What subwoofers have servos?

Velodyne

Genesis

My 17 year old F-1200 Velodyne is a servo sub. Still working great to this day!

From Rythmik's web site...

http://www.rythmikaudio.com/technology.html

Jeff

Yes, I know!! But according to Rythmik, they do servo differently!

Jeff

My old Infinity is a servo type.

Quick google brings up Martin Logan, Paradigm and even Yamaha, and that was just the first page of results.

I could be wrong but I thought the Paradigm subs in the newer models have no mention of servo.

Velodyne uses an accelerometer on the driver and associated electronics to give a corrective feedback signal. So, it is part mechanical and part electronic.

Rythmik uses something like 2 resistors and a capacitor for a very simple and inexpensive Servo that is all electronic, no mechanical accelerometer.

I believe each company claims benefits for their servo, which is what you would expect.

The Velodynes are low distortion and so are the Rythmiks.

It would be nice to see how the Seaton Submersive HP would compare with something like the Velodyne DD-18 Plus, which has been tested by Audioholics. The thing is the Velodyne has an MSRP of $5,000 while the Submersive is about $2,300.

Based on about 5 years of AVS member user experience with the Submersive, and more recent experience with the Submersive HP, it can hold its own with virtually every reasonable competitor, and do very well against subs costing much more than $2,300.

My thx ultra 2 rated Jamo d7 subs were servo controlled, they sounded great for music to me but never really excited me for home theater. The drivers in them looked pretty pathetic compared to some of the more robust drivers round these parts. There may be both a rythmic and a submersive at the upcoming southern California subwoofer meet. That ought to shed some thoughts on the question. Although unless done blind take comments with a grain of salt.

The only properly designed servo speakers are sealed box and use accelerometers to sense the cone motion. There are other ways to sense the cone motion but they have to be electrically and magnetically isolated from the main voice coil to work properly.

Infinity, Genesis, Velodyne, Phillips and Paradigm all use accelerometers in their servo speaker offerings for good reason

"comparing" one that is similar in price...

i had a paradigm servo 15 v2 (still lives on in my golf partners setup)... i would not trade one of my subm hp's to bring it back....

ymmv, that's just one man's opinion...

Rythmik Direct Servo - velocity and current sensing.

I only want to focus on the improvements of servo since it still takes a well-engineered driver. Servo is like the icing on a cake or whatever else you want it to be. It took me a while to write this since I wanted to make it simple enough to understand. Hopefully you understand some of the T/S parameters of drivers. If anybody notices that I mistyped something, please let me know. I tried to be as objective as possible.

The sensing coil reacts directly to velocity. It also sits in the same magnetic gap. It is able to react almost instantaneously. The latency is very short since and it does not need to convert measurements before it can correct. How quick the servo can respond is important in each of the points. The sensing coil weighs about 1 gram and is about 6 mils thick (6*2 for each side of coil is 12 mils total).


I am referring to sealed subs in general throughout this post to make it simpler besides the FV15HP link haha.



A) Coherence - The system does not suffer from amplitude modulation distortion as long as the amplifier has power on tap (aka power compression). The servo also creates a very linear response. Peaks and dips in the response cause extra ringing.

Under 25 Hz for the green line and the orange line are with amplifier clipping. Otherwise the response curve stays the same. Check both graphs in post 3.
http://www.hometheatershack.com/foru...html#post45117

Here is another example measured by Josh Ricci. Under 30 Hz is port compression (ported sub). I linked to the two port mode so it is easier to focus on the lack of thermal compression due to less port compression. Scroll down to the long term power compression sweeps and chart.
http://www.avsforum.com/avs-vb/showt...3#post20602943





B) The servo provides extra control to the cone. All servos reduce the Qts of the system but these use different methods (sensing coil, accelerometer, motor based). There are now three Q values. Every sub has a physical Q value. This is the Qtc of a driver in the box (The Qtc of a driver in a sealed enclosure is higher than the drivers Qts. The larger the box ,the smaller the Qtc of the enclosure. For Qtc to equal Qts, the enclosure would have to be near infinite in size). Non-servo subs system Q is the same as the physical Q. The Direct Servo reduces the system Q by about 3x. Those subs with EQ can also change the third Q or frequency response Q. After equalization, the Qtc of 0.5 theoretically produces no ringing. Here is a quick summary from Brian posted below.

