DIY 3D Speaker Scanner - the Mathematics and Everything Else

[No. 5]

Sorry, I don't mean to interrupt a sub-thread with my reflections question, but Mark's comment helped me clarify what's nagging at me. The sound field separation identifies reflections from the stationary surfaces in the measurement environment. The blue vertical member in my diagram isn't really stationary--it tracks around an arc at the same time the mic does. If the mic moves radially (red arrow) while the vertical member stays a fixed radial distance away (no blue arrow motion), will the processing remove the reflection from the vertical member? Or does the processing only remove truly stationary sources of reflection?

Few

Short answer: it should remove it.

Long answer: I believe that sound field separation works the same as a sound intensity measurement in that the system is aware of the arrival vector of any sound passing through the bounding measurement surfaces and can consequently ignore anything coming from the "wrong" direction. So for your concept, as long as the vertical microphone support is outside of measurement surfaces and far enough away from the microphone itself to allow a decent IR window, the system will "see" the reflection as coming from a fixed cylindrical surface, but coming from the same "wrong" direction as all the reflections off the walls and other surfaces in the room.

Edit: also, notice how similar your concept is to Klippel's NFS, so it's clearly workable. A side note on that, I've heard that their patent includes a method to address reflections of reflections (i.e. sound leaving the speaker, bouncing off of a surface then bouncing off the speaker and getting into the mic). I have no idea how important that is to include in a DIY version.

 

[mwmkravchenko]

Point taken on the Vector averaging. Getting a different method of making a measurement is indeed what we are all chasing. My guess is that this is baked into the mix on the Klippel system as well. You can argue that an anechoic measurement will only show you dips if they are truly from the device under test. What still happens in a math refined measurements is an approximation. Everyone that has ever measured more than once or twice realizes that what we get at that time, at that particular measurement will pretty much never get truly repeated. It is a representation of the device under test. Over about 20 years of measuring I have learned that there is good enough. I do my best to turn a blind eye to my OCD and get on with the work. This applies to what we can expect from anything that is doing an addition or subtraction of the measurements of our device under test.

Oh last little point. I believe that No. 5 has proven that as few as 5 measurements can be used with the moving mic and math functions in REW. That comparison was against a Klippel measurement of the same type of loudspeaker. Not the exact loudspeaker tested under the Klippel system. Might be good to throw that comparison up on this thread?
I put up my groundplane versus REW magic a while back in this thread.

Mark

 

[Few]

Videos of Klippel's crazy-complex mic positioner seem to me to show the vertical member (not the horizontal tube the mic is mounted to) moving radially. And the horizontal structure supporting the vertical member is surprisingly close to the mic and speaker. I'm a bit mystified that they get away with so many large surfaces at a fixed and short distance from the mic. My understanding is that the directionality of the sound is determined by comparing measurements made with different radial distances between the mic and sound source.

I guess the good news is that it implies it may not be necessary to go crazy trying to make the structure as acoustically invisible as possible. Nonetheless, it sure runs counter to my apparently faulty intuition.

Few

 

[mwmkravchenko]

A little more detail in between the sales pitch. I see a zero axis point measurement.

Mark

 

[No. 5]

I believe that No. 5 has proven that as few as 5 measurements can be used with the moving mic and math functions in REW. That comparison was against a Klippel measurement of the same type of loudspeaker. Not the exact loudspeaker tested under the Klippel system. Might be good to throw that comparison up on this thread?

Can do! Just had to find it again... I'm not great at documentation...

This is one of the three Behringer B2030P's that I have and measured via beamforming compared to the Klippel data from the ASR reviewed B2030P. Top to bottom is Klippel NFS; beamforming 41 points vector averaged (approximately 1" point spacing starting about 3' from the speaker); beamforming 4 points vector averaged (about 10" spacing). Not sure what's going on above 3kHz, if it's my B2030P or something else.

Edit: I should also note that all of my measurements were made outdoors. This method works better for me when it has less reflections to deal with.

 

[No. 5]

I'm a bit mystified that they get away with so many large surfaces at a fixed and short distance from the mic.

I am also a bit surprised by how close the structure is to the microphone. I know they are using sound field separation up to 1.5kHz, but you'd think you'd want enough breathing room around the mic to allow a 3 millisecond (or so) IR window... but maybe there is enough breathing room or they're using a shorter IR window, or there's something else at play.

My understanding is that the directionality of the sound is determined by comparing measurements made with different radial distances between the mic and sound source.

Here's a video describing a different application for the method of sound capture/field separation that, I think, makes a nice simple explanation of what's happening. One thing not mentioned in the video (but is important!) is that that one of the two microphone positions has inverted polarity with respect to the other.

I guess the good news is that it implies it may not be necessary to go crazy trying to make the structure as acoustically invisible as possible.

Indeed!

