Making Measurements

[greywarden]

Ok, so I'm using my laptop -> REW -> Blue Icicle -> Cross Spectrum Dayton EMM-6 and I'll be most likely be using one of my AMP-100s to power them. I've built a TALL (I think it's about 7 or 8ft) speaker stand out of 2x4s (it's uber sturdy) and when I did the initial Deadhorse measurements they were really clean to about 200hz or so.

Now, what I'd like to do is get clean measurements from the bottom (F3 of the new design is about 95Hz, so an octave or 2 lower than that) all the way to the top, 20kHz and I would also like the spl measurements to be absolute.

So I'll need to first adjust the amplifier to output 2.83v, correct? So how do I go about doing that?

Next, I'll need to take nearfield, midfield, and farfield measurements. Do I do the farfield first, then correct the volume levels or adjust the FRD later? Oh yeah, and some horizontal-plane off-axis measurements, too - advise on that as well.

And finally I'll need to take a measurement of both drivers together to ascertain my "z" measurement.

I'm thinking I'll build an attachment that can hold the mic at the 3 intervals, and at a predetermined angle, at the same time. Wheels are done spinning already - probably make that this weekend.

Basically I'd like to make some measurements that are "as professional as possible" with my setup.

Thanks in Advance!

[johnnyrichards]

Re: Making Measurements

When I used to do measurements like that, I would match everything to my 1 meter measurements at a frequency unaffected by the crossover and baffle step loss. PITA.

Nowadays, I tune BSC by ear (and correlate with the transfer function graphs in PCD) and only do 1M measurements. I used to do near field measurements and ground plane and all those, but after I got an *ear* for things, I found it less and less necessary to so. I also do not worry about sensitivity measurements. My rig *used* to be SPL accurate, but I fiddlef*****d with some settings and hosed everything lol. I need to spend some time and get it all back in tune.

As far as setting it to 2.83V, I have no idea of the proper technique to do so unless you hook up a driver and attach a DMM to it.

[[email protected]]

Re: Making Measurements

i think you need the dmm at the speaker connection on the amp and must use a test tone. not sure what the standard is, maybe 1k. if you put the speaker on a turntable, you can turn it to 15,30,45 ect.

[greywarden]

Re: Making Measurements

The idea in my head was to move the pic, instead of the speaker. Put a hinge of some sort on 2x4 or something that will hold a "mic stand" in 1 of three places along the 2x4 and attach a plate with holes marked 15*, 30*, 45*, etc. so I can put a bolt in there to hold it.

As far as geting the amplifier to 2.83v I assumed I would play pink noise through the speaker and use a DMM to check the voltage.

[cjd]

Re: Making Measurements

The idea in my head was to move the pic, instead of the speaker. Put a hinge of some sort on 2x4 or something that will hold a "mic stand" in 1 of three places along the 2x4 and attach a plate with holes marked 15*, 30*, 45*, etc. so I can put a bolt in there to hold it.

As far as geting the amplifier to 2.83v I assumed I would play pink noise through the speaker and use a DMM to check the voltage.

8ohm resistor, not a speaker, since the spec is [email protected]

[carlspeak]

Re: Making Measurements

The 1Watt spec. has been largely disregarded as of late. No speaker is exactly 8 ohms and thus setting the volume level so a DMM reads 2.83V will yield whatever SPL results. Some more favorable than others.
I do exactly what greywarden wrote; pink noise at 2.83V and take an SPL reading a 1M. Fini......

[billfitzmaurice]

Re: Making Measurements

So I'll need to first adjust the amplifier to output 2.83v, correct? So how do I go about doing that?

Don't bother. Measure at a level similar to normal listening. After you're done with the FR measurements run a 1kHz test tone, adjust that to 2.83v, measure it with a meter, use that to calibrate the chart.

[cjd]

Re: Making Measurements

The 1Watt spec. has been largely disregarded as of late. No speaker is exactly 8 ohms and thus setting the volume level so a DMM reads 2.83V will yield whatever SPL results. Some more favorable than others.
I do exactly what greywarden wrote; pink noise at 2.83V and take an SPL reading a 1M. Fini......

Yes, relatively speaking this is true.

However, if I want a consistent baseline you can compare against data acquired elsewhere, this is still a reasonable approach - dial for [email protected]

It really depends what you're after.

C

[charlielaub]

Re: Making Measurements

Ok, so I'm using my laptop -> REW -> Blue Icicle -> Cross Spectrum Dayton EMM-6 and I'll be most likely be using one of my AMP-100s to power them. I've built a TALL (I think it's about 7 or 8ft) speaker stand out of 2x4s (it's uber sturdy) and when I did the initial Deadhorse measurements they were really clean to about 200hz or so.