1) The physical Q value. Physical Q value determines the energy transfer ratio between our amps and speaker. For instance, even though 0 voice coil resistance can eliminate a lot of distortion sources, it causes an undesirable stall effect that the frequency response follows a -6db/oct starting from 100hz all the way to 5hz which is not very good at all. The best physical Q value is around 0.7 to 0.9 to ensure max energy transfer in the pass band.

2) The system Q value determined by the equivalent circuit should be as low as possible as it is proportional to "equivalent" Rvc. Currently it is 0.3 for 12" driver and 0.25 for 15" driver. Lower equivalent Rvc increases damping force. When you push the cone, you can feel a stronger resistive force pushing you back. This characteristic is best observed in free air because once you are in a box and you really need to push fast enough to emulate the frequencies above 10hz (otherwise, you will be just excercising the spring force of a sealed box). In a spring-mass-friction system, resistive force is the only one that dissipates energy (this is also how we are different accelerometer-based approach, which implements a very large spring force in the equivalent circuit, and spring force does not dissipate energy). It also means the cone would transfer less and dissipate more of the standing wave inside the enclosure. Moreover, if the system is accidentally push out of their linear operation range (such as clipping) or deviated from their linear path due to nonlearity of the system, it can recover 3x faster than nonservo as soon as it is back to linear range. The result is a sense of fast bass under almost all condition.

3) Final time domain/frequency domain Q value. Although system Q should be as low as possible as we push below 0.25, the final frequency response from the sub stay above 0.5. Anything below 0.5 can sound very lean. For this reason, we have provided three settings for everyone to try (0.5, 0.707, 1.0) and most listeners like 0.5.



To demonstrate the effectiveness of a lower system Q (more cone control and damping), Brian made a large enclosure. Inside the enclosure, he placed another driver. This inside driver is the active one and played the signal. The second driver faced the external environment as a typical sealed sub. The second driver was a reactor. The active driver inside played a sweep while SPL was recorded outside the reactor driver. This was done with the reactor driver without servo and also with it. This test shows how much of the bass energy from the back wave moves through the cone. The graph at the bottom of the link shows the increased damping of the servo sub. The number is the difference between the non-servo driver and the servo driver.
http://www.rythmikaudio.com/re-radiation.html

I uploaded this video to the Official Rythmik thread to show the improvement in real life. After flipping the switch to the servo, it feels as if the cone is sitting in a viscous liquid. There is some rocking of the driver when the servo is on. It was hard to do this video by myself with one hand pushing, another hand flipping the switch, and trying to hold the magnet of the driver with my feet. NOTICE - DON'T TRY THIS WITH AN ACCELEROMETER SUB. I DON'T WANT TO BE RESPONSIBLE IF ANYTHING HAPPENS.







C) Non-linear distortion. The Bl profile of a driver shows the motor force of the driver depending on the displacement of the cone from the center position. One equation for calculating motor force at the weakest point is Bl^2/Re. The servo reduces the Re by 3x. So, to solve for the 15 driver, we have (15.8^2)/(3/3) = 250 rounded up. Rythmik has a driver coil and also a sensing coil (weighs about 1 gram). The sensing coils Bl profile replaces the driver's profile (shown in the link). It is linear without reducing efficiency. A linear Bl curve helps reduce non-linear distortion.
http://www.avsforum.com/avs-vb/showt...9#post15541849

Here is a link to show the reduction of primarily spider and surround distortion. Scroll down to distortion reduction unless the whole article interests you.
http://www.rythmikaudio.com/memory.html





