Hi folks,
I've spent my whole Labor Day (US) holiday reading about servo subwoofers, the different types, their strengths and weaknesses etc. ...until my eyes can take no more.
I gather from several sources that the accelerometer servo variety is the superior type, but I see only (very) expensive completed/retail units available.
Is there in fact, a pre-built kit of this type that I can purchase and design an enclosure for? I thought if anybody would know it would be this group.
If there is, could someone point me in that direction? (I'm assuming there is not)
Then there is the voice coil variety that is offered by Rythmik. From what I can gather there are problems with this type acting as a transformer, especially in the upper bass area, and the amount of distortion reduction is not as good as the accelerometer type.
A step back: What draws me to servo is not so much the reduction in distortion, but the transients being better defined. Being able to tame big EQ in a small enclosure too. Don't get me wrong, very low distortion is desirable, but I believe that mostly 2nd order distortion is not that audible in the lower frequencies. So yes, the transients being 'tighter' is of much more importance.
Another thing is, that I'll be crossing this/these over at 50Hz. Do you folks think that the voice-coil type of servo would display the negative transformer behavior if crossed over at this frequency?
As you probably can guess, I'm on the fence here, and would very much appreciate any advice you could give me.
I did see the Analog servo sub thread that Armand posted, and while a lot of that went over my head, the results in the tests he provided were quite impressive. Given that though, how difficult is that circuit to build? and what other associated equipment would I need?
In conclusion: Being that I'm a busy guy with limited time, and limited knowledge about starting, (and more importantly, completing) an accelerometer type servo sub, do you think that the voice-coil type of servo would be a good 'middle ground' between a regular sub and an accelerometer style option? And at least provide the transient (and other) benefits I am after?
Again, I would greatly appreciate any and all recommendations, comments, advice etc. ..whether it be of a positive or negative nature.
Cheers
-Steve
I've spent my whole Labor Day (US) holiday reading about servo subwoofers, the different types, their strengths and weaknesses etc. ...until my eyes can take no more.
I gather from several sources that the accelerometer servo variety is the superior type, but I see only (very) expensive completed/retail units available.
Is there in fact, a pre-built kit of this type that I can purchase and design an enclosure for? I thought if anybody would know it would be this group.
If there is, could someone point me in that direction? (I'm assuming there is not)
Then there is the voice coil variety that is offered by Rythmik. From what I can gather there are problems with this type acting as a transformer, especially in the upper bass area, and the amount of distortion reduction is not as good as the accelerometer type.
A step back: What draws me to servo is not so much the reduction in distortion, but the transients being better defined. Being able to tame big EQ in a small enclosure too. Don't get me wrong, very low distortion is desirable, but I believe that mostly 2nd order distortion is not that audible in the lower frequencies. So yes, the transients being 'tighter' is of much more importance.
Another thing is, that I'll be crossing this/these over at 50Hz. Do you folks think that the voice-coil type of servo would display the negative transformer behavior if crossed over at this frequency?
As you probably can guess, I'm on the fence here, and would very much appreciate any advice you could give me.
I did see the Analog servo sub thread that Armand posted, and while a lot of that went over my head, the results in the tests he provided were quite impressive. Given that though, how difficult is that circuit to build? and what other associated equipment would I need?
In conclusion: Being that I'm a busy guy with limited time, and limited knowledge about starting, (and more importantly, completing) an accelerometer type servo sub, do you think that the voice-coil type of servo would be a good 'middle ground' between a regular sub and an accelerometer style option? And at least provide the transient (and other) benefits I am after?
Again, I would greatly appreciate any and all recommendations, comments, advice etc. ..whether it be of a positive or negative nature.
Cheers
-Steve
Last edited:
Yes, I have read the information on their website numerous times, and this may be the direction I take *IF* I cannot find a accelerometer based system that is:
1) Fairly simple to obtain parts for..
2) Fairly easy to build.
3) Is in a relative 'ballpark' price-wise.
I appreciate the link, but this is from 2000. Also, it may or may not be comparable to the DIY build that Armand posted in the http://www.diyaudio.com/forums/subwoofers/239941-analog-servo-sub.html thread, which being more current, likely to be using parts that are more easily acquired. Plus.. I see no purchasing information. (correct me if I am wrong).
I definitely appreciate the responses though... thank you...
Thanks for the response! Those use the 'direct servo' tech that Rythmik is known for, and IIRC, are available in some of their subs.
However, the 'direct servo' tech is the same as the voice-coil servo I describe in my first post. Which, as stated earlier, might be my only option if I can't find an accelerometer type servo tech that meets my criteria in post #4.
Understand that velodyne does this like no one else. Read past the subwoofer page. News
If you're going to do the box...do it right. I remember hearing a plain dd18 years ago...you can't do better than that. Louder...yes. Better...you need a sealed attic and a hole cut in your ceiling. Or an extra room to use as a sub enclosure. Infinite baffle...much more cost effective.
This hobby caused me to go back to college for electrical engineering. Some linkwitz guy and lx stuff...neurochrome...devialet basically.
