Go Back   Home > Forums > >
Home Forums Rules Articles diyAudio Store Blogs Gallery Wiki Register Donations FAQ Calendar Search Today's Posts Mark Forums Read

Software Tools SPICE, PCB CAD, speaker design and measurement software, calculators

Klippel Near Field Scanner on a Shoestring
Klippel Near Field Scanner on a Shoestring
Please consider donating to help us continue to serve you.

Ads on/off / Custom Title / More PMs / More album space / Advanced printing & mass image saving
Reply
 
Thread Tools Search this Thread
Old 18th February 2018, 07:32 AM   #21
Dave Zan is offline Dave Zan  Australia
diyAudio Member
 
Join Date: Nov 2010
Quote:
Originally Posted by aslepekis View Post
Exactly! That is pretty much the set-up that I had been think...
I had a think myself and maybe have a tidier idea than the Klippel mechanics.
Kind of a cantilever version of the Weinriech setup.
Speaker is mounted on on a post.
Version 1 - a strut from the bottom of the post, pivoted so it can rotate around the post, a tie wire from the top of the post to hold the strut out at radius.
This provides azimuth control with minimal structure to disturb the sound field.
Can be easily automated to step the azimuth for version 2.
The end of the strut has a horizontal pivot for a quarter-circular cantilever arm that sweeps out the polar co-ordinate, small mike on the end.
The arm could be carbon fibre, thin, not heavy, minimal disturbance but still accurate placement.
So for on-axis measurement the strut is to the side at 90, the arm is horizontal, around the equator.
For vertical polars the arm is rotated up and down, easy and cheap to automate for V3.
This provides true spherical coordinate measurements, which simplifies the maths.
Make sense?

Best wishes
David

I note that the Klippel system is not $20,00 as Earl claimed but nearer $50,000 - just for hardware.
Software is $60,000, presumably they consider it not an easy problem.
If you commercialize I definitely want royalties

Last edited by Dave Zan; 18th February 2018 at 07:49 AM.
  Reply With Quote
Old 18th February 2018, 04:06 PM   #22
gedlee is offline gedlee  United States
diyAudio Member
 
gedlee's Avatar
 
Join Date: Dec 2004
Location: Novi, Michigan
The Legendre polynomials Lgnd(x) are an orthogonal set defined between +1 and -1 along x. Fitting in x is best done on an equally spaced set of points along x. But to get into the spherical coordinate theta we must set x = cos(theta). Equal spacing along x is not equal spacing in theta. My spacing somewhat corrects that situation, but not entirely - it is still better than equal spacing. In reality equal spacing in theta works just about as well, but with maybe a few more points required for comparable accuracy.
__________________
Earl Geddes Gedlee Website
  Reply With Quote
Old 19th February 2018, 10:05 AM   #23
Dave Zan is offline Dave Zan  Australia
diyAudio Member
 
Join Date: Nov 2010
Quote:
Originally Posted by gedlee View Post
...spherical coordinate theta we must set x = cos(theta)...
You use theta for the azimuth? For spherical coordinates I remember most of the physics literature uses phi for azimuth and theta as polar, which is inconsistent with many maths books, so I just want be sure I have it clear.

If we want equal cos(theta) then wouldn't that imply symmetrical intervals for the back and front?
That's intuitively unreasonable, we care more about the front, but I can't see how the mathematics will play out.
Similarly, I have looked at Klippel's scanner presentations and the illustrations show a cylindrical scan - this seems sub-optimal, wouldn't a spherical scan with fewer azimuth steps near the poles be more efficient?
But promotional material probably doesn't have the most careful illustrations, some of the pictures of spherical harmonics look faulty too.

Best wishes
David


Best wishes
David
  Reply With Quote
Old 19th February 2018, 12:38 PM   #24
gedlee is offline gedlee  United States
diyAudio Member
 
gedlee's Avatar
 
Join Date: Dec 2004
Location: Novi, Michigan
The math usually assumes that the "north" pole is along the drivers axis, i.e. tilted towards the axis of symmetry (see Morse.) Yes, the cos(theta) does imply a finer resolution both forward and backwards, but backwards was clearly not of as much interest as forwards so I made it lessor resolution in the back. One could completely exclude the back without too much loss of information.
__________________
Earl Geddes Gedlee Website
  Reply With Quote
Old 20th February 2018, 01:20 AM   #25
aslepekis is offline aslepekis  United States
diyAudio Member
 
