|
|
| Re: 4Pi measurements [message #74240 is a reply to message #74239] |
Wed, 24 October 2012 00:33   |
 |
Wayne Parham
Messages: 18786
Registered: January 2001
|
Illuminati (33rd Degree) |
|
|
Your measurements look right except for two things:
1. The first speaker does have more peaking around 1.6kHz than it should. Could be the crossover, or it could be the midwoofer cone. Is it a new woofer? If not, you might measure it by itself and see if there is excessive cone breakup. Also, check your crossover and make sure you have the Zobel connected. A missing Zobel will do that. And check all the other values too, especially the capacitors in the woofer circuit and the coil in the tweeter circuit.
2. There is more rolloff up high than I would expect. They don't falloff this much in the top decade. I'm guessing it's something in the measurement system. It has what appears to be a first-order rolloff above about 2kHz. I don't really think your speakers are doing that, partly because it is consistent across multiple speakers, and partly because this is a fairly common measurement problem. I see it happen pretty often, people produce a curve that looks a lot like mass rolloff even on speakers that couldn't possibly produce that kind of response.
Other than those two things, your measurements look right. I think if you were to conjugate that abnormal first-order filter above 2kHz, your response curves would be exactly right.
You've done a very good job, and I'll bet they sound great.
Поздравляю!
|
|
|
|
|
|
| Re: 4Pi measurements [message #74244 is a reply to message #74241] |
Wed, 24 October 2012 14:17 |
|
Wayne Parham
Messages: 18786
Registered: January 2001
|
Illuminati (33rd Degree) |
|
|
Пожалуйста.
To measure the woofer, leave it in the box and just disconnect the tweeter. A sweep up to 2kHz or 2.5kHz is sufficient, but it wouldn't hurt to go to 4kHz. I would measure it two ways: (1) Measure the woofer connected directly to the amplifier (without crossover). This will tell you what the woofer is doing, all by itself. If it has excessive breakup, this will show it. Then (2) measure with the woofer crossover connected, which should show smooth rolloff above 1kHz. If it peaks at 1.6kHz, we have a problem with the Zobel or the core low-pass splitter filter.
As for the HF, you have the microphone positioned exactly right. In fact, it will be pretty much the same even if the microphone is moved towards the top or the bottom of the box, until you reach the point of the vertical nulls, which are pretty far apart. See the following thread, and click on the "Vertical Nulls" link for a video of the measurement process (just like what you're doing):
There are a lot of things that can cause the rolloff up high, but I think the most common one is just needing calibration. Seems like most uncalibrated systems are a little hot up high or a little rolled off. The microphone can be measured by a calibration house, who will return to you a disk that you install in the measurement system. The data on the disk is essentially a response curve of the microphone, which is then conjugated to make the resulting data true. Of course, this assumes the measurement system is accurate, including its internal preamp and ADC. It would be good to have this verified too.
|
|
|
|
|
|
| Re: 4Pi measurements [message #74246 is a reply to message #74245] |
Wed, 24 October 2012 15:11 |
|
Wayne Parham
Messages: 18786
Registered: January 2001
|
Illuminati (33rd Degree) |
|
|
I agree with you - I wouldn't worry about your calibration. I think that's most important when you need a solid reference, like if you were planning to publish your charts or make some sort of formal comparisons. But for what you're doing, I am already comfortable with what you have. In a way, your four π speakers can be used as a reference, because they look just like what I would expect except for an additional first-order LP filter function above 2kHz. You can mentally conjugate that curve and feel comfortable with it.
On the low-end, that's not uncommon either. At very low frequencies, a Delta-Sigma converter has a hard time, it's just the nature of the process. That's why some measurements systems take forever when measuring low frequencies. They're trying to compensate for that. An old Nyquist converter just uses successive approximation, so low frequencies are easy. But Delta-Sigma is a different process, and does not work well at very low frequencies.
|
|
|
|
|
|
|
|
|
|
|
|
Posting Rules
Members
Search
Help
Register
Login
Get a T-Shirt!