Anderson .Paak - Heart Don't Stand a Chance
If you're going to do the box...do it right. I remember hearing a plain dd18 years ago...you can't do better than that. Louder...yes. Better...you need a sealed attic and a hole cut in your ceiling. Or an extra room to use as a sub enclosure. Infinite baffle...much more cost effective.
This hobby caused me to go back to college for electrical engineering. Some linkwitz guy and lx stuff...neurochrome...devialet basically.
Anderson .Paak - Heart Don't Stand a Chance
Thanks for the response! Those use the 'direct servo' tech that Rythmik is known for, and IIRC, are available in some of their subs.
However, the 'direct servo' tech is the same as the voice-coil servo I describe in my first post. Which, as stated earlier, might be my only option if I can't find an accelerometer type servo tech that meets my criteria in post #4.
Motional Feedback Audio.
There was an accelerometer based project that hinted a kit would be available from tentlabs, but I don’t know if that ever happened.
https://www.by-rutgers.nl/PDFiles/Motional Feedback.pdf
DIY modules
At best, VC-MFB will reduce suspension dominated distortion that shows up at 20Hz and below by a factor of 2 or 3. Don’t expect much if any distortion reduction above the bottom octave. Most quality woofers will provide useful VC feedback below 50Hz that is untainted by transformer coupling with the driven VC. Getting up to 100Hz would require careful woofer selection. If you have the free-air impedance curve for a driver of interest, it is not too difficult to model when the coupling starts to dominate the VC feedback signal. I can post an example of how to do this later in the week if anybody is interested. Obviously any of the woofers offered by GR-Research will work just fine. Note that they also offer the amps with servo input already incorporated. Although they are a little pricey, that would provide you with a turnkey solution.
Did you find some data that suggested transients would be improved for a MFB sealed subwoofer if compared to a non-MFB sealed subwoofer EQ’d to have the same response? Or were you more talking about improved transient response for a woofer when it is driven past its Xmax.
Unless you use the same woofer/sensor/enclosure as the designer, accelerometer type servos will require you to be comfortable measuring loop gain & phase margin and adjusting filters in the feedback path to ensure stability. VC type servo is comparatively simple, requiring just a single summing opamp to add the servo loop to an existing plate amp and generally has no issues with stability.
Thanks very much Bolserst, both for the link provided and the advice.
(I looked at the tentlabs site and don't see a kit anywhere, but it looks like I could replicate the circuit if I thought I needed to do so.)
Also, a lot has been discussed on bentoronto's thread too, and your responses over there have been very educational as well.
So after I processed all of that to the best of my ability, I still have a question that is going to determine whether I cancel my order for the Rythmik kit(s).
(I still have a week)
Do you think that the amount of control over the transducer that Rythmik's servo provides, can be replicated with simple EQ/filters in a ported design?
Or do you think that Rythmik's VC servo system provides some advantage over and above what could be achieved with EQ/filters? Again, ported box here.
Thanks
-Steve
(I looked at the tentlabs site and don't see a kit anywhere, but it looks like I could replicate the circuit if I thought I needed to do so.)
Also, a lot has been discussed on bentoronto's thread too, and your responses over there have been very educational as well.
So after I processed all of that to the best of my ability, I still have a question that is going to determine whether I cancel my order for the Rythmik kit(s).
(I still have a week)
Do you think that the amount of control over the transducer that Rythmik's servo provides, can be replicated with simple EQ/filters in a ported design?
Or do you think that Rythmik's VC servo system provides some advantage over and above what could be achieved with EQ/filters? Again, ported box here.
Thanks
-Steve
There is another type which operates by mounting an electret mic capsule at the edge of the dust cap, measuring the pressure. I made a sub about 6 years ago using that concept, and it generally worked very well (actually better than some that I made using the ACH-01 accelerometers).
Some limitations, though -- the electret capsules are limited in how high a level they can handle, so you'd need to bias them in a 3-wire circuit (source follower) with higher voltage so they can have a bit more high level handling before they distort. And you need to use them in a large woofer (12" or more) or in a set of multiple slaved woofers. The pressure the capsule measures is local (not the total pressure seen by the cone) so a large woofer or several is needed to produce the desired useful levels while keeping the capsule linear.
There are also possible positioning problems, since the pressure they see can depend on what surfaces are nearby -- the setup can oscillate if the box is adjusted for open space in front but the woofer is placed facing a wall or floor. As with pretty much any of these circuits, you need to adjust the feedback level and possibly some filter parameters depending on the type of woofer and the box size to get the most useful amount of feedback with stability -- breakup resonances in the cone, for instance, can cause large amplitude peaks and wild phase variations at which oscillation can easily happen. That also limited the amount of distortion reduction I've ever been able to get -- getting to even 14dB loop gain (theoretically a distortion reduction of about 5x) at the lowest frequency isn't easy, but you don't reduce harmonic distortion by near that much because the loop gain drops about 6dB with every doubling of frequency so by the 3rd harmonic there isn't a lot of feedback in play. Rather than reduced distortion, the main benefits tended to be flatter, controlled frequency response, and reduction of output f1 compression (first order distortion rather than harmonic distortion).
