I don't believe it. I need more power?

[Irvrobinson]

I'm sitting here listening to one of my favorite rock albums, California Project, by Papa Doo Run Run, on Telarc. Beach Boys / Jan & Dean covers. The speakers are my trusty Revel Salon 2s crossed over to a Velodyne DD18+ at 60Hz low-pass, 80Hz high-pass, with the Revels powered by four channels of an ATI AT3005 plugged into a dedicated 20amp circuit. I just don't believe it, the sound is so clean and perfect, but the red "Peak" LED on the amp is lighting up momentarily now and then. ATI says of that light:

This indicator circuit continually monitors the output level of
your amplifier. This light will come on if the amplifier exceeds
its maximum output capability on any channel. While this
will not harm the amplifier, it does warn of potential harm to
your loudspeakers if the light stays illuminated continuously
for more than ten seconds. The volume setting from your
preamp or source device must be turned down if this occurs.

I'm busting the chops of at least one 300w channel? It couldn't be. No wonder those 125w/ch Levinsons sounded weak.

There goes that LED again. I'll be damned.

 

[Josuah]

What volume are you listening at? The Revel Ultima Salon2 speakers have a sensitivity of 86dB/2.83V/m. But the impedance drops as low as 3 ohms and the phase deviates at some of those dips as well (Revel Ultima Salon2 loudspeaker Measurements) which may present a more difficult load, regardless of actual SPL.

 

[TLS Guy]

I'm sitting here listening to one of my favorite rock albums, California Project, by Papa Doo Run Run, on Telarc. Beach Boys / Jan & Dean covers. The speakers are my trusty Revel Salon 2s crossed over to a Velodyne DD18+ at 60Hz low-pass, 80Hz high-pass, with the Revels powered by four channels of an ATI AT3005 plugged into a dedicated 20amp circuit. I just don't believe it, the sound is so clean and perfect, but the red "Peak" LED on the amp is lighting up momentarily now and then. ATI says of that light:



I'm busting the chops of at least one 300w channel? It couldn't be. No wonder those 125w/ch Levinsons sounded weak.

There goes that LED again. I'll be damned.

Why are you surprised?

Lets take a look and see where the problems come in a four way speaker like that with passive crossovers.

I can easily run my friends 400 watt per channel Macs out of gas driving his B &W 800 Ds.

Let us just see why you actually need 3000 to 4000 watts, but that if you provided it your crossovers would smoke.

Now you have a speaker that has a sensitivity of 86.4 db 2.83 volts I meter.

At 90 Hz the impedance of your speaker is 3.7 ohms and it does not go significantly above four ohms until 600 Hz. So it is a low impedance speaker where the majority of the power is.

So your speakers will draw 2.16 watts at 90 Hz from the amp to produce 86.4 db at I meter, versus 1 watt if it were 8 ohms. So the low impedance in the range where the power is has doubled the power requirement for your speakers versus 8 ohm ones.

Now it gets worse, your speakers have fourth order crossovers, and at a low frequency. Your insertion loss is1 db per order under the best of circumstances, assuming the use of the highest quality inductors. Since the crossover is far lower than can be recommended in a passive design I suspect you power loss is around 5 db and may well be as high as 6 db.

So, slightly over half your amp power is heating up the crossover.

Now lets take a look at an active situation. Lets take my speakers.

The bass units are each 8 ohm, since there are two drivers of 8 ohms, each directly coupled to an amp sensitivity is 93 db 2.83 volts 1 meter and that will take one watt of power. You will have to provide at least 8.4 watts of power to achieve the same spl, and more likely 10 watts.

Now even with that situation I still need to provide 750 watts per speaker, to keep things relaxed at concert levels.

That is why good monitor companies like ATC and PMC provide about 3 to 4 KW per speaker and that is with active crossovers.

Billy woodman showed me that most amps driving a passive three way speaker of average sensitivity are in almost continuous recovery mode.

What is required is active speakers with wide bandwidth drivers of good sensitivity.

I no longer believe it is possible to power a three way speaker to concert levels without horn drivers.

