Many moons ago I had a go at building the motional feedback subwoofer designed by Russel Breden in electronics world magazine, feb 97. It didn't work 
I was told that this may be because the circuit diagram in the article is wrong, and not being an electronics wizz I have no idea how to make it work.
Does anyone have plans for this subwoofer which they know work? I've bought the expensive DVC woofers and it would be a shame to waste them!
I was told that this may be because the circuit diagram in the article is wrong, and not being an electronics wizz I have no idea how to make it work.
Does anyone have plans for this subwoofer which they know work? I've bought the expensive DVC woofers and it would be a shame to waste them!
This was a very interesting project. As far as I know, there was no correction or update. Have you any more details about what could be wrong in, the schematics ? Does youy speaker work open loop ?
In AudioXpress, Dan Ferguson published a project with a similar idea without knowing previous Breden's work and using a Dayton dual voice coil driver. This may help.
In AudioXpress, Dan Ferguson published a project with a similar idea without knowing previous Breden's work and using a Dayton dual voice coil driver. This may help.
Check this post from me: http://www.diyaudio.com/forums/showthread.php?postid=987378#post987378
It's a working schematic for feedback with a microphone.
For use with a secondary voicecoil you should leave R11 and R12 out.
Connect the secondary to the two outer legs of a 10 potentiometer, connect the middle leg at the point where R11 and R12 are connected to C2.
Now one of the outer legs should be connected to ground so that the signal is in phase with the input signal.
Now you have to tune it, turn R4 all the way to the left or the right so that pins 1 and 2 of R4 are basically connected together.
Turn the circuit on and play a signal through the woofer.
Now turn the potentiometer connected to the secondary voicecoil so that the signal at the middel leg has about the same amplitude as the input signal.
Now you can start turning R4 until you start to get oscillation, then turn it back so the oscillation stops, now it ready for use.
If you have a very powerfull amplifier, connect a 60 watt lightbulb in series to protect the woofer when setting the feedback level.
This should work, assuming the signal from the secondary voicecoil represents the cone's acceleration.
I you don't understand something about it, just ask, or you might be building a one time smoke-machine.
Pherhaps you could send me the electronics world article?
It's a working schematic for feedback with a microphone.
For use with a secondary voicecoil you should leave R11 and R12 out.
Connect the secondary to the two outer legs of a 10 potentiometer, connect the middle leg at the point where R11 and R12 are connected to C2.
Now one of the outer legs should be connected to ground so that the signal is in phase with the input signal.
Now you have to tune it, turn R4 all the way to the left or the right so that pins 1 and 2 of R4 are basically connected together.
Turn the circuit on and play a signal through the woofer.
Now turn the potentiometer connected to the secondary voicecoil so that the signal at the middel leg has about the same amplitude as the input signal.
Now you can start turning R4 until you start to get oscillation, then turn it back so the oscillation stops, now it ready for use.
If you have a very powerfull amplifier, connect a 60 watt lightbulb in series to protect the woofer when setting the feedback level.
This should work, assuming the signal from the secondary voicecoil represents the cone's acceleration.
I you don't understand something about it, just ask, or you might be building a one time smoke-machine.
Pherhaps you could send me the electronics world article?
someone on here did a motion feedback system using a conductive foil inside the VC former and a metal ring ontop of the polepiece to make a basic air cored capacitor that changed value as the VC former moved, anyone remember and have a link to the project? I've looked for it a couple of times but not managed to find it.
If you look at the schematic in the pdf, you can see they used a 4th order lowpass in the feedback path, that's not a very good thing to do imho, because the phase shift of such a filter is more than 90 degrees.
A first order lowpass should be used so that a higher level of feedback can be achieved.
A first order lowpass should be used so that a higher level of feedback can be achieved.
Hi Acoustic1,
http://www.audioxpress.com/magsdirx/backissues/BISAXM.htm[/url]
audioXpress November 2003
* "A Servo Dual Voice Coil Subwoofer"
audioXpress September 2004
* "12" Dual Voice Coil Servo Subwoofer System"
http://www.audioxpress.com/magsdirx/backissues/BISAXM.htm[/url]
audioXpress November 2003
* "A Servo Dual Voice Coil Subwoofer"
audioXpress September 2004
* "12" Dual Voice Coil Servo Subwoofer System"
Hi Acoustics1,
Your information is right, there are some errors in Breden schematics.
There are two C8 : an electrolytic one in series with R11 and one in parallel with R21. This second one belongs to the integrator circuit. An adeqaute value should probably rely between 10 nF and 100 nF.
I found the connection of stability capacitor C16 (270 pF) to the low impedance outpout of A6 through C7 rather strange. I would suggest the insertion of a resistor (1 kOhm maybe) between the node (C7 - R10) and (C16 - inverting input of A5).
Your information is right, there are some errors in Breden schematics.
There are two C8 : an electrolytic one in series with R11 and one in parallel with R21. This second one belongs to the integrator circuit. An adeqaute value should probably rely between 10 nF and 100 nF.