D) Reduction of Memory Effect - This is tied to point 1 (Coherence). It is more complex but is talked about in the link above and the two links below.

http://www.avsforum.com/avs-vb/showp...&postcount=156

http://www.avsforum.com/avs-vb/showp...&postcount=170




A Quick Analogy: remember analogies are never completely correct.
Say we have a professional weight lifter who specializes in bench press and an Olympic gymnast who specializes in the still rings. Both weigh the same and are the best in their sport for producing upper body force. We now test them in the bench press. The weight lifter will most likely be able to put up more weight than the gymnast. Now we have both of these people perform a more complex test such as a planche on the rings. The weight lifter has the motor force but does contain the neurological coordination and feedback to stay true to the signal. Even if we went to a simpler movement on the rings, the weight lifter may come close but would still be shaking and producing unwanted movement. A gymnast also sometimes loses control (amp clipping) but is able to react faster to move back into position. Hope this helps if the above was too complicated. I wrote this real quickly after proofreading.

Accelerometer

I don’t really now much about this form of servo. Ask Benny how long it takes for the signal to reach the accelerometer after the signal reaches the motor. Then add how long it takes for the accelerometer to convert acceleration to pressure and then to electrical energy. I don’t know the answer.

What is the mass of an accelerometer?

You can look at these measurements of the DD-18+ and DD-15+

Click on the graphs so you can see if there is any power compression and to show if the response is linear. You might even see power expansion, which is when the next sweep has more output at certain frequencies due to the implementation of the servo.

http://www.audioholics.com/reviews/s...8-measurements

http://www.audioholics.com/reviews/s...ments-analysis

I also don’t know how much the accelerometer servo reduces non-linear distortion. Correct me if I am wrong but don’t accelerometer subs use limiters to protect the system from oscillation? If the limiter limits the output it should appear low so how can we say distortion is reduced? The DD-18+ servo gain also changes the frequency response so it is hard to get a fair comparison.

Perhaps Benny will chime in.

That's an absurd claim. Since the sense coil is wrapped tightly around the drive coil what about the effects mutual coupling, field and inductance modulation due to the position and current in the voice coil ? All of these things contribute to distortion and lack of coherence !! The only way a voice coil servo can work is if the sense coil is in a separate magnetic circuit away from the main driving coil. The sense coil should only respond to the velocity of the diaphragm only. It should NOT be contaminated with any other signals such as mutual coupling from the main drive coil and field and inductance modulation due to voice coil displacement.

In fact the greater the drive signal and current and the greater the diaphragm displacement the less effective this servo system is in reducing distortion simply because the sensing of the diaphragm becomes less effective and accurate. This system is far from perfect no matter how tightly the coil is stuck onto the driving coil !

A 3 to 5 gram accelerometer is hardly significant compared to a 100 to 500 gram diaphragm mass. Surely you are not going to tell me this is an issue ? And why would there be a delay in the signal from an accelerometer when the accelerometer is tightly coupled to the voice coil ?

And why do Velodynes have displacement limiters ?? Quite simply because no speaker can be driven past their mechanical limits no matter how good the servo system is !! Like any high gain servo system including negative feedback audio amplifiers, limits need to be placed in the feedback loop otherwise overload conditions can set the loop into oscillation. This phenomena is also observed on a poorly design amplifier which uses negative feedback where clipping causes the output to stick or momentarily burst into oscillation !! On a non servo or open loop system the speaker would just breakout into huge distortion or even damage but the apparent loudness due to the added distortion would make the speaker sound more impressive compared to a limited system but this is surely not what you want when you talk about coherence

And look what I just found !!

http://www.hometheatershack.com/foru...ealed-56l.html

Looks like there's a limiter in there somewhere or is it part of the measurement procedure that limits distortion

I honestly asked for you input on accelerometer subs since I know very little about these and you seem to know the theoretical advantages of using an accelerometer over an open loop system. I'm trying to help the OP since he asked and there are different implementations. I did not mind the DD-18+ when I heard it. It sounded pretty good but could only compare to others in the store. A few of us may get together this summer where both types of servo will be available in the same setting. If you live in the area, you are welcome to come by and enjoy.