 

[Few]

As I sketch mechanical designs for a mic scanning device, I’m reminded of an issue I raised probably 2-3 years ago—stepper motor noise. Mechanically, I think it would be helpful to mount the motor that moves the horizontal mic boom vertically, right at the joint between the mic boom and the vertical member (really wish I hadn’t introduced the word “member”). But that puts the stepper motor close to the mic. Are there any stepper motor experts who can chime in with a way to get the steppers to stop buzzing while they’re just holding a stationary position? That buzzing, if unavoidable, would need to be as far from the mic as possible, so I’m hoping there’s an easy way to avoid it.

Few

 

[fluid]

Some controllers have PWM controls to adjust the current chopping behaviour. I have tweaked this with my own Tic Controller and it can make quite a bit of difference to the "sound" of the stepper.

https://www.ti.com/lit/an/slvaes8/s...01017&ref_url=https%3A%2F%2Fwww.google.com%2F

 

[KLang1]

Just use a TMC 2209 or any other stepper motor driver with stealthchop. My printer is dead silent apart from the mechanical noise. There's plenty of cheap 3d printer boards that fit them and a Nema17 stepper is strong enough to move a microphone around. And if it isn't, add gearing.

 

[NTK]

Or use stepper (or servo) motors with brakes. Move to position, engage the brakes, turn off the motor, and complete silence.
The motors are more expensive, though. Example:

https://catalog.orientalmotor.com/item/2-phase-unipolar-stepper-motors/shop-pkp-series-2-phase-unipolar-stepper-motors/pkp244mu12m

 

[mwmkravchenko]

As I sketch mechanical designs for a mic scanning device, I’m reminded of an issue I raised probably 2-3 years ago—stepper motor noise. Mechanically, I think it would be helpful to mount the motor that moves the horizontal mic boom vertically, right at the joint between the mic boom and the vertical member (really wish I hadn’t introduced the word “member”). But that puts the stepper motor close to the mic. Are there any stepper motor experts who can chime in with a way to get the steppers to stop buzzing while they’re just holding a stationary position? That buzzing, if unavoidable, would need to be as far from the mic as possible, so I’m hoping there’s an easy way to avoid it.

Few

Why not put the stepper on the bottom, better center of gravity, and use a long timing belt loop to move your mic up and down. Also the motor with brake is kind of a one stop solution. As Klang1 mentions there seems to be another way too. Might balance out in the end though.

Mark

 

[mwmkravchenko]

Acoustic Measurement – Nicholas Martin

nicholasmart.in

Am I late to the party here? Maybe you guys already know about this. Just learned of it from diyaudio similar thread.

I understand what the reason for ARTA is now.

John Mulcahy when do you do a turntable interface? Maybe secondary lifting the mic interface?

 

[Few]

The reason I am thinking of having the motor move is that I like this “ServoBelt” design which essentially eliminates timing belt stretch even when long runs are needed. In case it’s not clear, there are two belts, one with the teeth facing up, and another with the teeth facing down, and meshing with the other belt. A short loop in the upper belt is wrapped around the driving gear on the motor. When the motor turns, that loop propagates up or down the upper belt so the motor, plus whatever it is attached to, moves along the belt axis. The only part of the belt that is under tension is that short loop, so there is essentially no stretch. Bell-Everman came up with this and commercialized it a decade ago or so. You just need to be sure to use belts whose teeth mate with themselves (not all profiles work).

Obviously this approach isn’t the only way to go, but I’ve always been intrigued and this seems like an opportunity to play with it.

Few

 

[mwmkravchenko]

VEVOR 2 x HSR 15-1500 mm Linear Guideway Rail with 4Pcs Pillow Block Carriage Bearing Block for 15 mm Slotted Block Square Block Linear Rail Routers CNC Set | VEVOR CA

Buy VEVOR 2 x HSR 15-1500 mm Linear Guideway Rail with 4Pcs Pillow Block Carriage Bearing Block for 15 mm Slotted Block Square Block Linear Rail Routers CNC Set at cheap price online, with Youtube reviews and FAQs, we generally offer free shipping to Europe, US, etc.

www.vevor.ca

Canukistani Vevor, but I'm pretty sure that it will localize to where you are. Two rails 1.5 meters in length. So you have an easy up and down and and in and out should you wish. Mounted to a suitable stiff plywood rear frame and you have a fairly easy system.

Right now it's the life of the self employed. two contracts on the go and neither will pay out until about a month. So dreaming only in terms of the buying stuff. I do however have a lot of plywood! So, That will be something that happens next.

Turntable I have posted before.

Amazon.com

www.amazon.com

If you wish there are ones with a USB interface.

 

[mwmkravchenko]

View attachment 348301
The reason I am thinking of having the motor move is that I like this “ServoBelt” design which essentially eliminates timing belt stretch even when long runs are needed. In case it’s not clear, there are two belts, one with the teeth facing up, and another with the teeth facing down, and meshing with the other belt. A short loop in the upper belt is wrapped around the driving gear on the motor. When the motor turns, that loop propagates up or down the upper belt so the motor, plus whatever it is attached to, moves along the belt axis. The only part of the belt that is under tension is that short loop, so there is essentially no stretch. Bell-Everman came up with this and commercialized it a decade ago or so. You just need to be sure to use belts whose teeth mate with themselves (not all profiles work).

