I am trying to determine the target sensitivity of speakers I want to build for some SE amps I am also building. So first I want to benchmark my current speakers to see how much more output I need. I have REW and UMC-1 mic. Did I do this right?
Placed mic 1m away from center of set of speaker drivers.
Played a 1khz tone with a measured amplitude at the amp output of 2.8V RMS.
Use REW SPL measurement to measure SPL, used both A and Z weighting (no difference)
Reading was 82db.
Speakers were in room.
Placed mic 1m away from center of set of speaker drivers.
Played a 1khz tone with a measured amplitude at the amp output of 2.8V RMS.
Use REW SPL measurement to measure SPL, used both A and Z weighting (no difference)
Reading was 82db.
Speakers were in room.
Hi,
I don't know if i got what you ask right but:
What is impedance of your loudspeakers?
Did you calibrate your input level and if yes what was your reference?
I don't know if i got what you ask right but:
What is impedance of your loudspeakers?
Did you calibrate your input level and if yes what was your reference?
I have not measured the impedance of my speaker at 1khz.
The input level to the speaker was measured on a scope. I trust the measurement to the rough precision I need.
I was asking if my general procedure is sound.
The input level to the speaker was measured on a scope. I trust the measurement to the rough precision I need.
I was asking if my general procedure is sound.
The way you do this depends on what you know. Do you know the sensitivity of your current speakers? Do you know how much amp output gives you the right listening levels? Do you know how much output your new amp will have? or are you building the speaker first and need an amp for it?
I've not been accurate enough: 2,83v is 1w for an 8r load, if your loudspeakers have different impedance your measurement won't reflect something 'real'.
If i ask about reference level and calibration it's because your computer can't guess what is the gain needed for the preamp to equal 94dbspl ( 1pa) which is usually the reference on which software can give you a reading corresponding to 'real' spl.
If i ask about reference level and calibration it's because your computer can't guess what is the gain needed for the preamp to equal 94dbspl ( 1pa) which is usually the reference on which software can give you a reading corresponding to 'real' spl.
That is what I am trying to estimate.
The opposite. I have built an SE amp which has about 6-8W output. I want more sound output than my current reference speakers. I'd like 10db more output with new speakers than my current speakers. I want to determine the target sensitivity for the new speakers. So first I need to estimate my current speaker sensitivity.
The method I described in my first post suggests my current speakers are ~82db/2.82v/1m. So this suggests my new speakers should be ~92db/2.82v/1m. My question is this approach of estimating a get me in the right ballpark?
Last edited:
With the same setup I measured at my speakers a mean sensibility of ~88dB.
The speakers are made by me.
Do you load the calibration file for your microphone?
The speakers are made by me.
Do you load the calibration file for your microphone?
Yes. I load the Mic calibration file. Did you measurement confirm what you think the sensitivity was, ballpark?
Edit: In my case 82db seemed low. I don't know the specs of my speakers but they are a ported 3 way tower. I thought they would be more than 82db.
Edit: In my case 82db seemed low. I don't know the specs of my speakers but they are a ported 3 way tower. I thought they would be more than 82db.
Last edited:
well about the only other thing you can do is repeat your measurement outdoors away from boundaries and see if there's a difference heck i'd do a sweep and see what Rew says for spl at 1 khz but then you'll also know if your speakers are more sensitive at other frequencies but i'd venture your original measurement sounds about right
Then measure the speaker with known T/S parameters - at exactly 1 meter, at 1 kHz and 2.83 V at the terminals.
Is there any data for this unknown loudspeaker that you can spare?
Did you measure T/S parameters of the woofer alone, free air?
What does a x/o look like?
Did you measure T/S parameters of the woofer alone, free air?
What does a x/o look like?
You might be saying that your old amp is 10dB more powerful (50W) and you need a speaker with which the new one can keep up.
Or maybe it's uncertain what you mean by output. A more sensitive speaker cannot give more output necessarily, although it can give the same levels for less power input.
My old amp is 125Wpc. I didn't need that much with my current speakers. My current speakers do ok with the 8wpc SE amp most of the time at modest listening levels. But once in a while I want that little more as Nigel Tufnel from Spinal Tap once said. I'd like "1" louder with the new speakers.
