Do servo subwoofers actually lower distortion?

[JohnYang1997]

Usually the combination of feedforward for speed and feedback control for error reduction and parameter drift is the optimal way, I don't know of such an implementation on loudspeakers though, only the known feedback woofers and the feedforward on the new Samsung soundbars.

To my understanding, Samsung must be using DSP in it. And not isolated to the woofer. This way the delay is not a problem anymore it is small enough to just ignore.

 

[thewas]

Yes, although I didn't say anywhere that delay would be a problem in feedforward control?

 

[Hayabusa]

With a well implemented feed forward controller you don't need to have any waiting of feedback as your pre-compensate any deviations so the output signal is already optimal from the start.

In MFB the feedback delay is much lower than the group delay of the woofer.
But I agree that with feed forward a slightly lower group delay could be made.
(But we are in the microseconds range here)

 

[JohnYang1997]

Yes, although I didn't say anywhere that delay would be a problem in feedforward control?

Here is the logic:
I don't see the possibility to precisely implement such controller in analogue.
So one must use DSP.
Since it's digital, there must be some delay.
When use solely for woofer the delay will cause issues.

The solution is to do everything including crossover in digital.

 

[DonH56]

Feedforward has some advantages but there are things it cannot compensate without a closed loop, like component aging or variances over PVT (process variation, voltage, temperature) that are not well-known and predictable in advance. It can also cause a system to go unstable so is not a complete panacea. Typically I have seen FF used with FB compensation to provide the best of both worlds.

 

[snapsc]

You do know this forum is heavy on numbers and not unfounded assumptions, yes?

I don't know the distortion levels but there are plenty of great subwoofers around these days and many ways to achieve low distortion. As for not using servo, it costs money -- you need a sensor (in Rythmik's case a special driver) and amplifier plus R&D to develop the systems. Velodyne still does, or did, use a servo design, and there have been a few others over the years. There are some things a servo can help (e.g. driver control to reduce distortion and "ringing", thermal issues in the voice coil) that are difficult to solve without a feedback loop. But a lot of that will go unnoticed at sub frequencies and of course optimized conventional designs may provide equal or better performance for the dollar.

Always trades...

Absolutely agree that numbers are important, which is why I added the data bass link...as you can see after looking at a lot of subs...it is possible to get low distortion with and without servo. I'm not saying servo is bad....I'm not saying servo doesn't sound good.

So....here are two interesting examples (and don't represent anything other than these two subs)....if you want the lowest bass you can get and don't care about distortion from 40-100 hz...then the Rythmik is the winner (both 1 port and 2 port mode). But if you want the lowest distortion in the 40-100 hz range, get the non servo HSU...but you give up some of the deepest bass capability).

It appears from this preview in Audioholics that what the servo is doing for Rythmik is allowing their subs to go deeper and still stay within the CEA distortion limits. Maybe that implies that servo can be used to get to the lowest distortion for a certain frequency range and maybe it can also be used to get to an acceptable distortion over a wider frequency range (go deeper).

It would be great if Brian Ding could provide more insight.https://data-bass.com/#/?_k=jt1hv7

 

[suttondesign]

What? To make a subwoofer, the box must contain a live monkey! Geez...

That does remind me of an old joke about creatures in boxes:

View attachment 48726

Cute. But as we now know, the cat is dead in one universe but alive in another. Whew.

 

[thewas]

Here is the logic:
I don't see the possibility to precisely implement such controller in analogue.
So one must use DSP.
Since it's digital, there must be some delay.
When use solely for woofer the delay will cause issues.

The solution is to do everything including crossover in digital.

Depends on the complexity, analogue feedforward control exists, but possibly not for such purpose, but the important thing is that a DSP delay by dt isn't a problem in feedforward control as it is known and you just output the control signal by dt earlier.

 

[thewas]

Feedforward has some advantages but there are things it cannot compensate without a closed loop, like component aging or variances over PVT (process variation, voltage, temperature) that are not well-known and predictable in advance. It can also cause a system to go unstable so is not a complete panacea. Typically I have seen FF used with FB compensation to provide the best of both worlds.

Yes, that's what I also said here.
As I said also a sensor can age and change its behaviour which is again a disadvantage in feedback control.
And with a detailed enough model you can even compensate partially some of the effects you mention above without measuring them directly, for example B&O does that already for thermal compression https://www.tonmeister.ca/wordpress/2014/01/16/bo-tech-thermal-compression-compensation/ and also many modern cars calculate the temperatures at some critical engine positions without measuring them.