I assume piezoelectric accelerometers are used. I thought these were placed on the bottom of the dust cap (this is why I thought there was some delay)? The pressure signal does not need to be converted? You once also said that Velodyne uses DSP to implement the servo. That conversion would certainly have some latency? I'm not trying to bash accelerometers, I was just curious what the latency is and now I am also curious if the accelerometer is placed on the dust cap.

It is not that absurd. You once told me all closed loop systems were free of thermal compression. How come the Rythmik maintains a more linear response and also is the best performing sub to date (5/19/2012) in the power compression tests. The Velodyne DD-18+ still shows power compression while the Velodyne DD-15+ actually shows power expansion. How come the Velodyne frequency response is not as linear.

Then later in the links I posted for the memory effect section, the Rythmik servo looks almost indistinguishable from the original signal using more complex signals. This form of servo appears to be linear time invariant.

So your main concern is Le(x) and Le(i). This can be reduced with more efficient drivers, properly implemented faraday rings, and less windings. Correct me if wrong but Le(x) and Le(i) distortion is usually low order (if I remember correctly, 2nd order?).

A good test for this would be the Irene scene in Black Hawk Down. The main frequencies are 6, 12, and 18 Hz with very little upper frequency content. The 36 and 54 Hz harmonic should be pretty easy to hear. At the Omaha Sub GTG, the Rythmik was clearly near its limits during the BHD scene. The group consensus is that none of us heard or felt anything from the Rythmik during that scene. If you look at all the summaries describing the performances of the day from people, most thought the Bic F12 sounded louder than the Rythmik. Most of the complaints about the Rythmik involved it sounding too quiet and not standing out.

The link you posted is the same one I did. There is no limiter used on those tests. That is the amp clipping. Ilkka even says that no limiter is used. On some of the current amps there is a limiter that you can turn on/off.

I actually can't remember saying anything to you but maybe that was because it was a while ago and I have forgotten about it.

You need to compare apples with apples. Likewise compare sealed boxes with other sealed boxes and not cheat with a vented box which we know limits cone excursion at the vent frequency, is more efficient for a given sized box but has much poorer transient response and usually requires extra filtering to limit cone excursion below the box frequency which again worsens the transient response. Good for home theater but not great for music ! You keep talking about latency. Why not compare the group delay of the sealed servo box with your filtered vented box and tell me which is the better one ?

No decent servo system mounts the accelerometer on a flimsy dust cap unless it is rigidly coupled to the voice coil. I believe Velodyne rigidly mount their accelerometers at the end of the voice coil. No latency there but how much is acceptable to you anyway and why is it such an issue ?


But mutual coupling can't be eliminated and is a major source of error in the feedback signal specifically because there are non linearities in the motor circuit that cannot be cancelled out !! You keep ignoring the direct error caused by mutual inductance between the drive coil and sense coil for which the proponent of this system conveniently ignores in his blurb. Why do you keep ignoring the obvious flaw in this design ? Hawksford and Mills wrote a paper on this and concluded that the error caused by mutual coupling was about 15 dB down from the velocity signal at 100Hz thus it was not insignificant !! At high driving current the situation becomes worse due to non-linearities in the motor circuit. This is in direct contrast to what the dude says about his technology. Read what he says carefully because you provided the link.

And regarding the link to the test results I posted, how come the compression becomes an issue for amplifier clipping only but a totally different phenomena is the cause for the compression in the Velodyne test ? Surely you must realize that ALL drivers have mechanical limits for which they should not be driven past !! The velodyne happens to provide more technology to deal with this issue whereas the simple opamp circuit in the voice coil servo you site cannot possibly do much !


