Tri-amping Qs:


I'm anticipating buying a pair of Emerald Physics EP-2.3s. I would ultimately like to tri-amp them. I'm thinking of starting with a single 100wpc amp, then adding a second amp for bi-amping, and finally a third for tri-amping. The speakers are 97dB efficiency. My average listening levels are 75-80dB and peaks are rarely over 105dB. A 100wpc amp will provide 20dBw of power. With the 97dB efficiency, max volume would be 117dB. My questions:
Is my math correct?
Will this give me enough headroom?
Will I have to use 3 identical amplifiers?
Or can I reduce the wattage of the mids and highs, say 60wpc for the mids and 30wpc for the highs?

My head says no.
Thanks GP
gary7m
First, I see that the 97 db spec is for an input of 2.83 volts, at a distance of 1 meter, and that the speaker’s impedance is 4 ohms. If the speaker’s impedance was 8 ohms, that would correspond to 97 db for 1 watt at 1 meter. However, since 2.83 volts into 4 ohms corresponds to 2 watts, the speaker’s efficiency is 94 db/1 watt/1 meter. Although if the 100 wpc amp you mentioned starting with is solid state, and that rating is for an 8 ohm load, chances are it would be able to provide somewhere between 150 and 200 watts into 4 ohms.

Second, neglecting room reflections the SPL produced by a box-type dynamic speaker such as those will fall off at 6 db per doubling of distance, which equates to 20 times the logarithm (base 10) of the ratio of two distances. So at a listening distance of say 10 feet the SPL produced will be approximately 10 db (actually 9.68 db) less than at 1 meter, neglecting the effects of room reflections.

Third, two speakers will be in use, which will add 3 to 6 db to the maximum SPL produced by one of them. The 6 db figure would apply if the listener is equidistant from the two speakers, and they are both reproducing the same signal.

So for an amp capable of providing 200 watts (23 dbW) into 4 ohms, at a 10 foot distance the maximum SPL (neglecting room effects) would be:

94 + 6 + 23 - 10 = 113 db.

This assumes, btw, that the speakers are not being driven hard enough for a long enough amount of time for "thermal compression" to set in and reduce their efficiency to a significant degree.

If the amp is rated at say 150 watts into 4 ohms, the result would be about 1.25 db less than that, or 111.75 db.

As far as biamping and triamping are concerned, while it is possible to obtain good results using different amps, using identical amps will greatly reduce the chances of poor results, that may result for example from loss of coherence due to differences in their sonic characters. Also, with different amps a means of matching their gains would most likely be necessary.

Finally, regarding the possibility of using amps having significantly different power capabilities, if an electronic crossover is not used "ahead" of the amps chances are you would be limiting the maximum power capability that could be utilized from the highest powered amp to not a great deal more than the power capability of the lowest powered amp. That is because while passive biamping (i.e., biamping without an electronic crossover ahead of the amps) reduces the amount of current and power each amp must supply, all of the amps still have to output voltages corresponding to the full frequency range of the signal, and at equal levels if gains are properly matched. So turning the volume control up high enough to utilize most or all of the power capability of the highest powered amp would most likely cause one or both of the lower powered amps to exceed its maximum output voltage capability, and therefore to be driven into clipping.

Regards,
-- Al
@almarg

Doubling of distance does not reduce output by 6dB, that’s if you were in free space (so more applicable for live sound setups); For residential setups, the loss is 3dB to 4dB, almost the inverse of doubling wattage.

Also, combing it both speaker outputs nets 3dB. 6dB is never gained. You may be confusing that with crossover design, where they have them intersect at -6dB to sum to zero, but that’s due to vector addition.

You shouldn’t factor that in anyway. Well, at least for movies, each speaker should be able to reproduce 0dBFS at the listening position. I assume the same applies for music. Most music is mastered with their peaks between -6dBFS to -3 dBFS, which corresponds to 99dBC and 102dB respectively.
@gary7m

I hope your peaks are never above 105dB, at that’s full scale, so anything abie that would cause clipping.

Firstly, I would demo this speakers. Crossing a 1” tweeter with a 12” midrange at 1kHz is going to lead to non-stellar soundstage and imaging in that region.

Those are 97dB and 4ohm nominal; which I believe is and a true anechoic spec and not a misleading in-room one. Now, they don’t state the minimum impedance it gets, maybe below 3ohm, but let’s just call it 4ohm for ease of calculation.

If you sit 4 meters away (~13ft), you’d lose anywhere from 6dB to 8dB, but keep in mind you get a boost of 2dB to 3dB for room boundary gain. So let’s just be overly cautious and say it’s now 90dB @ 1W 4ohm @ 4 meters.

