Home » Sponsored » Pi Speakers » 4 pi speakers + 3 pi subwoofers, pics & measurements (How to interpret the measurements?)
4 pi speakers + 3 pi subwoofers, pics & measurements [message #81352] Tue, 26 May 2015 13:29 Go to next message
vandevoordekoen is currently offline  vandevoordekoen
Messages: 18
Registered: January 2013
Chancellor
Hello,

Two years ago I started with the 4pi project thanks to Wayne who helped me a lot with his plans and answering my questions. Smile

Pictures!

1. Building the 4pi speakers


index.php?t=getfile&id=1700&private=0

index.php?t=getfile&id=1701&private=0

index.php?t=getfile&id=1702&private=0

index.php?t=getfile&id=1703&private=0

index.php?t=getfile&id=1704&private=0

index.php?t=getfile&id=1705&private=0

index.php?t=getfile&id=1707&private=0

index.php?t=getfile&id=1708&private=0

index.php?t=getfile&id=1706&private=0


2. Finishing


index.php?t=getfile&id=1709&private=0

index.php?t=getfile&id=1710&private=0

index.php?t=getfile&id=1711&private=0


3. New house, made new tv furniture --> new finishing (birch)


index.php?t=getfile&id=1720&private=0

index.php?t=getfile&id=1719&private=0


Measurements with ecm8000 mic (listening position): very poor low-end due to room modes and resonances:


index.php?t=getfile&id=1699&private=0


I've bought two lab12 subwoofers and made 2 testing enclosures. crossover with mini-DSP (30hz;48db/oct, 120hz;12db/oct):


index.php?t=getfile&id=1713&private=0

index.php?t=getfile&id=1714&private=0


A very geeky testing setup. My wife loves me when I do this kind of setups. Rolling Eyes


index.php?t=getfile&id=1715&private=0


Much better low-end frequency response at listening position!


index.php?t=getfile&id=1718&private=0

index.php?t=getfile&id=1717&private=0

index.php?t=getfile&id=1716&private=0


When I measure full range from the listening position, there is an obvious roloff at the high frequencies. Is this something to worry about? I used the behringer ecm8000 mic, used CAL-file (which makes the roloff even steeper) and the m-audio fast track which is also calibrated by REW.

Thank you in advance!



Re: 4 pi speakers + 3 pi subwoofers, pics & measurements [message #81354 is a reply to message #81352] Tue, 26 May 2015 13:39 Go to previous messageGo to next message
Wayne Parham is currently offline  Wayne Parham
Messages: 18670
Registered: January 2001
Illuminati (33rd Degree)

Those measurements look good for the most part. But I agree with you that something is wrong above 8kHz.

To isolate the cause, try putting the microphone right in the tweeter throat and see what the response looks like. If the response above 8kHz doesn't look like what's in our published spec sheets, then something is wrong with the compression driver, crossover or measurement gear.

Normally I'd say look at the crossover first. And I suppose it wouldn't hurt to isolate it by removing the crossover. Measure the compression driver, all by itself, with just a series capacitor for protection. But honestly, if we had a crossover problem, I'd expect the response to be bad starting a couple octaves below that. This kind of rolloff looks more like a problem with the driver or the microphone.

In any case, you'll have to do a process of elimination to find the culprit.

Re: 4 pi speakers + 3 pi subwoofers, pics & measurements [message #81355 is a reply to message #81352] Tue, 26 May 2015 13:47 Go to previous messageGo to next message
vandevoordekoen is currently offline  vandevoordekoen
Messages: 18
Registered: January 2013
Chancellor
Ok, thanks. I'll give it a try and will post the results here.
Re: 4 pi speakers + 3 pi subwoofers, pics & measurements [message #81357 is a reply to message #81354] Wed, 27 May 2015 08:05 Go to previous messageGo to next message
vandevoordekoen is currently offline  vandevoordekoen
Messages: 18
Registered: January 2013
Chancellor

To measure the tweeter right, I've put the microphone in the tweeter throat

index.php?t=getfile&id=1721&private=0


These are the measurements with the ecm8000 calibration file uploaded

1. Right tweeter

index.php?t=getfile&id=1725&private=0



2. Left tweeter

index.php?t=getfile&id=1723&private=0



Same measurements without ecm8000 calibration file


1. Right tweeter

index.php?t=getfile&id=1724&private=0


2. Left tweeter

index.php?t=getfile&id=1722&private=0


The measurements without calibration file look approximately like the published sheets (little more rolloff).

But what with the steeper roloff when measured at the listening position?

And what about the measurements with calibration file (which are more accurate, no?)

Could it be possible that I made a mistake when making the crossover so the CD compensation isn't working?

Thanks in advance.
Re: 4 pi speakers + 3 pi subwoofers, pics & measurements [message #81359 is a reply to message #81357] Wed, 27 May 2015 11:04 Go to previous messageGo to next message
Wayne Parham is currently offline  Wayne Parham
Messages: 18670
Registered: January 2001
Illuminati (33rd Degree)

I cannot help but wonder if the calibration file is wrong. One must make the inference that the microphone used to create the calibration file has rising response, so the conjugate was to reduce HF by the same amount to compensate. But from what I'm seeing in the response curves, that doesn't appear to be the case. The raw measurements look closer to what I would expect, just like you said. So I lack confidence in the measurements.

