Ask an Acoustic Engineer (me)

[Ge0]

Great response to my question Justin and thank you for taking the time to write that. I loved the driving analogy! I’m sure this will help many members here understand the theory. I couldn’t get the DIYMA link to work. I did a bunch of research into regulated vs non regulated amplifiers a long time ago. Not sure if it’s entirely accurate but the consensus was that non regulated power supplies gave amplifiers a slight edge in dynamic power over regulated power supplies but I’m not so sure that would be audible. Just think of the power drop you get when shutting off your engine with the audio playing. Your voltage could go from 14 volts to 12.5 volts yet you likely won’t notice a difference from the reduced power as a result.

I hate to further complicate things, but . Anyone remember the old Bob Carver amps that put out hellacious power in a tiny rack mount chassis? This was before the class D amplifier days. How did he do this? Regulated variable tracking power supplies. The power supply adjusts its rail voltage rapidly in sync with the incoming signal. This kept power supply switching losses to a minimum.

Here is some info off the internetwebthing"

1) Magnetic Field Power Amp = a fancy (maybe even copyrited) name just to stand out among other amps and dazzle (baffle) the uninformed consumer.

2) The only thing related to a magetic field is the use of a TRIAC plus additional circuitry to modulate the primary of the power transformer to limit power/secondary voltage. It doesn't appear to be very well regulated and only seems to "stiffen" the AC line under heavy loads. Not real fancy, not real effective, and can be very noisy due to the low frequency AC lines and spikes caused by high voltage switching.

3) All Carvers I've seen run 'Class G' (rail switching) circuitry in their output stage. They run low/med/high Vcc rails with the 'Class A' drivers running on the highest rails and the additional output devices connected to the higher rails by fast recovery diodes. The entire output stage is controlled by a single op-amp through global feedback. This gives the benifit of lower idle power thus the need for smaller heatsinks, along with high peak-power. The best example of this is their famous little cube amp rated at 200WPC with virtually no heatsink.

Today's pint sized amps use this same technology. Mix that with a multi-stage class D output stage and your overall losses are reduced to a bare minimum. Really cool stuff. It can't quite beat classic amp topologies in sound quality quite yet. But, you would be hard pressed to detect a 0.01% distortion vs 0.1%. The reduction in size is worth the trade-off to me in my daily driver.

Ge0

[Ge0]

So.... Just so I understand at a more "basic" level. :-)

- An amplifier with a "regulated" power supply will provide the same amount of power regardless of the speaker impedance.

- An amplifier with an "un-regulated" power supply will provide less power to speakers with higher impedance.

Are those two statements true?

The reason that I ask is because I'm still not clear why my JL XD600/6v2 (which does NOT have a RIPPS power supply) provides 100W RMS at 2 ohms, but 75W RMS at 4 ohms. Why is there such a small difference between the 2 and 4 ohm outputs? Don't most "un-regulated" amps provide twice the amount of power at 2 ohms (compared to 4 ohms)?

While the JL XD line of amps don't have the "RIPPS" power supplies, do they maybe have some "lesser" version of a regulated power supply instead? I honestly don't know if they are regulated or un-regulated.

Thanks - and sorry if this was already explained earlier - maybe I just didn't understand it.

A regulated power supply will hold a constant voltage at its power supply rails regardless of small (or large) fluctuations in system voltage. For instance it will keep the rails at +/- 20V whether Vbat is at 9V or 15V. The power you supply to a fixed load will not vary.

An unregulated power supply can be considered an input voltage tracking supply. Say it is designed to produce +/- 20V at 13.5V. At 9v you may only get +/- 15V power supply rail voltage. At 15V maybe it's +/- 28V. The power you supply to a fixed load will vary in regards to vehicle system voltage.

This does not have much to do with amp protection circuitry. Each amp design has a maximum power envelop. This is dictated by power supply design and heat sink cooling capacity. Similar technology to what is used on modern video cards and microprocessors. The amp will protect itself when it gets near its thermal limits. Also, driving lower impedance loads raises current draw and thermal loss. So, doubling of power may not be possible by halving the load. The amp just couldn't deliver that much power safely so it throttles back. Hope this makes sense.

Ge0

[JCsAudio]

What you’re describing to me Ge0 sounds like a class H amplifier design which is a Class AB with variable power supply from my understanding. Arc Audio currently uses a variation of this technology in some of their amplifiers.

[Cutaway]

@justin zazzi
I read a great discussion a few posts back and wanted to revise the topic of Loud Speaker Design and innovation.
Just curious on your thoughts/opinions on why we haven't seen anything like the phoenix gold cyclone or CV Strokers in today's car audio? I don't know what the PG performance was but CV did really well with that design.