Obviously this approach isn’t the only way to go, but I’ve always been intrigued and this seems like an opportunity to play with it.

Few

Very cool idea. Something just occurred to me. An ACME thread is self locking. If a lead screw was used, it would be a simple application of what works without to much experimentation. Again just throwing out bones to see what gets picked.

Mark

 

[mwmkravchenko]

 

[Few]

Yeah I considered a threaded rod of some sort (acme, ball screw, etc.) but I’m assuming the vertical range of motion has to be pretty big. As I write that, though, I don’t have any firm numbers to base that on—what vertical range do we actually need?

In any case, a large range of motion would necessitate a very long screw of whatever sort, and they tend to be prone to ”whipping” if they’re spun too fast. During a measurement I don’t think there would be much risk, but if the system were asked to make a quick jog to the “home” position, for example, the rate of that motion might be quite limited if a long skinny screw were used. For diy purposes, just living would a slow homing process would likely not be a big deal, but it’s one of the compromises to bear in mind.

I also looked into racks for a rack and pinion system, but long ones aren’t cheap.

Few

 

[mwmkravchenko]

Yeah I considered a threaded rod of some sort (acme, ball screw, etc.) but I’m assuming the vertical range of motion has to be pretty big. As I write that, though, I don’t have any firm numbers to base that on—what vertical range do we actually need?

In any case, a large range of motion would necessitate a very long screw of whatever sort, and they tend to be prone to ”whipping” if they’re spun too fast. During a measurement I don’t think there would be much risk, but if the system were asked to make a quick jog to the “home” position, for example, the rate of that motion might be quite limited if a long skinny screw were used. For diy purposes, just living would a slow homing process would likely not be a big deal, but it’s one of the compromises to bear in mind.

I also looked into racks for a rack and pinion system, but long ones aren’t cheap.

Few

Ame is self locking negating the need for a brake. A ball screw is the polar opposite. Whipping, well fixed on two ends and can be fixed in the moving nut area just as you see on a lathe lead screw.

Height. I have been a pondering. Ideally you want the DUT to have a rather central location in your measurement room. Driver spacing and cabinet dimensions into account we can come up with a little bit of dimensions. Floor standers are basically best designed with the tweeter. Let's call it 42 inches or 1007mm. A book shelf will need to be on a stand. You loose about 4 inches about 105mm to the turntable. We are at a top cabinet height now of 46 inches or 1170mm. Considering that a top to bottom height for a cabinet is only about a meter or 39 3/8th inches you need a useable travel of this dimension. That would need the length of the mic boom holding part added on the top of the travel and the bottom of the travel. Generously we could do it in about 12 inches total or about 310mm. So, total height? 58 inches. Something 1500mm would work perfectly that's basically 60 inches. If you have a taller cabinet you can use a plinth under the mic boom and get to where you need in height.

Those Vevor rectangular linear rails fit the bill nearly perfectly. Both for vertical and horizontal. Just saying! Hard to beat Vevor for cost versus utility.

Now, to find an appropriate ACME all thread. There's McMaster Carr. The load is tiny so the smallest that they have is probably perfect. A hunting I will go.

Mark

 

[mwmkravchenko]

V-Slot NEMA 17 Linear Actuator Kit T8 ACME Leadscrew drive with Nema17 Motor | eBay

This NEMA 17 Screw-Driven Linear Actuator System is great for high thrust, stiffness and tight precision in high-load, moderate-speed applications. V-Slot NEMA 17 Linear Actuator Kit (Lead Screw). Nema 17 Stepper Motor-76oz in 1.

www.ebay.com

Covers one meter in length, but that is not too bad. It's an all in price. So that makes it rather interesting.

Mark

P.S. still hunting for the 1.5m

V-Slot® NEMA 17 Linear Actuator Bundle (Belt Driven) | eBay

This V-Slot® Linear Actuator (Belt Driven) is a fast and easy solution to a V-Slot® Gantry System. (1) NEMA 17 Stepper Motor. Actuator Specs VersatileV-Slot Linear Rail. Bundle Contents GT2-2M Timing Belt for top speed.

www.ebay.com

Loading…

www.aliexpress.com

 

[mwmkravchenko]

Acoustic Measurement – Nicholas Martin

nicholasmart.in

Am I late to the party here? Maybe you guys already know about this. Just learned of it from diyaudio similar thread.

I understand what the reason for ARTA is now.

John Mulcahy when do you do a turntable interface? Maybe secondary lifting the mic interface?

Automatic Polar Measurements using ARTA, stepper motor and Tic Controller

I have been in the process of researching and preparing to build an automatic measurement table for a while as making a large number of measurements manually is no fun and very time consuming. I came across a great website I thought had everything I needed laid out and documented...

www.diyaudio.com

It's seems so late to the party was I that there is a solution to the problem!

Mark