Watts to cover the few highest peaks. With typical (not low-efficiency) speakers, we would not want to average even 1 Watt over the work, so 100 Watts or less covers the peaks.
> power needed for headroom is a matter of opinion, but there is general agreement that it can be less for valves than SS.
Quite large amounts of clean clipping are inaudible. In classical music production I may find one or two single-cycle peaks 6dB higher than anything else. I routinely apply few-milliSecond fades to bring these in line without altering the timbre, to allow 6dB higher average levels on CD and on playback systems. Tube amps tend to clip gently; transistor amps tend to splatt. The clipping is subtle but the splatt is horrendous. So it isn't surprising that tube amps can run 3dB to 6dB into clipping, and that we can use amps of 1/2 to 1/4 the nominal power that we need in sand-state to avoid splatt.
> it seems that 35w will be just enough.
There will be some fine listening in 35W, even 15W, even with 83dB speakers. If you hang with live orchestra, then try to get the same level at home, you may have some audible peak-mashing in crescendos; however most people don't listen at home as loud as a full orchestra plays.
rmihai said:
It won't say it will be much better. The difference between 35 and 60 watts is under 3 dB and not far from the minimum audible. In your shoes I would weigh condition, cosmetics, absence of mods, certainly price and sonic reputation as heavily as the power difference. (65 watts should do fine.)
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.