Now, what I'd like to do is get clean measurements from the bottom (F3 of the new design is about 95Hz, so an octave or 2 lower than that) all the way to the top, 20kHz and I would also like the spl measurements to be absolute.

So I'll need to first adjust the amplifier to output 2.83v, correct? So how do I go about doing that?

Next, I'll need to take nearfield, midfield, and farfield measurements. Do I do the farfield first, then correct the volume levels or adjust the FRD later? Oh yeah, and some horizontal-plane off-axis measurements, too - advise on that as well.

And finally I'll need to take a measurement of both drivers together to ascertain my "z" measurement.

I'm thinking I'll build an attachment that can hold the mic at the 3 intervals, and at a predetermined angle, at the same time. Wheels are done spinning already - probably make that this weekend.

Basically I'd like to make some measurements that are "as professional as possible" with my setup.

Thanks in Advance!

Your tall "stand" is similar to putting a speaker on top of a 6' step ladder. I have done this, albeit indoors, and found I could obtain gated impulse frequency response measurements that were valid almost down to 100 Hz. This would give you a FR measurement 100Hz-20kHz. This will be good for all your "farfield" measurements and off-axis measurements. I use ARTA and once you get the setup working you can pretty quickly go back and forth between adjusting mic-to-speaker angle and recording a new measurement.

Even outdoors on your stand you will get reflections off of the ground, so practically speaking below that use a nearfield or ground plane measurement technique and then splice the two together using one of the tools available for that purpose (Jeff Bagby's Frequency Response Modeler or my FRD Response Blender). This will give you your full "wide band" frequency response. Since you know the level of your "farfield" measurement will be correct, you only need to adjust the level of the nearfield during the splicing procedure.

I typically work with minimum phase measurements, and then use the three-measurement technique to get the driver offsets. This has worked very well for me.

Whether you are driving a resistor or speaker, 2.83V is 2.83V from a constant voltage source like an audio amplifier. You will need to measure the voltage at the speaker terminals with a good voltmeter and then make sure not to re-adjust the gain anywhere along the chain.

-Charlie

[greywarden]

Re: Making Measurements

Your tall "stand" is similar to putting a speaker on top of a 6' step ladder. I have done this, albeit indoors, and found I could obtain gated impulse frequency response measurements that were valid almost down to 100 Hz. This would give you a FR measurement 100Hz-20kHz. This will be good for all your "farfield" measurements and off-axis measurements. I use ARTA and once you get the setup working you can pretty quickly go back and forth between adjusting mic-to-speaker angle and recording a new measurement.

Even outdoors on your stand you will get reflections off of the ground, so practically speaking below that use a nearfield or ground plane measurement technique and then splice the two together using one of the tools available for that purpose (Jeff Bagby's Frequency Response Modeler or my FRD Response Blender). This will give you your full "wide band" frequency response. Since you know the level of your "farfield" measurement will be correct, you only need to adjust the level of the nearfield during the splicing procedure.

I typically work with minimum phase measurements, and then use the three-measurement technique to get the driver offsets. This has worked very well for me.

Whether you are driving a resistor or speaker, 2.83V is 2.83V from a constant voltage source like an audio amplifier. You will need to measure the voltage at the speaker terminals with a good voltmeter and then make sure not to re-adjust the gain anywhere along the chain.

-Charlie

Awesome thank you Charlie, that seems like the easiest way to do it. So I'll take On-axis - Near/farfield; Off-axis - Near/farfield; and then the "3" for finding out the AC measurement.

Is there a concrete prescribed method for nearfield measurements? Or is it just "really damn close"?

[ErinH]

Re: Making Measurements

you don't need to do individual nearfield for off-axis. you'll be stitching below where the driver begins to beam (at least to any notional degree). Just do nearfield at a close enough point and your farfield results can be merged with the farfield.

This is a great link on NF/FF merging:

http://www.klippel.de/fileadmin/klippel/Files/Know_How/Application_Notes/AN_39_Merging_Near_and_Farfield_Measurements.pdf

There are two aspects when taking measurements covered in the above link.

1. One is the window time you have before reflections, which is typically limited to floorbounce and around 3ms or 333hz. It'll vary a bit but it's in this ballpark unless you're able to set the speaker/driver up higher off the floor (and then you're probably running in to ceiling bounce unless you're outside). This is your farfield minimum frequency limit.
2. The other aspect is the critical point where the driver is out of it's pistonic range. This is where your nearfield maximum frequency limit is. This is all dependent upon the driver size. If you have a 7" driver, let's say with 18cm effective diameter you are limited to 5475/(18/2) = 608hz.