Some limitations, though -- the electret capsules are limited in how high a level they can handle, so you'd need to bias them in a 3-wire circuit (source follower) with higher voltage so they can have a bit more high level handling before they distort. And you need to use them in a large woofer (12" or more) or in a set of multiple slaved woofers. The pressure the capsule measures is local (not the total pressure seen by the cone) so a large woofer or several is needed to produce the desired useful levels while keeping the capsule linear.
There are also possible positioning problems, since the pressure they see can depend on what surfaces are nearby -- the setup can oscillate if the box is adjusted for open space in front but the woofer is placed facing a wall or floor. As with pretty much any of these circuits, you need to adjust the feedback level and possibly some filter parameters depending on the type of woofer and the box size to get the most useful amount of feedback with stability -- breakup resonances in the cone, for instance, can cause large amplitude peaks and wild phase variations at which oscillation can easily happen. That also limited the amount of distortion reduction I've ever been able to get -- getting to even 14dB loop gain (theoretically a distortion reduction of about 5x) at the lowest frequency isn't easy, but you don't reduce harmonic distortion by near that much because the loop gain drops about 6dB with every doubling of frequency so by the 3rd harmonic there isn't a lot of feedback in play. Rather than reduced distortion, the main benefits tended to be flatter, controlled frequency response, and reduction of output f1 compression (first order distortion rather than harmonic distortion).
Hi wreckingball,
Take a look at patent #5,537,479 (1996, Kreisel, Miller and Kreisel) it describes a combination of push-pull driver mounting and voice coil mounted sensor motion feedback. The data provided shows basically what can be achieved w/ and w/o MFB.
A further improvement can be had by using djk's PPSL arrangement which might just work well w/ the method bwaslo recommends in Post #13.
Regards,
Take a look at patent #5,537,479 (1996, Kreisel, Miller and Kreisel) it describes a combination of push-pull driver mounting and voice coil mounted sensor motion feedback. The data provided shows basically what can be achieved w/ and w/o MFB.
A further improvement can be had by using djk's PPSL arrangement which might just work well w/ the method bwaslo recommends in Post #13.
Regards,
Thanks guys, duly noted, and I'll look closer at those options later.
Since this is my thread, I guess I'm entitled to ask a potentially stupid question: In bentoronto's thread in post #224 here: http://www.diyaudio.com/forums/subw...back-woofer-available-sort-6.html#post4826153
Can someone confirm for me that KSTR is describing:
"a" as accelerometer based servo/MF system
"b" as voice coil based servo/MF system
"c" as just EQ'ed system
Is this correct?
If so, it would seem that a and c have potentially more stability problems than b. And that a and c are most suited for sealed enclosure.
*edit* @ tb46: Excuse my ignorance, but what is "djk's PPSL arrangement" ?
Cheers
-Steve
Since this is my thread, I guess I'm entitled to ask a potentially stupid question: In bentoronto's thread in post #224 here: http://www.diyaudio.com/forums/subw...back-woofer-available-sort-6.html#post4826153
Can someone confirm for me that KSTR is describing:
"a" as accelerometer based servo/MF system
"b" as voice coil based servo/MF system
"c" as just EQ'ed system
Is this correct?
If so, it would seem that a and c have potentially more stability problems than b. And that a and c are most suited for sealed enclosure.
*edit* @ tb46: Excuse my ignorance, but what is "djk's PPSL arrangement" ?
Cheers
-Steve
Last edited:
Hi wreckingball,
I should have posted a link:
http://www.diyaudio.com/forums/subwoofers/177905-thread-those-interested-ppsl-enclosures.html
PPSL stands for push-pull slot-loaded, it refers to a mounting arrangement of pairs of woofers in a plenum, and can be applied to any of the basic enclosure formats (sealed, vented, tapped, horn).
Regards,
I should have posted a link:
http://www.diyaudio.com/forums/subwoofers/177905-thread-those-interested-ppsl-enclosures.html
PPSL stands for push-pull slot-loaded, it refers to a mounting arrangement of pairs of woofers in a plenum, and can be applied to any of the basic enclosure formats (sealed, vented, tapped, horn).
Regards,
I saw you question in the other thread, but was holding off answering hoping KSTR would respond. My understanding is that he was saying there are three possible amplifier output impedances you can use to drive a woofer. “b” is the usual voltage source amplifier, so very low output impedance, “a”is negative output impedance which if carefully managed can cancel out VC resistance and improve damping. The negative output impedance can be generated with positive current feedback .“c” is if you use a current source amplifier which has a very high output impedance. It is ill-suited to drive woofers thru their resonance range. The high output impedance can be generated with negative current feedback.
I always wondered if this patent using dual accelerometers to cancel odd harmonics and push-pull mounting to cancel even harmonics was an effort to get around the Velodyne patent. I don’t believe they ever offered any Hi-fi products using the technique, although I recall some pro-equipment that did. If you mount one of the woofers with magnet facing out, you need to make sure the mechanical suspension and noise from air pumping thru the voice coil gap is low for the woofer you choose. Otherwise, as I found out, you may have greatly reduced even harmonic distortion but can’t appreciate it because of all the harmonically unrelated noises generated from the back end of the outward facing woofer.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.