The time is now to ditch passive crossovers.

 

[cpp]

Turn the volume down and save your money.

 

[PENG]

Your insertion loss is1 db per order under the best of circumstances, assuming the use of the highest quality inductors. Since the crossover is far lower than can be recommended in a passive design I suspect you power loss is around 5 db and may well be as high as 6 db.

So, slightly over half your amp power is heating up the crossover.

I would think that any insertion loss or other losses have already been factor in when Revel specified xdB/2.83V/m.

 

[TLS Guy]

I would think that any insertion loss or other losses have already been factor in when Revel specified xdB/2.83V/m.

The issue is that that sensitivity is low, but when you have a speaker that has a low impedance you have to be aware how much the the speaker draws from the amp.

So if you quote sensitivity at 2.83 volts 1 meter, then a four ohm speaker draws twice as much power from the amp as the 8 ohm one.

Now if you quote sensitivity at 1 watt 1 meter, then the four ohm speaker will be 3 db less sensitive than the 8 ohm speaker.

So his speakers are 83.4 db 1 watt 1 meter is sensitivity.

Years ago all speakers were quoted at 1 watt, 1 meter, which is the honest watt to do it, The 2.83 volts 1 meter was introduced to pull the wool over the eyes of the consumer.

The 1 watt, 1 meter and 2.83 volts 1 meter give the same sensitivity reading at 8 ohms, as 2.83 volts produces 1 watt with an 8 ohm load.

The other issue is that power lost in the passive crossover is power not available to the drivers, and I was using that to illustrate the waste of amp power and pointing out that if it were an active design, then the sensitivity and spl would go up five to six db.

 

[PENG]

I'm busting the chops of at least one 300w channel? It couldn't be. No wonder those 125w/ch Levinsons sounded weak.

There goes that LED again. I'll be damned.

It can be true! I did a quick calculation (wasn't checking on TLSG). If you sit 4 meters away, assuming you have a room gain of 3 dB, listening at 86.4dB average you would need approximately 13.3W, double that for every 3dB so for peaks of 101.4 dB you need 424W, the next 3 dB increase will draw 850W and lit up the LED for sure. So it all depends on how far you sit and how loud you listen to. For me I listened to average SPL of less than 80, sitting only about 10 ft away in a smaller room. I have never seen my LED's lit up on any of my Adcom, Anthem and Bryston (only 300/500W). I have two 20A circuits run just for the power amps. Time to get something like a Bryston 28B SST2 at the minimum just so you can hit >100 dB without the LED brinking at you. Of course if you are happy to listen at 80dB average then you will rarely see those lights on.

 

[PENG]

The issue is that that sensitivity is low, but when you have a speaker that has a low impedance you have to be aware how much the the speaker draws from the amp.

So if you quote sensitivity at 2.83 volts 1 meter, then a four ohm speaker draws twice as much power from the amp as the 8 ohm one.

Now if you quote sensitivity at 1 watt 1 meter, then the four ohm speaker will be 3 db less sensitive than the 8 ohm speaker.

So his speakers are 83.4 db 1 watt 1 meter is sensitivity.

Years ago all speakers were quoted at 1 watt, 1 meter, which is the honest watt to do it, The 2.83 volts 1 meter was introduced to pull the wool over the eyes of the consumer.

The 1 watt, 1 meter and 2.83 volts 1 meter give the same sensitivity reading at 8 ohms, as 2.83 volts produces 1 watt with an 8 ohm load.

The other issue is that power lost in the passive crossover is power not available to the drivers, and I was using that to illustrate the waste of amp power and pointing out that if it were an active design, then the sensitivity and spl would go up five to six db.

I agree with everything you are saying except the insertion loss part. I am saying that when Revel says 2.88V/m to give you 86.4dB, they must have meant 2.88V signal applied to the speaker terminal so any insertion loss would have all been factored in that 2.88V requirement. It is important that this part is clear if you are talking about an impact of 5 dB due to the crossover. Again, my point is, Revel sell the whole speaker with the crossover built-in, how can they tell people they are getting 86.4dB without telling them they are actually getting 81.4dB because of the crossover? May be that's not really what you are saying..