[Bnlcmbcar]

From the diagrams that you posted, I'd be concerned as you are (especially in a car) that you'd narrow the image, because you are pulling some of the L and R content into the center - which, most troublingly - brings some of the L content to your R side, if you are in the driver's seat. Conversely, for a passenger, brings some of the R content to the L of them. Eek.

Yes indeed. This is the case for scenario 1 in the diagram. I ‘illustrated’ what you described above in the first attachment where the stage collapses. But if all 3 speakers are the same can we get the stereo sum to sound wider with speaker placement by playing stereo differential signal from the R and L speakers (scheme 2)?

What I'd like to see in a center channel for a car, would be "(L+R)-(L-R)". That way, it's only bringing the content from both speakers to the center, and even if there's a little content that's on both channels but simply louder on one than the other, it still mitigates that by attenuating that sound at the center channel. I'd believe that would at least help keep the stage width - hopefully as wide as with just a stereo pair.

I know what you mean but it cannot be derived that way. There is complex math that needs to be crunched to calculate a center signal.

"(L+R)-(L-R)" = L+R-L+R = 2R (not desired center)

Even if you take it further to using L-R and R-L to extract center from combined L+R you wind up with L+R again:

(L+R) - [(L-R)+(R-L)] = L+R - L + R - R + L = L+R

Fear not though, you can actually have that center channel you describe with a select few DSP’s that have center channel processing based on calculations to deduce the common information from both the right and left channels and steer them to a center channel based on their proprietary algorithms.

But I think really, in a car, the KISS rule applies... there's already glass, and absorbant upholstery, and plastic, all pointing different directions - you could have ONE speaker in a car and end up with a nightmare of multiple pathlength distances, direct and reflected (with each of those having a 180 degree shift - plus pathlength difference offset!) creating anything but a flat response as it arrives at your ears.

So I subscribe to the "the fewer speakers the better" theory for car audio. There's exceptions - for example, three way components where you actually aim the mid and tweeter - can provide better imaging. But my default recommendation is simplicity over complexity, for those pathlength reasons.

I’m just very curious for information on the theories of whats achievable. The main premise and goal would ultimately be geared towards creating a stage for more than 1 passenger to enjoy.

I'm really interested in this myself, but in no way for creating a center channel (IMO, there's already enough direct and reflected pathlength sounds wreaking havoc on image-killing phase interactions as they all ultimately arrive at the listening position)...
I want to make a "L-(L+R)" channel and a "R-(L+R)" channel to add some rear fill (something I otherwise also don't believe in, for those same image-killing phase interaction reasons) plus some additional delay, so I'm going to be researching this soon myself. My DSP will only help with delay and passband. I'd be interested if you find any good threads.

Check out the miniDSP 8x12DL and the APL1012. These have some powerful features involving FIR filters to help account for the harsher environment of the car audio with its many reflections. The Helix Ultra DSP has a center channel algorithm in conjunction with a virtual channel architecture that enables varying degrees schemes like 3 and 4.

Eventually stronger processors/algorithms will emerge (or become cheaper) enabling more options to the table for car audio than whats more traditionally used.

[geolemon]

@justin zazzi
I read a great discussion a few posts back and wanted to revise the topic of Loud Speaker Design and innovation.
Just curious on your thoughts/opinions on why we haven't seen anything like the phoenix gold cyclone or CV Strokers in today's car audio? I don't know what the PG performance was but CV did really well with that design.

Sorry to hijack - I happen to own two PG Cyclones (one broken), and had the honor of making pen-pals with Tom Danley back in the early '00s, with the goal of making my own rotary servo subwoofer (NOT to make a product to sell, just a fun one-off, for engineering bragging rights). Unfortunately my business partner (fellow audio tech geek) and I burned out the motors that Tom Danley personally had suggested to us, and we put it on the shelf to focus on other things... that would have been fun though. Basically I'd taken a flat bar stock for the center, and had my buddy CNC machine some motor mount adapters - I mounted one motor on each end, unlike the Cyclone with had one motor at one end. The motors were each about double the size of the Cyclone's, so I made a center vane out of foam core and kevlar, hollowed out with stiffening channels, and then vacuum-bagged it for a super light vane. Instead of a screw shaped wave guide, I was just going to build it into a quad-chamber box that would manage the wave phases. Would be fun to pull tat project off the shelf, but it was becoming an expensive experiment...

Anyway, I'm not seeing the similarity between the PG Cyclone and the CV Stroker:

The Cyclone was a remarkable one-of-a-kind joint effort between PG Cyclone, specifically licensing Tom Danley's patented servo-drive technology - interesting for that rotary implementation that ServoDrive doesn't even sell. It was a one of a kind.
I can tell you from my experience with the Cyclone, that it is good for about 40hz and down (way down), but basically then you need another more traditional subwoofer to pick up the 40hz and up to wherever your midbasses really pick up. It was also very expensive, to not do everything that a subwoofer of today can do. And the real advantage was displacement - this was before "high excursion" subs really existed. I'd have to find the specs again, but this was before subs like the W7 or XBL^2 existed, so they were advertising the Cyclone had 3x the displacement of a standard 12"... which is about what a W7 does, or a CSX-12, and even a W7 is less expensive.