 

[Rollomoto]

Feedforward has some advantages but there are things it cannot compensate without a closed loop, like component aging or variances over PVT (process variation, voltage, temperature) that are not well-known and predictable in advance.

Just have a look at Klippel controlled sound: https://www.klippel.de/products/klippel-controlled-sound.html
There is some learning mechanismen to compensate aging, etc.

 

[DonH56]

Just have a look at Klippel controlled sound: https://www.klippel.de/products/klippel-controlled-sound.html
There is some learning mechanismen to compensate aging, etc.

Good point. I'm fundamentally an analog guy and did not think about CPU/DSP/etc. control systems. You'd think with the past decade or two of dealing with DFE/FFE/FEC and the like I'd remember...

 

[dped90]

i believe brian is sending one of his subs to amir for testing.

Did it actually happen yet??

 

[fpitas]

I’d prefer a well designed enclosure with a far more robust driver.

I also question how audible the distortion reduction is. Obviously less distortion is good, but...

 

[ErVikingo]

Ah Servo Controlled woofers! I can’t say if mine reduces distribution per se (well apart from keeping my woofers from trying to go under 12 or so hz) but there is a distinct difference in the impulse control when I switch off the servo controller on my Infinity IRS Beta.

There are so many components inside the controller that I don’t know what else is going on.

 

[suttondesign]

Did it actually happen yet??

yes. it got reviewed some time ago.

 

[DonH56]

Rythmik L12 Subwoofer Review

This is a review and detailed measurements of the Rythmik L12 subwoofer. It is kindly purchased by a member and drop shipped to me. The L12 costs US $559 in two colors (black oak and matte) and US $619 (gloss white). If you buy a white one to stick in your home theater, you and I will never...

www.audiosciencereview.com

Review: Rythmik F12G Direct Servo Subwoofer

I decided to split this one off from my other thread here because it is a full-up review of the Rythmik F12G Direct Servo Subwoofer and I thought some of you guys who haven't been interested in my other updates might still care about this review. I wrote this for HTS years ago but the photos...

www.audiosciencereview.com

Rythmik Audio F12G Direct Servo Subwoofer In-Depth Review

Rythmik Audio F12G Direct Servo Subwoofer In-Depth Review

www.erinsaudiocorner.com

 

[dped90]

yes. it got reviewed some time ago.

Yes, I forgot about this one. https://www.audiosciencereview.com/forum/index.php?threads/rythmik-l12-subwoofer-review.12140/

I don't think I finished reading that review, though I wished the sealed F12 sub had been reviewed instead. OTOH, the Rythmik website says that the L12 is their cheapest, so if distortion between 60 and 30Hz @ ~ 100db truly is impressively low then as Jim Salk, who offered better braced cabinets said "what more could you ask for"? https://www.salksound.com/model.php?model=Rythmik 12 Subwoofer

As these are my midwoofers, and which only play down to 70Hz, I'm understandably very concerned about distortion within that lower bandwidth, though only for music, not HT.

 

[Compact_D]

As far as I know, none of the big full range speakers have a servo as part of their designs

There was such an attempt:

Radiotehnika S-70

It was manufactured in Latvia (then USSR) in the early 80's and had a Piezo feedback sensor (right bottom in the diagram). As at that time "profit" was not such an important thing, there was some room for experiments.

There were measurements of course done during development which are now probably lost, but subjectively bass sounded way better compared to the passive variant with bass reflex instead of servo system. The drawback was that at very high levels it was possible to overload the servo system with some unpleasant results.

and a fun video (not technical) :

 

[Keith_W]

There are some nice insights from Brian Ding about his servo subwoofer design in this AudioCircle thread.

 

[dped90]

I was curious if anybody has measurements that show that a servo control loop on a subwoofer actually lowers distortion at frequencies in the 20Hz-50Hz region. I'm sure it would need to be the same test setup where the loop is turned off and on to show the difference. There is a lot of marketing push surrounding this technology and I thought it would be insightful to see some actual measurements.

Would be very interested in your evaluation of this review on the Rythmik F12G sealed subwoofer.

Rythmik Audio F12G Direct Servo Subwoofer In-Depth Review

Rythmik Audio F12G Direct Servo Subwoofer In-Depth Review

www.erinsaudiocorner.com