So, in order to hit clipping (even though most songs don’t hit 105dB, they usually peak at ~100dB), you’d need 32W into 4ohm.

I usually advise having 2x as much wattage as you’d need, so you aren’t constantly running at full power, which can reduce longevity, bringing you to 64W. If you never plan to do DSP, then 64W into 4ohm is perfect.

Why would you want amps with different wattage ratings?. If you wanna fix some portions that are too soft or loud due to room modes, use DSP and/or room treatment. Bringing an entire region up/down is a very poor practice.

Now, I’m betting the tweeter is actually around 8ohm (as most compression drivers are), so you’d want an amp that still gets decently close with doubling wattage from 8ohm down to 4ohm.

The Emotiva PA-1 monoblocks are using the same ICE module as in the PS Audio M300, and it’s $600/pair, hard to beat. Hypex also have their NC400 monoblock kits which are <$2000 and it takes less than 2 hours probably to assemble and doesn’t need any soldering or anything.

But yeah, most any amp would be powerful enough for those, so just focus mainly on the quality (linearity, distortion, SNR, etc.). 
 
i wouldn’t even think about bi or tri amping, just a waste of money unless using low power SET amps, which I wouldn’t advise.
Doubling of distance does not reduce output by 6dB, that’s if you were in free space (so more applicable for live sound setups); For residential setups, the loss is 3dB to 4dB, almost the inverse of doubling wattage.
Note that I stated several times that I was ignoring the effects of room reflections.

Also, combing it both speaker outputs nets 3dB. 6dB is never gained. You may be confusing that with crossover design, where they have them intersect at -6dB to sum to zero, but that’s due to vector addition.
I was not thinking of crossover design, and I believe that my reference to a 6 db gain under the specific conditions I stated was correct. My statement was:

... two speakers will be in use, which will add 3 to 6 db to the maximum SPL produced by one of them. The 6 db figure would apply if the listener is equidistant from the two speakers, and they are both reproducing the same signal.
At various points in the room, if the two speakers are reproducing identical signals and again putting aside the effects of room reflections, the addition of a second speaker will cause SPL to increase by amounts ranging from minus infinity (i.e., no sound) to 6 db, depending on whether the phase relation of the two outputs at a particular vantage point causes them to sum constructively, destructively, or somewhere in between. Averaged throughout the entire room the increase will be 3 db, of course, corresponding to the doubled amount of power that is supplied to the speakers.

Some years ago I read a particularly excellent paper written by some folks in Australia which explained exactly this, in greater detail. I can’t find the link, though.

Regards,
-- Al
Gentlemen,
Thanks for your responses. I appreciate them very much and I appreciate you taking the time to write those long answers. I can now go looking for an amp that will have enough headroom @ approximately 75-100w.

Al: Thanks for your two responses. That was thoughtful. Your answers gave me a way to think differently about my questions and gave me a basis to compare amplifiers.

mzkmxcv: Thanks for taking the time to respond. Your answer added more to my knowledge and provided an answer to what power I would need to deal with headroom and also extended my understanding of how to determine how powerful an amp should be to drive particular speakers. Now I know, in the future, that I double the wattage as a good rule of thumb.

When I said "peak," I should have said, "On my chintzy iPhone-based dB meter, A-weighted, the needle infrequently taps105dB," usually in orchestral music. The EP 2.3s have an onboard DSP, so I can use your information about that. I also appreciate your opinion about tri-amping. That will save me a pile of cash and allow me to put more money into a higher quality single amp. I guess it's the geek in me that thought tri-amping would be better than using a single amp.

Again, thanks to you both,
GP
You’re welcome, Gary. Glad I was helpful.

Upon re-reading my previous post, to be precise I should make a slight tweak as follows (the words shown in caps have been added):

From: "At various points in the room, if the two speakers are reproducing identical signals and again putting aside the effects of room reflections, the addition of a second speaker will cause SPL to increase by amounts ranging from minus infinity (i.e., no sound) to 6 db, depending on whether the phase relation of the two outputs at a particular vantage point causes them to sum constructively, destructively, or somewhere in between.

To: "At various points in the room, if the two speakers are reproducing identical signals and again putting aside the effects of room reflections, the addition of a second speaker will cause SPL to increase by amounts ranging from minus infinity (i.e., no sound AT SOME FREQUENCIES) to 6 db AT SOME FREQUENCIES, depending on whether the phase relation of the two outputs at a particular vantage point causes them to sum constructively, destructively, or somewhere in between.

Regards,
-- Al
iPhone-based dB meter, A-weighted 

Being A-weighted would cause a discrepancy, C-weight should be used. On my iPhone I use the SLA Lite app, I have a calibrated mic but I haven’t compared the two to see how off the iPhone is.