So, the above two criterion mean you have a maximum nearfield of 608hz and a minimum farfield of 300hz. There's about 300hz of wiggle room there. If you have a smaller driver, such as a 3" full range with a diameter of closer to 10cm your nearfield max increases to 5475/(10/2) = 1095hz. So, you have about 800hz range from farfield to nearfield bounds (1095-300hz).


All that said, this is what I would do:

• Set the amplifier output to 2.83v with a 1khz, 0dB tone.
• Connect the speaker
• Put the mic in the farfield, measure and determine via impulse the first reflection. You'll see the spike and it will probably be in the 3ms range, unless as noted above you are able to place the speaker further away from a reflective boundary
• Save this FR result
• Knock down the voltage to something a bit more minimal. this is done mainly for mic protection in case it's not able to withstand higher spl like you would get in the nearfield (every doubling of distance is a 6dB increase/decrease. If you're at 1m and measure within a couple inches of the driver, you've increased SPL measured by the mic by about 18dB. This can damage a mic if the spl is already loud in the farfield.
• Move the mic to a couple cm within the speaker or between speakers.
• Measure and set the window wide... something like 100-200ms is sufficient.
• Compare the NF and FF results and look for where the trends are close to the same in the min/max passband you get from reflection/NF max.
• Merge here by bringing the NF SPL down to match the FF result. This will give you a anehoic-ish response. I typically find I merge around the 500-800hz range, depending on speaker size.

The one thing that hasn't been discussed is baffle but you're measuring a speaker rather than a raw driver so it's not as big a deal to note this.

The other thing that I really encourage you to consider is that anything you set the speaker on - any boundary such as a ladder, stand, etc - can affect your measurements. Troels has shown this. I found the same to be true. Make sure your mic doesn't have the holder right behind it. I built out my mic on a piece of PVC and smoothed the transition from PVC to mic so there would be no reflection back on the mic.
Troels' site: http://www.troelsgravesen.dk/measurements.htm


HTH



I should note that the above is how I measure drivers and speakers. Truthfully, for measuring speakers there may be a better method but it's worked for me and is extremely straightforward. I'll post up my speaker test baffle and show you how I do my off-axis measurements... just promise not to laugh. It involves a hook and a dog leash. :D

[greywarden]

Re: Making Measurements

Wow that's some awesome information, Eric, thanks alot! I'm curious to see this baffle, haha

[tyger23]

Re: Making Measurements

I guess I'm not sure what's so difficult about 1W @ 1M. Take the square root of the impedance of the speaker in question at the test frequency being used, and set the output of the amplifier (in volts) to that number. Voila.

I.e. - impedance of the test speaker is 7.1 ohms @ 1KHz. Set the AC output volts of the amplifier (WHILE SUPPLYING SIGNAL ACROSS THE ACTUAL SPEAKER IN QUESTION) equal to SQRT(7.1) = 2.66V.

Use a TrueRMS multimeter set to measure AC volts to make this measurement.

It does help to make sure the frequency output is set to the low point, or "trough" of the impedance curve. This will vary from speaker to speaker, but 1KHz is pretty close for *most* woofers/mids. Tweets should mostly be a bit higher, like 3KHz or so.

[carlspeak]

Re: Making Measurements

I've read 1 kHz test tone in a number of posts now. That's all well and good for a 2-way speaker. What about a 3-way? At 1 kHz, the woofer is out of the reading completely. IMO, not good. That's why I use pink noise for 3-ways.

I also liked erinh's succinct explanation on testing. IMO, it sums up much of what it took J. D'Appolito's rambling, equation riddled tome (Testing Loudspeakers) to do and even then he followed up with two papers explaining how to interpret the test results. :-)

[ryanbouma]

Re: Making Measurements

When I measure outdoors, I usually do my nearfield measurements about 6 to 12" from the dustcap. Then smooth the measurements 1/12 or use a very long gate time. Indoors I measure nearfield around 1" away.

I agree with others, that if you're 7' up, then just gate down to around 150hz or so. And splice the nearfield. You'll have pretty much anechoic results.

When I set up REW to do sensitivity, I adjusted my amplifier to get 2.83V with a DMM at the speaker terminals with a 120hz sine tone. That happened to be -21.5db on my receiver. I now know that's what it is and always go back to that setting. I double check sometimes, but it's always right. Then I used the steps in REW with an SPL meter to set the mic level and left the gain on the blue icicle at FULL. Matches manufacturer SPL everytime and is easily repeated.