Thread moved to subwoofers section
WoW, This is incredible work!!
I would love to implement this type of sub system in my ESL setup!!
Keep up the great work!!
jer 🙂
I would love to implement this type of sub system in my ESL setup!!
Keep up the great work!!
jer 🙂
Yes it is ideal for this sort of application as the ESL's are known for their low distortion so they should be teamed up with a low distortion bass reproduction system. With the built in active crossover and the 8 band parametric bass equalizer you'll have the ability to tweak the system to your hearts content 😉
regards
david
regards
david
Negative feedback looping around the speakers: the final frontier.
I experimented and enjoyed motional feedback for a few decades. Benefits are substantial including to distortion and bass extension. Might also have some mysterious benefits to room acoustics but I have yet to see that explored. Like with a lot of enhancements, biggest benefits are to worsest and cheapest systems and that's what often attracts commercial interests.
But I always argued that it made little sense to control a precise $150 woofer using a junky 50-cent accelerometer. (Of course the Philips economics are using a 50-cent accelerometer to control a $5 driver, so that made a bit more sense, at the time. Philips also had an RCA patent they had to avoid crossing.) But your new system looks promising. (Until recently, I've advocated using feedback derived from putting the voice coil in a bridge.)
Don't forget, any way you derive cone motional feedback, the theory applies only to stiff cones in sealed boxes and true infinite baffles (maybe horns too especially with sealed boxes behind the driver).
Thanks for great thread. Your replies are very helpful and terribly polite (could you really be Canadian?).
Ben
I experimented and enjoyed motional feedback for a few decades. Benefits are substantial including to distortion and bass extension. Might also have some mysterious benefits to room acoustics but I have yet to see that explored. Like with a lot of enhancements, biggest benefits are to worsest and cheapest systems and that's what often attracts commercial interests.
But I always argued that it made little sense to control a precise $150 woofer using a junky 50-cent accelerometer. (Of course the Philips economics are using a 50-cent accelerometer to control a $5 driver, so that made a bit more sense, at the time. Philips also had an RCA patent they had to avoid crossing.) But your new system looks promising. (Until recently, I've advocated using feedback derived from putting the voice coil in a bridge.)
Don't forget, any way you derive cone motional feedback, the theory applies only to stiff cones in sealed boxes and true infinite baffles (maybe horns too especially with sealed boxes behind the driver).
Thanks for great thread. Your replies are very helpful and terribly polite (could you really be Canadian?).
Ben
Last edited:
Revision B is on the drawing board as I type this. I have changed the amplifier interface to a more general purpose interface that requires an external adaptor board which is designed for each specific amplifier type. Not all amplifiers will be suitable as there are certain house keeping signals and feedback which is needed by the dsp board for proper interface and sequencing. However the recommended amplifiers will be off the shelf devices sourced from third parties 😉
regards
david
Yes you are quite correct. By its very nature motional feedback can only correct motor distortion and not fix up a poorly designed diaphragm which has bad resonances and vibration modes. Because these resonances occur above the effective range of the feedback these resonances will appear whether or not you use feedback, and give the speaker its characteristic bad or colored sound. The feedback will however correct distortion in the piston range of the driver. The other negative aspect of these resonances is that they can effect the stability margin of the feedback system and can turn it into an oscillator sometimes at very high frequencies 🙁
My preference is to use stiff metal coned diaphragms which are made out of aluminum or titanium etc. These will maintain integrity of the diaphragm to a much higher frequency allowing much higher loop gain and bandwidth therefore reducing distortion further whilst minimizing distortion from cone flexure at low frequencies. They should sound better as well 😉 This probably explains one of the reasons why Genesis Loudspeakers use metal cone woofers in their servo systems 😉 You can get away with a softer cone material provided it has lots of damping but then it may physically distort and flex at lower frequencies which causes distortion that the feedback cannot correct. Other materials such as a mixture of Kevlar and paper as well as rohacell should be considered.
regards
david
Last edited:
Have you considered acoustical feedback from a microphone directly in front of the woofer? There are very small microphone capsules available today so they would not interfere with the sound of the driver.
Check out Mr. Eraths work.
I have the last two servo models he made before he passed and possibly the last unit he made.
His were "Electronic suspension" systems and used a simulated speaker as a reference (in parallel with the real woofer) then made real time changes to the feedback section with the error signals.
I have the last two servo models he made before he passed and possibly the last unit he made.
His were "Electronic suspension" systems and used a simulated speaker as a reference (in parallel with the real woofer) then made real time changes to the feedback section with the error signals.
Not sure as I know what you are saying.
However, with voice-coil-in-a-bridge feedback, if you are kind of anal, you make one of the legs of the bridge resemble the various lumped non-active electric elements of the driver. In other words, the goal is to balance the bridge for DC and steady-state conditions (AKA veridical motion). That way, the only unbalance (AKA as the feedback signal) arises from false motions of the voice coil (AKA errors in the driver's acoustic output).
There is some obfuscation in this thread between (1) pre-emptive correction based on modeling driver misbehaviour and (2) the miracle of negative feedback. You can have both in a system but they are unrelated concepts.
Ben
Mr. Erath was doing Servo woofers in the 60's, and yes he used the miracle of negative feedback.
The "new" miracle of DSP would compliment his analog model nicely.
The "new" miracle of DSP would compliment his analog model nicely.
Sounds like he was using exactly Werner bridge method I described (RCA patent). Frankly, I think the modeling is grad-school or Camden Lab stuff or maybe added to make the efforts patentable (a common motivation in the commercial world that DIYer sometimes fail to perceive) and does not materially improve results.
If the Erath woofers used the Werner-type feedback*, they no doubt were good speakers.
Sure, DSP is nice but I don't see any connection.
Please share with us your reasons for lionizing this Mr. Erath (who I never heard of before today). Publications? References? Links to URLs besides his devoted followers?
Ben
*they seem to make big bass from small boxes (AKA bass extension resulting from motional feedback) which is characteristic.
If the Erath woofers used the Werner-type feedback*, they no doubt were good speakers.
Sure, DSP is nice but I don't see any connection.
Please share with us your reasons for lionizing this Mr. Erath (who I never heard of before today). Publications? References? Links to URLs besides his devoted followers?
Ben
*they seem to make big bass from small boxes (AKA bass extension resulting from motional feedback) which is characteristic.
Last edited:
AudiogoN Forums: Louis W Erath - In Memoriam
diyAudio.com Wiki - projects by fanatics, for fanatics
The great grand pappy of servo speakers...
diyAudio.com Wiki - projects by fanatics, for fanatics
The great grand pappy of servo speakers...
Last edited:
Nonsense. Erath may have been a good fellow but he is a near total non-entity in the history of motional feedback. His 60's product(s) are likely based on Werner's 1957 JASA (or JAES) RCA-lab publication... and all kinds of forerunners like experiments with positive current feedback in earlier decades.
If you want to make a useful contribution to this discussion, why don't you sketch his circuit for us - I assume it is made from discrete components.
Please, let's get back to this great thread.
Ben
Last edited:
Unless he used some sort of independent sensing device to provide a signal proportional to the cone motion the amount of distortion reduction from this type of arrangement would be minimal. From what I have seen of the patents nothing like this was ever employed in any of his designs.
regards
David
Yes have tried this but the feedback loop tries to correct for any low frequency pressure disturbances in the room which is not related to the audio signal and makes the cone gyrate wildly sometimes. Also if there is more than one bass reproducer in the room the interaction between them could pose a problem with this arrangement. The accelerometer solution doesn't suffer from this problem because it doesn't respond to pressure.
regards
David
Last edited:
- Status
- Not open for further replies.