 

[Alex2507]

Instead of more amplification you could always buy some Infinity Primus 360's like I have.

 

[TLS Guy]

I agree with everything you are saying except the insertion loss part. I am saying that when Revel says 2.88V/m to give you 86.4dB, they must have meant 2.88V signal applied to the speaker terminal so any insertion loss would have all been factored in that 2.88V requirement. It is important that this part is clear if you are talking about an impact of 5 dB due to the crossover. Again, my point is, Revel sell the whole speaker with the crossover built-in, how can they tell people they are getting 86.4dB without telling them they are actually getting 81.4dB because of the crossover? May be that's not really what you are saying..

You are absolutely correct on the last point. I'm just trying to point out where the power goes, and that since there is no huge magnet + voice coil movement, that crossover burn out is now a problem we did not encounter often, until high powered amp became more readily available.

Revel won't say how much power those Salon 2 take, but probably not much more than the OP is driving them with. I would doubt he can give them enough power to solve his problem.

To make a high powered speaker really does require an active solution, at least in the lower crossover regions.

I would really like Irvrobinson to measure the DC resistance of those salon 2s at the speaker terminals, and see how close the DC resistance is to 3.7 ohms. If the DC resistance should be above 3.7 ohms, then that fourth order crossover is ringing badly.

I'm highly suspicious of passive crossovers below 350 Hz and especially fourth order ones. Fourth order crossovers are very prone to ringing as it is and at 150 Hz you set your self a huge climb to make a satisfactory crossover. This has never worked for me, and wonder how well it has really worked for the HK team.

I leaned from experience 30 years ago to make any crossover below the 350 400 Hz range active, especially those over first order.

 

[Irvrobinson]

What volume are you listening at? The Revel Ultima Salon2 speakers have a sensitivity of 86dB/2.83V/m. But the impedance drops as low as 3 ohms and the phase deviates at some of those dips as well (Revel Ultima Salon2 loudspeaker Measurements) which may present a more difficult load, regardless of actual SPL.

I didn't have the OmniMic set up at the time. I've previously recorded peaks of over 106db at my listening position, and I think was playing this album a bit louder. The album is especially well recorded, and I shamelessly admit to being a Beach Boys fan.

 

[PENG]

I'm just trying to point out where the power goes,

Got it, iow, one of the reasons why the sensitivity could be that low! I used to have speakers with that kind of low sensitivity but most of them were acoustic suspension type.

 

[Irvrobinson]

Why are you surprised?

Intolerable ignorance, obviously.

At 90 Hz the impedance of your speaker is 3.7 ohms and it does not go significantly above four ohms until 600 Hz. So it is a low impedance speaker where the majority of the power is.

I know, and in this frequency range each is channel is rated for 450W.

So your speakers will draw 2.16 watts at 90 Hz from the amp to produce 86.4 db at I meter, versus 1 watt if it were 8 ohms. So the low impedance in the range where the power is has doubled the power requirement for your speakers versus 8 ohm ones.

Now it gets worse, your speakers have fourth order crossovers, and at a low frequency. Your insertion loss is1 db per order under the best of circumstances, assuming the use of the highest quality inductors. Since the crossover is far lower than can be recommended in a passive design I suspect you power loss is around 5 db and may well be as high as 6 db.

So, slightly over half your amp power is heating up the crossover.

You've lost me. Which crossover is far lower than can be recommended in a passive design? Revel's 150Hz 4th order crossover between the woofers and the lower midrange driver?

Now lets take a look at an active situation. Lets take my speakers.

The bass units are each 8 ohm, since there are two drivers of 8 ohms, each directly coupled to an amp sensitivity is 93 db 2.83 volts 1 meter and that will take one watt of power. You will have to provide at least 8.4 watts of power to achieve the same spl, and more likely 10 watts.

Now even with that situation I still need to provide 750 watts per speaker, to keep things relaxed at concert levels.