Ok.
A quick google search revealed some infos on your loudspeakers:
Question about Merlin speakers. : audiophile
So nominal impedance 6.2r, max output 113db and max amp power 200w.
With this we could estimate they should be in the 89db sensitivity (113db - 24db (gain from 200w relative to 1w approx)= 89db). ( each time you double power you gain +3db: eg. 1 to 2w: +3db, 2 to 4w: +3db, 4 to 8w:+3db,...).
So yes your measurements seems a bit on the low side.
So as you use a 125w amp max output should be around 110db spl ( with your current loudspeakers and amp at max level).
If you want same output capability you'll need something along 100/104db for your new loudspeakers.
( 110db - 9db (8w) or 110- 6db (4w) ).
From the start i think you could estimate things differently: find the most dynamic track you have. Set your playback volume to your target level at listening point.
Then you play pink noise and use a spl meter ( an appli on your phone or tablet will be enough) at listening spot ( use some protection for your ears it'll probably be loud!).
You then ad the level loss from distance ( you loose 6db each time you double distance: eg 1 to 2m : -6db, 2 to 4m: -6db,...).
Done.
Could be wise to make pink noise -20db, you'll then ad 20db to spl you'll read. Much less chances you'll develop tinitus that way. 😉
A quick google search revealed some infos on your loudspeakers:
Question about Merlin speakers. : audiophile
So nominal impedance 6.2r, max output 113db and max amp power 200w.
With this we could estimate they should be in the 89db sensitivity (113db - 24db (gain from 200w relative to 1w approx)= 89db). ( each time you double power you gain +3db: eg. 1 to 2w: +3db, 2 to 4w: +3db, 4 to 8w:+3db,...).
So yes your measurements seems a bit on the low side.
So as you use a 125w amp max output should be around 110db spl ( with your current loudspeakers and amp at max level).
If you want same output capability you'll need something along 100/104db for your new loudspeakers.
( 110db - 9db (8w) or 110- 6db (4w) ).
From the start i think you could estimate things differently: find the most dynamic track you have. Set your playback volume to your target level at listening point.
Then you play pink noise and use a spl meter ( an appli on your phone or tablet will be enough) at listening spot ( use some protection for your ears it'll probably be loud!).
You then ad the level loss from distance ( you loose 6db each time you double distance: eg 1 to 2m : -6db, 2 to 4m: -6db,...).
Done.
Could be wise to make pink noise -20db, you'll then ad 20db to spl you'll read. Much less chances you'll develop tinitus that way. 😉
Last edited:
I knew the developer of those speakers, Bobby Palkovic, he was local to me. I know most of those specs are WAG's.
But I think sensitivity is closer to 89 than 82. I need to figure out why my technique gave a low number. When playing music my system recorded 99 db peaks with the 8W amp, probably was clipping so that wasn't 8 clean watts.
I'll try more measurements. I have ear protection.
If you have ear protection then use it with the method i gave ( even with -20db pink noise!).
I repeat what i stated in my second message: if you have not calibrated your preamp level to a known (and reliable) spl your analysis software can't give you accurate absolute reading. Relative db difference will be reliable ( as long as your in a range out of noisefloor -acoustic or electronic), not the reference level. 😉
Food for thoughts ( and one of the source of the method given/ it is used in theater too with minor difference, and more widely in pro world):
Honor Roll - Digido.com
I repeat what i stated in my second message: if you have not calibrated your preamp level to a known (and reliable) spl your analysis software can't give you accurate absolute reading. Relative db difference will be reliable ( as long as your in a range out of noisefloor -acoustic or electronic), not the reference level. 😉
Food for thoughts ( and one of the source of the method given/ it is used in theater too with minor difference, and more widely in pro world):
Honor Roll - Digido.com
Last edited:
My preamp is not in the loop at all. I am using a signal generator feeding the amp directly. I know the input signal and measure the amplitude at the speaker terminals with an oscilloscope. There is no calibration (other than the scopes) needed. I know I am putting 2.82V rms on the speaker terminals.