Low distortion, DSP based high gain servo controlled woofer controller. - Page 2 - 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 18th June 2013, 04:01 AM   #11
diyAudio Member
 
Tranquility Bass's Avatar
 
Join Date: Jun 2013
Location: Australia
Quote:
Originally Posted by Barleywater View Post
Impressive indeed.

Does this enable speaker to make nice 20Hz square wave?
As long as the speaker uses a rigid diaphragm it is quite possible to accurately produce a 20Hz square wave

regards
david
  Reply With Quote
Old 18th June 2013, 06:51 AM   #12
diyAudio Member
 
Join Date: Nov 2012
Location: Melbourne
Quote:
Originally Posted by Tranquility Bass View Post
Hi,

4th image - Distortion results at 20Hz before feedback,
5th image - Distortion results at 20Hz after 30 dB feedback applied,

Hi David,

The level of fundamental (20Hz) in 5th image is about 2dB lower than in 4th image (before feedback).

Are you able to provide feedback results for fundamental of exactly the same level as in the 4th image?.

Best Regards,
Bohdan
  Reply With Quote
Old 18th June 2013, 07:41 AM   #13
diyAudio Member
 
Tranquility Bass's Avatar
 
Join Date: Jun 2013
Location: Australia
No problems
Attached Images
File Type: png Lanzar MAXP154D 30dB@20Hz.png (35.7 KB, 502 views)
  Reply With Quote
Old 18th June 2013, 08:05 AM   #14
diyAudio Member
 
Tranquility Bass's Avatar
 
Join Date: Jun 2013
Location: Australia
Default Speaker/Accelerometer 15 Hz Single Tone Test

This test shows a worst case scenario which emphasizes suspension and motor distortion and the effectiveness of feedback to clean it up This speaker produces loads of distortion below it's resonant frequency. The lower the frequency the more dominant the distortion mechanisms. I would not recommend it for HiFi use

1st Image - Distortion Spectrum with Voltage Drive and no feedback
2nd Image - Distortion Spectrum with Servo and 30 dB loop gain
3rd Image - Distortion Spectrum with Servo and 30 dB loop gain + boost
4th Image - Oscilloscope shot with Voltage Drive and no feedback*
5th Image - Oscilloscope shot with Servo and 30 dB loop gain*

* Oscilloscope top yellow waveform is the accelerometer output
* Oscilloscope bottom blue waveform is the voice coil current
Attached Images
File Type: png Lanzar MAXP154D V-Drive@15Hz.png (37.1 KB, 479 views)
File Type: png Lanzar MAXP154D 30dB@15Hz.png (37.0 KB, 464 views)
File Type: png Lanzar MAXP154D 30dB+Boost@15Hz.png (36.5 KB, 461 views)
File Type: gif LANZAR MAXP154D V-Drive@15Hz.gif (18.0 KB, 464 views)
File Type: gif LANZAR MAXP154D 30dB@15Hz.gif (17.5 KB, 117 views)

Last edited by Tranquility Bass; 18th June 2013 at 08:18 AM.
  Reply With Quote
Old 18th June 2013, 08:11 AM   #15
diyAudio Member
 
Tranquility Bass's Avatar
 
Join Date: Jun 2013
Location: Australia
Default Speaker/Accelerometer 110 Hz Single Tone Test

The following measurements compare the effectiveness of current drive alone in reducing high frequency distortion. About a 14 dB reduction in distortion can be achieved with current drive depending on the drive level. Unfortunately current drive is worse than voltage drive at frequencies below the resonant frequency which is where the global feedback is effective

1st image - Distortion Spectrum with Voltage Drive and NO global feedback
2nd image - Distortion Spectrum with Current Drive and NO global feedback
Attached Images
File Type: png Lanzar MAXP154D V-Drive@110Hz.png (36.1 KB, 155 views)
File Type: png Lanzar MAXP154D I-Drive@110Hz.png (37.2 KB, 119 views)

Last edited by Tranquility Bass; 18th June 2013 at 08:22 AM.
  Reply With Quote
Old 18th June 2013, 09:18 AM   #16
Armand is offline Armand  Norway
diyAudio Member
 
Join Date: Aug 2005
Impressive!
Will for sure be interested in buying one of these when you finish!
5 years ago I made an analogue servobass system with the same accelerometer and was able to drop THD from 26% to 5% on my Peerless XLS 10" woofer. This was at full X-MAX of 12mm and at 23Hz. I ran into problems with the soft dust-cap that created resonances at about 500Hz. I had to mount the accelerometer directly over the voice coil which again created wrong balance in the cone. The accelerometer weighs 4 grams and when the g-forces reached about 70g the coil crashed into the magnet
Anyway, it seems your system has much better potential. Have you looked into mounting one of these super-small chip accel availiable?
http://www.analog.com/static/importe...ts/ADXL193.pdf
  Reply With Quote
Old 18th June 2013, 10:34 AM   #17
superR is offline superR  Netherlands
diyAudio Member
 
Join Date: Nov 2006
Location: Rotterdam
Can you show me the open loop in a bode plot? It looks nice, but the DSP*Wooferdynamics is what is the really interesting part.