Join Date: Sep 2011
Quote:
Originally Posted by Dave Zan View Post
I had a think myself and maybe have a tidier idea than the Klippel mechanics.
Kind of a cantilever version of the Weinriech setup.
Speaker is mounted on on a post.
Version 1 - a strut from the bottom of the post, pivoted so it can rotate around the post, a tie wire from the top of the post to hold the strut out at radius.
This provides azimuth control with minimal structure to disturb the sound field.
Can be easily automated to step the azimuth for version 2.
The end of the strut has a horizontal pivot for a quarter-circular cantilever arm that sweeps out the polar co-ordinate, small mike on the end.
The arm could be carbon fibre, thin, not heavy, minimal disturbance but still accurate placement.
So for on-axis measurement the strut is to the side at 90, the arm is horizontal, around the equator.
For vertical polars the arm is rotated up and down, easy and cheap to automate for V3.
This provides true spherical coordinate measurements, which simplifies the maths.
Make sense?
I like the simplicity, but I'm not sure I'm picturing it correctly. I sketched how I saw the description (the one on the left). Clarifications are welcome.
The one on the right is one of my early ideas for the scanner hardware.
Quote:
Originally Posted by Dave Zan View Post
I note that the Klippel system is not $20,00 as Earl claimed but nearer $50,000 - just for hardware.
Software is $60,000, presumably they consider it not an easy problem.
If you commercialize I definitely want royalties
If I commercialize, you deserve royalties.

The first article I read about the NFS pointed out that it's a fifth of the starting price for a decent anechoic chamber, so it's a value priced system... like a used Rolls Royce.

Quote:
Originally Posted by Dave Zan View Post
...I have looked at Klippel's scanner presentations and the illustrations show a cylindrical scan - this seems sub-optimal...
My assumption is that they chose a cylindrical scanning surface so that the hardware would be easier to build and be more portable.
Quote:
Originally Posted by gedlee View Post
One could completely exclude the back without too much loss of information.
That sounds like it could then be worth also perusing a variation of this device that can measure a raw driver in a baffle. I recall you had mentioned something along the lines of using your measurement technique for drivers in a baffle in the other thread.
Attached Images
File Type: jpg IMAG0922.jpg (865.7 KB, 171 views)
  Reply With Quote
Old 20th February 2018, 01:40 AM   #26
Dave Zan is offline Dave Zan  Australia
diyAudio Member
 
Join Date: Nov 2010
Quote:
Originally Posted by aslepekis View Post
I sketched how I saw the description (the one on the left). Clarifications are welcome...
Almost, the strut and tie is correct
Now put the strut out to the side of the speaker.
On the end of the strut is a pivot and arm to scan in theta
The arm is a quarter circle that follows the equator to the front of the speaker.
The mike is perpendicular at the end of the arm, pointed at the speaker.
The mike could be manually moved in and out, to do a 2 surface scan.
That makes the mike mount very small for minimal disturbance, and one simple adjustment when necessary.

Picture later if this is not clear.
The rationale is that there is minimum structure and simple pivots, much cheaper and easier to automate than linear slides.
The Klippel mechanics seem quite clumsy, to be honest.


Best wishes
David

Last edited by Dave Zan; 20th February 2018 at 02:09 AM.
  Reply With Quote
Old 20th February 2018, 06:17 AM   #27
Dave Zan is offline Dave Zan  Australia
diyAudio Member
 
Join Date: Nov 2010
Quote:
Originally Posted by gedlee View Post
The math usually assumes that the "north" pole is ...the drivers axis
Thanks, that clarifies it.
The Klippel scanner illustrations have a vertical axis so I just used the same.
But your assumption makes more sense for an axisymmetrical driver.
My own speakers are not axisymmetric, perhaps that unconsciously influenced me too.

Best wishes
David.
  Reply With Quote
Old 21st February 2018, 02:40 AM   #28
aslepekis is offline aslepekis  United States
diyAudio Member
 
Join Date: Sep 2011
Quote:
Originally Posted by Dave Zan View Post
Almost, the strut and tie is correct
Now put the strut out to the side of the speaker.
On the end of the strut is a pivot and arm to scan in theta
The arm is a quarter circle that follows the equator to the front of the speaker.
The mike is perpendicular at the end of the arm, pointed at the speaker.
The mike could be manually moved in and out, to do a 2 surface scan.
That makes the mike mount very small for minimal disturbance, and one simple adjustment when necessary.