I found the connection of stability capacitor C16 (270 pF) to the low impedance outpout of A6 through C7 rather strange. I would suggest the insertion of a resistor (1 kOhm maybe) between the node (C7 - R10) and (C16 - inverting input of A5).
If one was to use an accelerometer, what range should the accelerometer be capable of measuring?
They come in ranges of 1,7G to as high as 250G.
If you can't buy an accelerometer anywhere like me, you can buy a wiimote or the nunchuck attachment for the wiimote, it contains an ADXL330 accelerometer which is capable of measuring +/- 3.6g.
Soldering to it might be difficult, since it comes in a 4 mm × 4 mm × 1.45 mm LFCSP package.
They come in ranges of 1,7G to as high as 250G.
If you can't buy an accelerometer anywhere like me, you can buy a wiimote or the nunchuck attachment for the wiimote, it contains an ADXL330 accelerometer which is capable of measuring +/- 3.6g.
Soldering to it might be difficult, since it comes in a 4 mm × 4 mm × 1.45 mm LFCSP package.
Looking at different ways of servoing a driver, I found that velocity feedback has an interesting feature. At resonance, the velocity is in phase (or at 180°) with the input voltage.
It has been said that using one of the coil of a Dual Voice Coil driver as sensor looses efficiency due to the halving of Bl. This effectively looses the reference efficiency as defined by Richard Small.
But remembering that Duddley Harwood was complaining about too powerful driver motors, it occurs to me that subs are dealing with low frequencies and that there are good reasons to have an underdamped resonance because it enhances the local efficiency around the resonance. The feedback will bring back the desired damping.
Ideally, the Qt of a driver in a closed box sub should be around 1.1, because it is the value for which the efficiency is the highest in the resonance zone.
Considering the average spectral distribution of audio signals, I wonder if there exists an ideal resonance frequency in the same manner as there is an ideal value for Qtc.
Another interesting feature of using one of the coils of a DVC driver is that its output voltage is quite high and can be directly fed to the power amp. The Rythmik audio subs of C.-S. Ding are using this idea.
It has been said that using one of the coil of a Dual Voice Coil driver as sensor looses efficiency due to the halving of Bl. This effectively looses the reference efficiency as defined by Richard Small.
But remembering that Duddley Harwood was complaining about too powerful driver motors, it occurs to me that subs are dealing with low frequencies and that there are good reasons to have an underdamped resonance because it enhances the local efficiency around the resonance. The feedback will bring back the desired damping.
Ideally, the Qt of a driver in a closed box sub should be around 1.1, because it is the value for which the efficiency is the highest in the resonance zone.
Considering the average spectral distribution of audio signals, I wonder if there exists an ideal resonance frequency in the same manner as there is an ideal value for Qtc.
Another interesting feature of using one of the coils of a DVC driver is that its output voltage is quite high and can be directly fed to the power amp. The Rythmik audio subs of C.-S. Ding are using this idea.
Motional feedback is the way to go. It has been the progress of humankind to complete the feedback loop - in this case, right out to the speaker and beyond.
But very hard to do and not possible commercially. Feedback requires precise alignment of phase (you gotta be 180 degrees off, eh) and your means of sensing error has to good. Given how speakers change with the weather, furniture, etc. tricky to keep stable, except for a DIY person.
Doing it acoustically or using a 50-cent accelerometer (or home-brew capacitor made by gluing stuff to a dust cap) seems challenging. But if you are using speakers where coil/cone motion is fairly well related to sound wave, you have a fighting chance. That mean you have to use a sealed box (which includes, I think, the sealed box in a corner horn). But then with MF you can pump up the bass with less distorition in order to compensate for loudness shortcomings.
I don't know if using a second voice coil is more trustworthy than doubling back an error signal from the first voice coil. But I've been using a very well-performing MF system with my Klipschorn for 20 years solid-state and 20 years before that with tubes. It is a bridge design where the speaker VC is one leg of a bridge - not unlike some current feedback circuits.
Can't recommend it highly enough to bass lovers for cleanliness as well as tight bass.
Basic reference was, I think, J. Audio Eng. Soc, maybe 1954, RCA labs.... pre-internet, I know.
But very hard to do and not possible commercially. Feedback requires precise alignment of phase (you gotta be 180 degrees off, eh) and your means of sensing error has to good. Given how speakers change with the weather, furniture, etc. tricky to keep stable, except for a DIY person.
Doing it acoustically or using a 50-cent accelerometer (or home-brew capacitor made by gluing stuff to a dust cap) seems challenging. But if you are using speakers where coil/cone motion is fairly well related to sound wave, you have a fighting chance. That mean you have to use a sealed box (which includes, I think, the sealed box in a corner horn). But then with MF you can pump up the bass with less distorition in order to compensate for loudness shortcomings.
I don't know if using a second voice coil is more trustworthy than doubling back an error signal from the first voice coil. But I've been using a very well-performing MF system with my Klipschorn for 20 years solid-state and 20 years before that with tubes. It is a bridge design where the speaker VC is one leg of a bridge - not unlike some current feedback circuits.
Can't recommend it highly enough to bass lovers for cleanliness as well as tight bass.
