Low distortion, DSP based high gain servo controlled woofer controller. - diyAudio
Go Back   Home > Forums > Loudspeakers > Subwoofers

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 17th June 2013, 01:52 PM   #1
diyAudio Member
 
Tranquility Bass's Avatar
 
Join Date: Jun 2013
Location: Australia
Default Low distortion, DSP based high gain servo controlled subwoofer controller.

Hi,

This is a work in progress based on an initial prototype (REV:A) and nothing is set in concrete yet so Id thought Id share some of the preliminary results I have achieved thus far which look very promising.

I was going to put this in the subwoofer section except the main goal of this device is for accurate bass reproduction of music using a servo controlled woofer in a sealed box, open or infinite (large wall or floor mount) baffle systems. Typically it would be used as part of a multi-way passive or fully active speaker system or a stand-alone subwoofer system used to complement an existing full range speaker that cannot cover the bottom octaves. It has a fully programmable active crossover feature that can be used as a two-way stereo system with two dedicated bass servo channels. Alternatively it can be set up as a mono 3 or 4 way active system including one dedicated bass servo.

The board is designed to interface directly to a highly linear and low noise industrial grade accelerometer, namely a Measurement Specialties ACH-01-03 accelerometer. A novel mounting method directly couples the accelerometer to the voice coil which minimizes base strain and thermal coupling to the accelerometer. This accelerometer features an integral shielded flexible lead and a +/- 150g capability with a 30KHz mechanical resonance. This is the same accelerometer used in a $50,000 reference speaker system for the same type of application. It can be purchased from a number of sources including Digikey and Mouser for as little as $20.

A simple design procedure using fully automated measurement to characterize the speaker and box simplifies the entire design process without requiring any specialized or expensive equipment. No PhD required in electronics or electro-acoustics A servo woofer design can be implemented and operating in less than 30 minutes !!

Initial testing and measurements are promising with a maximum of 43 dB reduction in motor distortion achievable. Most commercial systems only provide between 6-12 dB reduction in THD.


Any questions and suggestions are welcome and any of the specifications are subject to change. I hope to make this thread a blog on my progress with this project so any feedback is welcome.

1st image - DSP servo controller,
2nd image - Test loudspeaker with accelerometer mounted to the voice coil,
3rd image - The whole setup,
4th image - Distortion results at 20Hz before feedback,
5th image - Distortion results at 20Hz after 30 dB feedback applied,

Regards
David


Attached Images
File Type: jpg P1020863.jpg (160.8 KB, 1061 views)
File Type: jpg P1020864.jpg (126.0 KB, 1032 views)
File Type: jpg P1020868.jpg (144.4 KB, 1005 views)
File Type: png Lanzar MAXP154D 20Hz V-Drive.png (36.8 KB, 989 views)
File Type: png Lanzar MAXP154D 30dB 20Hz.png (36.6 KB, 970 views)

Last edited by Tranquility Bass; 6th July 2013 at 05:59 AM.
  Reply With Quote
Old 17th June 2013, 02:20 PM   #2
Pano is offline Pano  United States
diyAudio Moderator
 
Pano's Avatar
 
Join Date: Oct 2004
Location: Milliways
Blog Entries: 4
Impressive results! You've even cleaned up some of the harmonics above 100Hz.
Would love to see a test frequency where the harmonics don't line up with the 50Hz mans.
Some 2 tone tests would be fun to see, as well.
__________________
Take the Speaker Voltage Test!
  Reply With Quote
Old 17th June 2013, 02:52 PM   #3
diyAudio Member
 
Tranquility Bass's Avatar
 
Join Date: Jun 2013
Location: Australia
I have some tests that I did at 15Hz that I will post tomorrow. Because this speaker is a car speaker its linearity is very bad below its resonant frequency which is about 30Hz. It has a lot of motor and suspension non linearity and probably a very limited X-max of 10mm or less. There is no flux shorting ring so it also has a lot of high frequency distortion as well. The drive current to produce all of the distortion was not very much which makes it ok for car use because only SPL matters in that application

However for the purposes of comparison of before and after it is the ideal test candidate. Also it can handle lots of power so mishaps are not going to kill it. Quite often I have driven the Power Physics 1000 watt amp into clipping when I was developing the DSP code

regards
David
  Reply With Quote
Old 17th June 2013, 02:59 PM   #4
Pano is offline Pano  United States
diyAudio Moderator
 
Pano's Avatar
 
Join Date: Oct 2004
Location: Milliways
Blog Entries: 4
Cool! It does look like the ideal test driver for what you are doing. Looking forward to seeing more results. Also very interested to see what your DSP will do once the speaker is mounted in an enclosure. Running before and after DSP sweeps with something like ARTA-Steps or HOLMImpulse would show how well the DSP feedback handles any problems coming from the box itself.