I suppose you could have done something in the crossover that would cause this, because the response droops about 10dB through the decade of operation, e.g. 1kHz to 10kHz. That's what mass-rolloff looks like generally, although it is usually flat to about 4kHz and we're seeing falloff around 2.5kHz. So that makes me wonder about the compression drivers.

If you want to isolate crossover from driver, just disconnect the crossover and measure without it, using a direct connection to the compression driver.

In any case, I think we should be suspicious of the measurements and refrain from making too many judgement calls based on them. We're just not sure if the measurements are valid.

Re: 4 pi speakers + 3 pi subwoofers, pics & measurements [message #81360 is a reply to message #81359] Wed, 27 May 2015 16:23 Go to previous messageGo to next message
vandevoordekoen is currently offline  vandevoordekoen
Messages: 18
Registered: January 2013
Chancellor
Disconnecting the tweeter is something for tomorrow or the weekend. I measured the right tweeter (still connected to the crossover) with another mic.

This is the result:

index.php?t=getfile&id=1726&private=0


The mic that was used for this measurement isn't perfect, but compared to the measurement with the ecm8000 mic + CAL-file, the same roloff appears...

earlier measurement with ecm8000 and CAL-file

index.php?t=getfile&id=1725&private=0


I guess measuring the tweeter disconnected will give a better image of the sound. To be continued...

Re: 4 pi speakers + 3 pi subwoofers, pics & measurements [message #81361 is a reply to message #81360] Thu, 28 May 2015 11:19 Go to previous messageGo to next message
Wayne Parham is currently offline  Wayne Parham
Messages: 18670
Registered: January 2001
Illuminati (33rd Degree)

I agree with you that using the second microphone tends to confirm the original measurements. The trend shows in both measurements. The second microphone isn't as flat, but the trend is still evident. I would consider this to be corroboration and it gives me more confidence in the tests, or at least in the microphone. I'm still not 100% confident the rolloff isn't from something else besides the compression driver and/or crossover, but I'm becoming more convinced that it is.

It will be intersting to see the measurements sans crossover. That will narrow this down further.

Re: 4 pi speakers + 3 pi subwoofers, pics & measurements [message #81362 is a reply to message #81361] Thu, 28 May 2015 11:32 Go to previous messageGo to next message
chrisR is currently offline  chrisR
Messages: 36
Registered: May 2009
Baron
Wayne, I bought a calibrated PE mic from these guys:
http://www.cross-spectrum.com/measurement/calibrated_dayton.html
Assuming we trust their calibration, I also get similar measurements in the upper octave (10K-20K), where there's a roll-off at ~17KHz. I'll try to figure out how to post the plots.
Chris
Re: 4 pi speakers + 3 pi subwoofers, pics & measurements [message #81363 is a reply to message #81362] Thu, 28 May 2015 12:59 Go to previous messageGo to next message
vandevoordekoen is currently offline  vandevoordekoen
Messages: 18
Registered: January 2013
Chancellor
Here are the measurements of the tweeter directly connected to the amp. ECM8000 mic with calibration file was used.

Frequency response with horn

index.php?t=getfile&id=1728&private=0


Frequency response without horn

index.php?t=getfile&id=1727&private=0


Same rolloff...

I guess the mic is the problem and the similar roloff with the other mic is coincidence?
  • Attachment: driver.jpg
    (Size: 64.19KB, Downloaded 6623 times)
  • Attachment: tweeter.jpg
    (Size: 66.80KB, Downloaded 6688 times)
Re: 4 pi speakers + 3 pi subwoofers, pics & measurements [message #81364 is a reply to message #81363] Thu, 28 May 2015 14:36 Go to previous messageGo to previous message
Wayne Parham is currently offline  Wayne Parham
Messages: 18670
Registered: January 2001
Illuminati (33rd Degree)

I think we can rule out the microphones, because two different ones show the same trend. And I think we can rule out the waveguide, both because its properties are known and also because measurements show it hasn't modified response. We would expect no response change from a waveguide or CD horn. But that still leaves the compression drivers and the measurement system.

The DE250/H290C driver/waveguide combination has less rolloff up high than what I would expect, which is why the crossover has no capacitor in position C1. Traditionally, one would expect to have compensation for mass-rolloff in the form of a first-order high-pass filter in the crossover's tweeter circuit. But not all drivers are like that.

Mass-rolloff is expected, and it's really more unusual to not have it on a CD horn or waveguide. So the DE250/H290C combination is unusual.

Which leads me to this question: Is the driver you're testing a DE250? If not, you might consider installing a capacitor in position C1 to provide mass-rolloff compensation. The response curve shown looks like it could use it. So if the driver isn't a DE250, this may be all that needs to be done.

Previous Topic: 4Pi Crossover Update
Next Topic: 2pi plans
Goto Forum:
  


Current Time: Fri Mar 29 01:51:40 CDT 2024

Sponsoring Organizations

DIY Audio Projects
DIY Audio Projects
OddWatt Audio
OddWatt Audio
Pi Speakers
Pi Speakers
Prosound Shootout
Prosound Shootout
Smith & Larson Audio
Smith & Larson Audio
Tubes For Amps
TubesForAmps.com

Lone Star Audiofest