Basic reference was, I think, J. Audio Eng. Soc, maybe 1954, RCA labs.... pre-internet, I know.
Hello!
I want to warn - I do not speak English!
I noticed that only here discusses the application of motional feedback in the subwoofer, so here and decided to write.
Many years ago I created a very simple position sensor movement, which in my opinion may be suitable for applications as a sensor of motional feedback.
On the gauges have a generator to throttle the oscillatory circuit that would bring the metallic object can drastically change its amplitude generation without changing the frequency of oscillation. It is the Eddy current sensors.
Running the generator at a frequency of about 5kHz, its amplitude frequency response is uniform in the range from 0 to 300 Hz and it has a sensitivity of approximately 10 volts per millimeter. For yet his characteristic sensitivity is not linear, but it is easy to minimize.
The sensor has some disadvantages, but they'll talk if you will find interesting.
In my video sensor is powered at 5 volts and the output is 60 volts rectified through the diode voltage. Screwdriver attachment to the throttle sensor causes a drastic change in the readings - follow the arrow voltmeter.
Further, we can see that the oscillation frequency remains unchanged. Then, the throttle sensor is close to metal objects mounted on the speaker cone, and it connects to the source of 50 Hz.
My letter in English to help translate the Google. I hope that he did it good.
Well, please watch my video: YouTube - ??????? ?????? ???????????-?????????
If you have any questions for me, please ask them clear and short, so I could understand them.
Thank you very much for your attention.
I want to warn - I do not speak English!
I noticed that only here discusses the application of motional feedback in the subwoofer, so here and decided to write.
Many years ago I created a very simple position sensor movement, which in my opinion may be suitable for applications as a sensor of motional feedback.
On the gauges have a generator to throttle the oscillatory circuit that would bring the metallic object can drastically change its amplitude generation without changing the frequency of oscillation. It is the Eddy current sensors.
Running the generator at a frequency of about 5kHz, its amplitude frequency response is uniform in the range from 0 to 300 Hz and it has a sensitivity of approximately 10 volts per millimeter. For yet his characteristic sensitivity is not linear, but it is easy to minimize.
The sensor has some disadvantages, but they'll talk if you will find interesting.
In my video sensor is powered at 5 volts and the output is 60 volts rectified through the diode voltage. Screwdriver attachment to the throttle sensor causes a drastic change in the readings - follow the arrow voltmeter.
Further, we can see that the oscillation frequency remains unchanged. Then, the throttle sensor is close to metal objects mounted on the speaker cone, and it connects to the source of 50 Hz.
My letter in English to help translate the Google. I hope that he did it good.
Well, please watch my video: YouTube - ??????? ?????? ???????????-?????????
If you have any questions for me, please ask them clear and short, so I could understand them.
Thank you very much for your attention.
Your English was OK
Subwoofer cones are heavy enough and slow enough to carry standard industrial position sensors, LVDT, linear optical grating and rotary encoder. Closed loop control of a subwoofer does not seem that difficult provided the loop response is quick, using current drive of the voice coil, would speed up the loop response by charging the VC inductance faster also it would eliminate power compression. I used a current drive VC in a current to pressure transducer once.
Linear optical grating would be my choice, 0.01mm positional accuracy, linear at any excursion and fast enough, a rotary encoder would be easier mechanically. The linear encoder could be placed inside the hollow centre pole used on many subwoofers and linked to the VC cap with some elastomer.
Digital output is a bonus making it possible remove analogue entirely from the signal processing chain, a dac is still needed to control the current amplifier but that is inside the feedback loop.
Hi, metasculptor!
If I understand you correctly, you think that my sensor must be installed on the speaker cone. But it is not. On the speaker cone, you must install only a thin strip of metal bent to the required shape and the sensor placed on the frame of the speaker (as was shown in my video). Its value lies in the fact that it is very easy to install and besides, it's very cheap.
The detected its signal can be fed into the amp without any digital conversion. May need to set notch filter at 5 kHz and bandwidth of the amplifier to limit the frequency of 300Hz.
I do not know what a VC.
Thank you.
If I understand you correctly, you think that my sensor must be installed on the speaker cone. But it is not. On the speaker cone, you must install only a thin strip of metal bent to the required shape and the sensor placed on the frame of the speaker (as was shown in my video). Its value lies in the fact that it is very easy to install and besides, it's very cheap.
The detected its signal can be fed into the amp without any digital conversion. May need to set notch filter at 5 kHz and bandwidth of the amplifier to limit the frequency of 300Hz.
I do not know what a VC.
Thank you.
Sorry you misunderstood, I was suggesting to install an industrial optical sensor on the speaker cone.
I understand how your eddy current sensor works. Eddy current sensors are very popular for motor vehicle detection. The pickup coil size must get larger as the subwoofer cone excursion increases to maintain sensitivity.
The notch filter cannot be too sharp or it will exclude the sidebands of the signal being detected in this case 5Khz +/- 300Hz
The amplifier bandwidth is limited by the active crossover network which drives it.
Voice Coil
Cheers
Metalsculptor.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.