Ideally a separate DSP implementation and a separate Woofer-model.

Next to that, have you measured the latency of the DSP from analog in to analog out?
  Reply With Quote
Old 18th June 2013, 12:53 PM   #18
diyAudio Member
 
Join Date: Dec 2003
Location: Mumbai
Dear David ,

Kindly drop your email id to dasgupta.s@outlook.com .

Suranjan
  Reply With Quote
Old 18th June 2013, 03:08 PM   #19
diyAudio Member
 
Tranquility Bass's Avatar
 
Join Date: Jun 2013
Location: Australia
Quote:
Originally Posted by superR View Post
Can you show me the open loop in a bode plot? It looks nice, but the DSP*Wooferdynamics is what is the really interesting part.

Ideally a separate DSP implementation and a separate Woofer-model.

Next to that, have you measured the latency of the DSP from analog in to analog out?
A lot of it is proprietary so can't share too many details, suffice to say that I use a novel approach that allows me to define the frequency response or speaker alignment independent of the servo amplifier response. Most servo designs have to build the closed loop frequency response around the servo amplifier and the requirements of stability. Currently I use a simple 2nd order high pass shaping filter and I can select the cutoff frequency and Q independent of the servo amplifier bandwidth and gain. You could use other shaping filters such as 3rd or 4th order Butterworth or Bessel etc which would give additional subsonic filtering but poorer group delay and transient response.

Regarding latency the total group delay or transit time of the adc and dac chain is 7.5/fs @192KHz. This is about 37.5uS. But add to that the extra dsp processing and delay in the analog circuitry gives about a net 4 degree phase shift at 100Hz. To put that into perspective the phase shift of the compensation of the servo amplifier swamps the group delay by a large margin whilst still maintaining unconditional stability

regards
david
  Reply With Quote
Old 18th June 2013, 03:27 PM   #20
diyAudio Member
 
Tranquility Bass's Avatar
 
Join Date: Jun 2013
Location: Australia
Quote:
Originally Posted by Armand View Post
Impressive!
Will for sure be interested in buying one of these when you finish!
5 years ago I made an analogue servobass system with the same accelerometer and was able to drop THD from 26% to 5% on my Peerless XLS 10" woofer. This was at full X-MAX of 12mm and at 23Hz. I ran into problems with the soft dust-cap that created resonances at about 500Hz. I had to mount the accelerometer directly over the voice coil which again created wrong balance in the cone. The accelerometer weighs 4 grams and when the g-forces reached about 70g the coil crashed into the magnet
Anyway, it seems your system has much better potential. Have you looked into mounting one of these super-small chip accel availiable?
http://www.analog.com/static/importe...ts/ADXL193.pdf
Yes I bought some NOS ADXL250's some time ago from ebay but didn't use them in the end. The problem with some of the mems devices is they have much more noise compared to the piezo devices. The ACH-01-03 is a very low noise device which makes it ideal for this type of app.

There are even better Measurement Specialities devices I have looked at such as the 805 and 805M1 but they are more expensive and need to be mounted on a pcb etc and don't have an integral flexible coax lead.

Mounting of the accelerometer is crticial. As soon as you mount it on a dust cap you create a mechanical resonance depending on the weight of the accelerometer and the mechanical compliance of the dust cap. This resonance can vary from a few hundred Hz to the KHz region and makes this type of mounting totally futile in most cases. The end result is to make the system unstable and turn it into an oscillator.

I have an 8 inch aluminum coned woofer and I mounted the accelerometer onto the aluminum dustcap using bluetack of all things. It works quite well but exhibits a sharp peak resonance of 15 dB at 1.3KHz. Without filtering the system will oscillate at about 1.3KHz. But I have built in a user defined fully customizable anti-modal filter which can essentially null out this resonance and eliminate any instability At the end of the day the accelerometer should be tightly coupled to the voice coil but even then you have to deal with modal resonances in the speaker cone. Again the anti-modal filter built into the dsp software comes to the rescue

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 09:07 PM
Low distortion & noise digitally controlled volume chip abraxalito Analog Line Level 6 13th August 2011 02:14 AM
Amazingly low distortion -- Radford-based design ray_moth Tubes / Valves 3 13th November 2010 03:30 PM
Low cost dsp based crossover module Crossoverman Multi-Way 38 20th February 2010 12:25 PM
Is there a no/gain preamplifier schematic using high-distortion EC90? Layberinthius Tubes / Valves 6 10th May 2004 07:13 AM


New To Site? Need Help?

All times are GMT. The time now is 12:26 PM.


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