Measuring actual sensitivity

[Blazin]

How does one go about this, for my latest project I had a projected results of.. 90dB +/- 3dB with an impedance dropping no lower than 3 Ohms nominally 4 ohms..

I've achieve +/- 3dB 'flat' and nominal 8ish ohms but I'd like to convert my SW graphs into actual rather than relative Magnitude..

RS SPL Meter?

[augerpro]

Basic idea goes like this:

1)Pick a tweeter with a flat-ish and consistent response that you know the sensitivity of, I used a Scanspeak 9800 for this
2)Run a MLS signal and adjust SW until teh response plots are the same level as the manufacturer's plots
3)Don't change anything, than do your other measurments. The response now reflects the actual sensitivity.

or

1)Pick a tweeter with a flat-ish and consistent response that you know the sensitivity of, I used a Scanspeak 9800 for this
2)Run a MLS signal and adjust SW until teh response plots are the same level as the manufacturer's plots
3)Do a loopback measurement. Plot the response, it should be just a flat amplitude with a flat phase. Note the level of the response, we'll call this the Ref level, and the mic gain position. Now all you have to do in the future is fire up SW, run a loopback measurement and tweak SW until it's at the Ref level, ensure the mic gain is the same, and you can now do calibrated measurements.

[Blazin]

Basic idea goes like this:

1)Pick a tweeter with a flat-ish and consistent response that you know the sensitivity of, I used a Scanspeak 9800 for this
2)Run a MLS signal and adjust SW until teh response plots are the same level as the manufacturer's plots
3)Don't change anything, than do your other measurments. The response now reflects the actual sensitivity.

or

1)Pick a tweeter with a flat-ish and consistent response that you know the sensitivity of, I used a Scanspeak 9800 for this
2)Run a MLS signal and adjust SW until teh response plots are the same level as the manufacturer's plots
3)Do a loopback measurement. Plot the response, it should be just a flat amplitude with a flat phase. Note the level of the response, we'll call this the Ref level, and the mic gain position. Now all you have to do in the future is fire up SW, run a loopback measurement and tweak SW until it's at the Ref level, ensure the mic gain is the same, and you can now do calibrated measurements.

Run an MLS, loop play record, adjust Mic gain or do a Scale in SW? Adjust what exactly. Thanks auger

[Dennis H]

Or just use the RS meter, C weighting, 1 meter, sine waves at 500-1000 Hz, with a 2.83V signal. They are pretty accurate in that range.

[jkrutke]

Or just use the RS meter, C weighting, 1 meter, sine waves at 500-1000 Hz, with a 2.83V signal. They are pretty accurate in that range.

I wouldn't use the RS meter simply because it picks up all the harmonics and the dB reading is simply the level of the entire spectrum. It's not accurate, and higher distortion speakers will seem to be more sensitive than they actually are.

The Phonic PAA3 Handheld Audio Analyzer is a better tool simply because it can measure the spl of the fundamental in RTA mode without picking up the harmonics and adding them in. (not to mention it's calibration level is much better than the rat shack meter) I have one of these. It's nice, though not the cheapest.

[Dennis H]

I wouldn't use the RS meter simply because it picks up all the harmonics and the dB reading is simply the level of the entire spectrum. It's not accurate, and higher distortion speakers will seem to be more sensitive than they actually are.

Well, I'll bow to your experience John but it's unlikely Blazin's RS mids in series would have more than 1% THD in the 500 to 1K range driven by a 2.83V signal (1.4V for each driver). But, of course, what you say is true as a general principle. 1% THD would be a 0.09dB error and 10% THD would be a 0.8dB error.

[TacoD]

Calibrate your mic.

[jkrutke]

Well, I'll bow to your experience John but it's unlikely Blazin's RS mids in series would have more than 1% THD in the 500 to 1K range driven by a 2.83V signal (1.4V for each driver). But, of course, what you say is true as a general principle. 1% THD would be a 0.09dB error and 10% THD would be a 0.8dB error.

True. I think I've just tested too many garbage speakers in my day. On the other hand, I've compared the RS meter levels to the PAA3 levels with a known driver and the RS was off by 4 dB on the high side. I never really thought about it and just attributed it to HD but now that I think about it, it was likely just a calibration issue.

[Dennis H]

On the other hand, I've compared the RS meter levels to the PAA3 levels with a known driver and the RS was off by 4 dB on the high side.

Yikes! I thought they were better than that but I've never calibrated one to know for sure. If they're really 4dB off that's no good for what Blazin wants to do (adjust his curves to 2.83V for his senior project report).

Blazin, I'm pretty sure you don't have the budget to buy a $400 meter like John uses or a $200 reference tweeter like Brandon uses. Maybe you could work from the PE measurements of the RS drivers and figure out a fudge factor for your graphs. I think they are on a large baffle, 2.83V, 1m.

Another thought, did your mic come with a calibration number (x mV/Pascal)? If so you could measure the gain of the mic preamp, measure its output during a measurement and calculate the SPL from there. 94 dB is 1 Pascal of pressure so

dB = 94 + 20*log(measured pascals) [use base 10 log].

[Blazin]

Hmm i'm sure it did.. its the Behringer ECM8000

[Dennis H]

Excellent! I think they usually specify the mic sensitivity at 1kHz so that would be a good test tone to play to check your SPL with a DVM. I don't know anything about SW. Can you calibrate it so you read real SPL when you do a measurement? Like play a tone with a known SPL and change something in the setup menus until it reads right?

[Blazin]

you could work from the PE measurements of the RS drivers and figure out a fudge factor for your graphs. I think they are on a large baffle, 2.83V, 1m.

So using FRD trace with the PE supplied FR graphs.. and my original On Axis measurements.. it appears that my graphs are 100dB off +/- 3dB haha


Sooo should I scale my final measurements up 100dB? sure.. terrible noobish way to do it.. as its a guesstimate.. but if I'm consistant, it should be consistant?

[Dennis H]

So using FRD trace with the PE supplied FR graphs.. and my original On Axis measurements.. it appears that my graphs are 100dB off +/- 3dB haha

Too late now I guess but most of the drivers have .txt files you can download.

Remember the PE files don't show baffle step so pick a frequency above there to match them. Just get one of the drivers to line up well in some frequency range (say the mid around 1K) and then boost all the drivers the same amount.

[Blazin]

Looks like I hit 87 +- 3dB seems about right considering the RS150s are 88dB anyway