What the immediate goal? Will you be telling us enough about the system that we can build one ourselves?
__________________
Take the Speaker Voltage Test!
  Reply With Quote
Old 17th June 2013, 03:27 PM   #5
diyAudio Member
 
Tranquility Bass's Avatar
 
Join Date: Jun 2013
Location: Australia
Yes it's aimed at diyers or OEMs etc who want to build high quality servo controlled (sub)woofers without requiring a PhD in electronics or electro-acoustics I have done all of the ground work for them and encapsulated it in the dsp hardware/software as well as providing a user friendly windows app that communicates with the board via a USB connector

Whether the speaker is mounted in a sealed enclosure or open baffle etc ultimately the whole package is designed to quickly measure, analyse and commission a complete servo controlled woofer system from start to finish without requiring any specialized or expensive test equipment. Also the design is not tied to a particular loudspeaker or amplifier vendor although it would be unwise to use a cheap speaker with a high resonant plastic cone etc. Whilst the servo can fix up motor problems it can't perform miracles on a poorly designed speaker with bad diaphragm vibration modes and resonances

However there are no need for special drivers with sensing coils that can only be purchased from the one source. Hopefully some driver manufacturers may jump on board and build the accelerometer into their speaker so they can add more value to their products since you can now buy these accelerometers off the shelf for less than $20 in one off quantities which is a bargain compared to what you would pay for the equivalent device 10 years ago. If not the diyer can still mount the accelerometer themselves just I have done on the prototype.

It's still early days yet and the dsp code is an ongoing concern but its all starting to converge to a final end point hopefully

I will post some more measurements tomorrow as I am not at my other computer at the moment.

P.S. The same accelerometer is used in a $50,000 reference speaker system for the same type of application. Hopefully my system gives much better performance

regards
David
  Reply With Quote
Old 17th June 2013, 04:00 PM   #6
diyAudio Member
 
Join Date: Jan 2008
Impressive indeed.

Does this enable speaker to make nice 20Hz square wave?
  Reply With Quote
Old 17th June 2013, 05:44 PM   #7
Pano is offline Pano  United States
diyAudio Moderator
 
Pano's Avatar
 
Join Date: Oct 2004
Location: Milliways
Blog Entries: 4
Thanks David.
FYI, that accelerometer is $38 at Mouser here in the USA. Not bad, and it may be a useful tool for all sorts of other things, like panel resonances.
__________________
Take the Speaker Voltage Test!
  Reply With Quote
Old 17th June 2013, 05:53 PM   #8
kessito is offline kessito  Netherlands
diyAudio Member
 
Join Date: Sep 2010
Location: Amsterdam
Interesing stuff, nice job!

What is theoretical upper frequency the device can "fix" based on accellerometer and samplerate/converter latency?

cheers,
Kees
  Reply With Quote
Old 18th June 2013, 02:35 AM   #9
diyAudio Member
 
Tranquility Bass's Avatar
 
Join Date: Jun 2013
Location: Australia
Quote:
Originally Posted by kessito View Post
Interesing stuff, nice job!

What is theoretical upper frequency the device can "fix" based on accellerometer and samplerate/converter latency?

cheers,
Kees
Hello Kees

I run the ADC's and DAC's at the fastest rate of 192 KHz which minimizes total group delay to 7.5/fs or about 37.5 uS. Then add to that delay through the dsp processing and analog sections and you can expect about 4 degrees phase shift at 100Hz. Believe it or not most of the phase shift comes from the driver because above the piston range of the driver the diaphragm acts as a transmission line with many different vibration modes and exhibits a very complex impedance depending on how it is constructed and how it is terminated at the surround. This is why one vendor of servo systems uses drivers with rigid metal cones. I have an 8 inch driver with aluminium cone and I can confirm it performs much better than paper.

For stability reasons I purposely limit the maximum open loop gain bandwidth to 120Hz. Above this frequency, current feedback reduces high frequency distortion due to inductance modulation. Limiting the open loop gain bandwidth means I don't need to use drivers with special construction because the vibration modes will fall outside the bandwidth of the servo amplifier

I will post some more measurements.

regards
David

Last edited by Tranquility Bass; 18th June 2013 at 02:52 AM.
  Reply With Quote
Old 18th June 2013, 02:49 AM   #10
diyAudio Member
 
Tranquility Bass's Avatar
 
Join Date: Jun 2013
Location: Australia
Quote:
Originally Posted by Pano View Post
Thanks David.
FYI, that accelerometer is $38 at Mouser here in the USA. Not bad, and it may be a useful tool for all sorts of other things, like panel resonances.
A few weeks ago Digikey had them for $20 whilst Mouser was $40 but now they have gone back up again. Bummer I should have bought a heap of them

regards
david
  Reply With Quote

Reply


Hide 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
Trackbacks are Off
Pingbacks are Off
Refbacks are Off


Similar Threads
Thread Thread Starter Forum Replies Last Post
Passive Radiator for Low Excursion, Low Displacement & High Sens Woofer? goody75 Subwoofers 15 7th May 2012 08:07 PM
Low distortion & noise digitally controlled volume chip abraxalito Analog Line Level 6 13th August 2011 01:14 AM
Amazingly low distortion -- Radford-based design ray_moth Tubes / Valves 3 13th November 2010 02:30 PM
Low cost dsp based crossover module Crossoverman Multi-Way 38 20th February 2010 11:25 AM
Is there a no/gain preamplifier schematic using high-distortion EC90? Layberinthius Tubes / Valves 6 10th May 2004 06:13 AM


New To Site? Need Help?

All times are GMT. The time now is 10:50 AM.


vBulletin Optimisation provided by vB Optimise (Pro) - vBulletin Mods & Addons Copyright © 2014 DragonByte Technologies Ltd.
Copyright 1999-2014 diyAudio

Content Relevant URLs by vBSEO 3.3.2