I'm a big fan of active crossovers too. One reason I was originally drawn to listen to the Linkwitz Orions. It looked like SL employed some inherently better thinking their design, which always fascinated me.

That is why good monitor companies like ATC and PMC provide about 3 to 4 KW per speaker and that is with active crossovers.

Which models? I spent quite a bit of time listening to the ATC SM150SL, and it was nowhere near that powerful. The website says 350w divided between the drivers. The imaging of the SM150 wasn't very impressive, incidentally. The speakers never disappeared in a pretty good set-up.

The time is now to ditch passive crossovers.

Maybe, but for those of us that aren't into DIY speakers the choices are very limited.

 

[AcuDefTechGuy]

I've never had this issue.

Are you using Audyssey or other EQ?

Are your speaker channel levels set to below the 00dB point in your processor?

DenPureSound had this issue with his Klipsch. Audyssey set his speaker channel levels to like -9dB.

 

[Irvrobinson]

Turn the volume down and save your money.

What fun is there in that? One could make that argument about my entire system.

 

[Irvrobinson]

It can be true! I did a quick calculation (wasn't checking on TLSG). If you sit 4 meters away, assuming you have a room gain of 3 dB, listening at 86.4dB average you would need approximately 13.3W, double that for every 3dB so for peaks of 101.4 dB you need 424W, the next 3 dB increase will draw 850W and lit up the LED for sure. So it all depends on how far you sit and how loud you listen to. For me I listened to average SPL of less than 80, sitting only about 10 ft away in a smaller room. I have never seen my LED's lit up on any of my Adcom, Anthem and Bryston (only 300/500W). I have two 20A circuits run just for the power amps. Time to get something like a Bryston 28B SST2 at the minimum just so you can hit >100 dB without the LED brinking at you. Of course if you are happy to listen at 80dB average then you will rarely see those lights on.

That's the challenge. Since the ATI is 450w/ch to get a meaningful bump in headroom I'll need at least 1000w/ch. A Crown Macro-Tech or I-Tech seems more like the ticket, but they have cooling fans which are audible, and the specs (other than very high power) aren't encouraging compared to the ATI.

 

[Irvrobinson]

I've never had this issue.

Are you using Audyssey or other EQ?

Are your speaker channel levels set to below the 00dB point in your processor?

DenPureSound had this issue with his Klipsch. Audyssey set his speaker channel levels to like -9dB.

No EQ beyond the sub. No processor. The Salon2 woofers are driven by a the digital 80Hz / 6db/oct high-pass feed from the DD18+. The mid-high portion of the Salon2 crossover is driven by a direct feed from the preamp. Like you I probably run the DD18+ a little hot, but that's not a factor for the ATI amp, obviously. According to OmniMic my in-room response at my listening position is about +/- 5db at my listening seat.

 

[AcuDefTechGuy]

This probably won't help. But at this point, might as well try everything.

Instead of bi-amping, go back to single amp per speaker.

You recently bought the ATI and went to bi-amping. This was not an issue before?

 

[jjackkrash]

"This light will come on if the amplifier exceeds
its maximum output capability on any channel. While this
will not harm the amplifier, it does warn of potential harm to
your loudspeakers if the light stays illuminated continuously
for more than ten seconds."

I wonder if blink or a flash with the read light is really a big deal. Its certainly a far cry from "illuminated continuously for more than ten seconds." I'd call ATI and ask them about what a blink or flash means and if you are really out of headroom at that point. I'd be curious what they say.

 

[Irvrobinson]

This probably won't help. But at this point, might as well try everything.

Instead of bi-amping, go back to single amp per speaker.

You recently bought the ATI and went to bi-amping. This was not an issue before?

No, I was using a similar set-up before, but with twin Mark Levinson 334 stereo amps. The only difference with the ATI is 3-4db of additional headroom before clipping. That extra headroom is apparently important, because I find myself listening louder with the ATI amp than with the Levinsons.

If I went back to two-channel operation I'd have to run the Revel woofers full-range, and that sounds muddy in my room. As I posted previously, that alternative also measures poorly. The positions of the Revels are not very good for bass smoothness.