This is a classic straw man fallacy. What relevance does this test have to do with normal operation ? Similarly moving an accelerometer based servo system around the room and watching the diaphragm respond to the movement is also another strawman because it has absolutely no bearing on normal operation. It is just a silly comparison.


You can't honestly be serious when you quote this. Are there any objective measurements to confirm that this is a major issue or do we just have to believe the fancy animated graphics ?

To bad if the driver doesn't have a vented pole piece

I find it difficult to take a lot of these claims seriously. Most of them would not stand up to peer review by a respected authority such as the AES.

I look for power compression at frequencies greater than >40 Hz. The impedance minimal point is generally in this area. Yes, group delay is derived from the phase. I never said latency was an issue. I don't have the engineering background to know what is acceptable phase shift. I just wanted to know.

Curious question, what subs have you had in the same conditions to determine if sealed is great for music and a ported sub is not great for music? There are great examples and bad examples of both sealed and ported subs.



Brian already addressed the Hawksford and Mills paper in response to you. His Direct Servo is not the same. I don't know the patent of direct servo or the trade secrets but every question you asked in a thread was already answered. You are still probably confused so further questions like this are pointless. It keeps turning into an unsound circular argument.

Yes, if you want to stay within linear range or need more output, add more subs.



It depends if you think the back wave affects the experience. Do you believe that the enclosure is too small for these large waves to be created? How can two subs with a linear response down to the teens sound so different when listening to these with your ears within a couple feet of the driver? Maybe we need more tests to help correlate what we hear. Why does the enclosure vibrate? The driver is free from this force? Can't we build the enclosure out of construction paper with a solid baffle since you think it is absurd the back wave can (re-)radiate through the cone or enclosure?

The chart at the bottom is created by measure, record, change variable, measure, record, calculate difference, plot. Is that not objective? You must of went to a prestigious school if that is not objective. If there is difference in the test, does that not provide some reason even if it is in the early stages? So was that a fallacy fallacy ? Oh, my driver has a vented pole piece.

If you don't have the engineering background then why do you blatantly accept what is written on some website as though it's the truth ?


If it's not a negative feedback servo in the traditional sense then what is it ? In his own words I have seen him refer to it as a variation of the Audio Pro ACE BASS method which was never claimed to be a servo by the original inventor. The whole thing is based on a marketing gimmick and a play on words to make people think it's an improved version of the traditional servo woofer.


Jees, all of this worry about a small hole in a pole piece whilst ignoring the major radiating area of the cone surface !! Hello ? Am I missing something here ? Re-radiation through a small orifice in a pole piece and a meaningless chart to back it up with no practical time domain measurements and comparisons. I think he is having a lend of you.

I bought the kit being somewhat skeptical but bought it based on funds. I had it in >10 rooms including demo room of one audio store and also outside. Other subs seem to subjectively have problems with time-smearing. Actually I was talking about the cone and ignoring the pole vent. Thats why I said, oh it has a pole vent at the end since I almost forgot. Where are all the flaws you mention during third party measurements. He states the draw backs of pure positive current feedback and pure sensing coil based. You actually brought up ACE last time. He explained the differences. Now either there is a communication barrier or someone is trolling. Why should I believe you over him.... I asked my old EE professor questions last year but that could be another fallacy.

We are back at the beginning of the unsound circular argument. Best of luck.

We do not have servos on a drum set, why need them on a subwoofer? The one velodyne sub i listened to with a servo did not turn me on. In fact it turned me off.

Your arguments are unsound because they are based on false premises. You state that you have no engineering background but you keep trying to quantify differences by regurgitating others bogus claims that you are in no position to substantiate. And I love how the 'troll' argument gets wheeled out once again as soon as there is someone that disagrees with the majority fan club

Most of the claims made on that website are just a whole lot of fanciful audio fool speak designed to appeal to the naive. Like I said, most of it wouldn't stand up to peer review by a respected authority such as the AES.