Picture later if this is not clear.
The rationale is that there is minimum structure and simple pivots, much cheaper and easier to automate than linear slides.
The Klippel mechanics seem quite clumsy, to be honest.


Best wishes
David
Ahh, I think I've got it now. See attached (I didn't sketch the pivot for theta very well, but it's at the side of the speaker). I like the simplicity of it, certainly preferable to trying to do something like Klippel's Z axis.

I'm thinking the length of the microphone mount (D1 in the sketch) can assist in reducing the acoustic disturbance; with enough distance a long enough impulse window can be had to remove any reflections from the hardware but still provide enough resolution to provide detail down to around 1kHz or lower... dimensions depending, of course. Making those dimensions favorable to a long IR window could easily make the whole thing awkwardly large. I'm supposing that a two layer scan with field separation would help even more, we just have to get to that point.
Attached Images
File Type: jpg IMAG0923_BURST002.jpg (400.1 KB, 157 views)
  Reply With Quote
Old 21st February 2018, 04:08 AM   #29
Dave Zan is offline Dave Zan  Australia
diyAudio Member
 
Join Date: Nov 2010
Quote:
Originally Posted by aslepekis View Post
...See attached (I didn't sketch the pivot for theta very well, but it's at the side of the speaker). I like the simplicity of it
Thank you, I think it's a pretty neat solution, if I may say so myself.

Quote:
...impulse window can be had to remove any reflections from the hardware...a two layer scan with field separation would help even more...
It should be possible to have cleaner IR than the Klippel structure and still keep the size reasonable.
Even without 2 layer scan the results should be as accurate as any other technique short of an anechoic chamber.

Quote:
we just have to...
I see the problem in 3 steps.
First - 1-d horizontal scan with Legendre polynomial solution, essentially what Earl's PolarMap does. Room reflections are removed by time window.
Second - Full spherical scan with spherical harmonic solution, as Earl notes in the PolarMap doco, this is rather more difficult.
Third - 2 layer scan to remove room reflections, this requires the radial expansion in Hankel functions (or equivalently Bessel).
I still don't have an intuitive feel for the third step partly because I don't know how Klippel combines time window and 2 layer measurements.
Does Klippel have a manual for their scanner?
@Earl, do you have any ideas about this?

Best wishes
David

I will have a look at the Klippel patents and see what I can learn.

Last edited by Dave Zan; 21st February 2018 at 04:18 AM.
  Reply With Quote
Old 21st February 2018, 11:51 AM   #30
Dave Zan is offline Dave Zan  Australia
diyAudio Member
 
Join Date: Nov 2010
@Aaron, I looked at the Klippel patent, fortunately it covers exactly the part I don't understand, the 2 layer reflection extraction process... unfortunately it's not very simple.
Probably why that module of software is ~ $24,000.
I think I need to find a copy of "Fourier Acoustics" by Williams.

Best wishes
David
  Reply With Quote

Reply


Klippel Near Field Scanner on a ShoestringHide this!Advertise here!
Thread Tools Search this Thread
Search this Thread:

Advanced Search

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off

Forum Jump

Similar Threads
Thread Thread Starter Forum Replies Last Post
Klippel R&D Measurement Methodology SUBduction Multi-Way 3 11th December 2016 04:37 PM
TB W3-1878 Klippel graph aarvin2 Full Range 3 18th November 2012 03:04 PM
Why doesn't DIYAUDIO have a klippel? JZatopa The Lounge 26 14th November 2011 02:40 AM
Vinyl Scanner? Nehesi Analogue Source 37 20th July 2011 08:29 AM
TD15M, does anybody have one willing to donate for Klippel testing? thadman Multi-Way 71 4th September 2009 03:10 AM


New To Site? Need Help?

All times are GMT. The time now is 04:25 AM.


Search Engine Optimisation provided by DragonByte SEO (Pro) - vBulletin Mods & Addons Copyright © 2018 DragonByte Technologies Ltd.
Resources saved on this page: MySQL 14.29%
vBulletin Optimisation provided by vB Optimise (Pro) - vBulletin Mods & Addons Copyright © 2018 DragonByte Technologies Ltd.
Copyright ©1999-2018 diyAudio
Wiki