Also did you realise that the inventor of this idea produced further patents which attempted to fix up its major short comings ? Obviously those improvements were hard to implement otherwise he would have done it. It's obviously easier just to promote gimmicks on a website and nitpick at the opposition rather than implement solid engineering solutions to a problem

That's because the distortion a drum produces is part of it's sound. However when reproducing that sound from a speaker in its entirety you don't want a speaker to add anymore distortion to the sound than what is in the original sound. All speakers distort and one way to lower its distortion is to incorporate the speaker in a feedback loop which will reduce distortion.

And by the way most amplifiers use negative feedback for the same reason so nearly every amplifier is essentially a servo system

Yamaha uses servos on their subs and have been doing so for a very long time (20+ years) I think they sound fine and work as intended.

I transferred from the EE department after a year since I did great on the theoretical work but sucked in the lab. What was the point of doing all that work just to save a company some money by simplifying circuits.

Brian already addressed the differences and shortcomings between his servo and ACE and also pure current feedback. The direct servo is different than those approaches. That is great if an engineer can improve on their work. It wasn't frowned upon when I helped with new tests for determining critical power. Also if you can measure a significant difference with trying the best to only change one variable, there should be further investigation.

You told me, "... all other servo designs will have immunity from voice coil resistance changes with temperature...". That did not seem to be the case of those Velodyne measurements. So do I trust the person that apparently has some troubles with multiple reading comprehensions on the differences of implementations, was wrong with some of their past statements, and only states theoretical issues or do I trust the engineer that has put his work up to third party measurements and actually measures in the real world. I really enjoy the ambiance it adds to music. Some people said all subs sound practically the same when flat but I have not found that the case even when I listen nearfield to reduce the rooms effects. No one to date has called my sub boomy and full sounding. The majority of people who heard it think that I have it turned off even though its signal was slightly elevated compared to the main speakers. You are welcome to join us in a blind listening test and state the issues of each sub based on listening.

Even if his servo fit your exact definition (even though it is not pure current feedback or ACE), where in the third party measurements can you prove all these shortcomings? If anything it would be second order distortion. If people are listening without abusing the sub, it should be below the detection threshold for most people. If you further want to decrease non-linear distortion, add more subs.

The drums I have seen also don't have motors. Are you using a variety of sub cones in a drum set? If so, I understand your case . When I have time, I'm hoping to build a new enclosure with no cutout for a plate amp so a group of us can run a non-servo version EQ with JRiver and also quickly compare it with the servo version. I also hope to hear a velodyne side by side with other subs someday.

why wouldn't a servo speaker such as the velodyne have immunity to changes in voice coil resistance and why is this such an issue anyway ?

Since you are very good at telling me what I posted in the past why don't you ask your good mate why he deleted posts of mine and then barred me from posting on his thread (that someone else started anyway) when I highlighted the error in his reasoning which as I recall went something like this :-

"Even a 100% increase in voice coil resistance would only result in a 3dB drop in output which is hardly noticeable.... !!"

Your good mate said it was 6dB which is point blank WRONG !! Then he made some personal snide remark aimed at myself as to why I was wrong when I wasn't !! He should go back to school for making that error !! It even gets worse for his case because if you enclose the speaker in a feedback loop with a relatively high loop gain such as the velodyne then that drop in output will now be totally negligible simply because negative feedback desensitizes the closed loop transfer function to changes in parameter variations so again your claims are highly exaggerated as most of the rythmik claims are !!

And you talk about measurements !! What is so good about those measurements on hometheater shack ? There are plenty of non 'servo' speakers that exceed such a mediocre objective performance.

And you are entitled to your subjective opinions. Nobody is denying you that at least not me. Plenty of people will have good and bad stories about other subs including velodyne

I suggest that you focus on expounding the virtues of your own design rather than nit picking at insignificant limitations of other designs. But when most of your claims are highly exaggerated then what is there to talk about ?