The Cerwin Vega Stroker is basically just a traditional subwoofer. It had a unique little gimmick, but it isn't really wild or radical IMO: in place of a dustcap, they used a second spider, that clamped onto a rod that poked forward of the pole piece. I say "clamped" because there was a set screw that allowed you to slightly adjust the center "at rest" position of the cone unit in the magnetic gap - some say that was to compensate when you mounted the subwoofer horizontally, to actively avoid cone droop... but that's really solving something that isn't usually a problem anyway. It did make the suspension quite unique, because really there was both an upper and lower spider, and even the surround was spider-like. But essentially, it was a big, heavy SPL sub with a fancy dustcap. Definitely not as far-out as a Cyclone.


Do you mean "Why aren't there more unique subwoofers"?

There are, if you look past the kids who only care about "power handling" (despite the trade-offs in efficiency that imparts) from their salad-bar contract manufactured subwoofers, there's lots of unique geek stuff - that's the stuff I love.
I'd suggest reading up on:

• underhung vs overhung subwoofers (just as a baseline on motors - but underhung are far less common)
  
  • optional components like shorting rings and sleeves (seeing DUMAX measurements of BL curves will help understand that topic - as well as the ones below: )
  
  
• the JBL GTi DDD motor - which had two separate magnetic gaps, and two opposite-wound windings on the former, to not only give it linearity, but a braking force if driven to over-excursion.
• the XBL^2 motors - essentially a groove in an underhung-sized top plate, with a voice coil winding length perfectly balanced so that as coils leave magnetic field 1, they enter magnetic field 2, keeping BL linear - but in a MUCH simpler way than JBL did
  
  • Side note for googling: check out where the W7 is cross-drilled in it's long excursion motor, and it's winding length. Some say that's to create an XBL^2-like effect, without running afoul of the XBL^2 patents.
  
  
• Speaking of the W7 - it's own front suspension is very innovative, and allows for a larger cone - therefore more cone area, and a larger surround - therefore more excursion capability, previously somewhat mutually exclusive.
• Shallow subwoofers - some of these stuff the motor inside where the cone used to be - well done ones are interesting engineering, and worth looking into the trade-offs on excursion and how they overcome them.

If it's not just unique subwoofers you are interested in, search out:

• plasma tweeters. In pursuit of the lowest mass possible, it's literally sound from an electrical flame.
• Electrostatic drivers, and magnetic-planar drivers. It's a panel suspended in either a massively high-voltage static field, or a magnetic field, rather than a cone.
• Distributed Mode Loudspeakers. It's literally just a motor, designed to vibrate something that otherwise wouldn't be a speaker
• Even - phase plugs and whizzer cones on midranges. That search will get into a great discussion on speaker directionality.
  
  • which also makes me think of compression drivers - in the late 80's and early 90's they were morphed into "waveguides" that dominated IASCA SQ competitions back then

I'm sure I could come up with others (I own a dome subwoofer with a 10" voice coil but that was never on anyone's radar), but if you are researching to understand them personally, and then to understand why these aren't all commercial successes, that list above will keep you busy for a couple years.
Usually though, commercial failure comes down to two things:

• Relatively unproven technology is uncomfortable to most people. That's a barrier to sales and to adoption rates.
• Low volume sellers, and complex assemblies, both add to cost. Expense is always a barrier. Even me - I'm personally a "bang for the buck" fan way more than I'm a "geek engineering" fan.

...and that always should make an engineer ask themselves:

1. What is the goal we're trying to achieve? Speakers simulate air - is there something different you are trying to achieve? Why?
2. Is there an existing way to achieve that goal, that reaches the limits we want, for less money? Then why engineer a more expensive way?

Sorry - /hijack.

(EDIT for visuals: )
JBL GTi - really two opposing motors folded into one - you can see both coils, both magnetic gaps - which are opposite wound, because those gaps are opposite phase:

XBL^2 motor cutaway - showing "at rest" position, but as the cone moves in either direction, coils leave one gap and enter the other, leaving BL constant (linearity) while making huge excursion:

Compare that to those baseline "overhung" and "underhung" cutaways.
And the best I can find of the W7 internals... see that sneaky pole vent cross drilling in there? It's right at the center of where the coil is "at rest", having the same type of effect as XBL^2 [as that would weaken the field at that spot]... but also that innovative basket, suspension, all custom parts that the little guys just can't do:


EDIT edit: (apologies)
I can't mention ServoDrive and how different the Cyclone is, without showing what they actually sell:

It's interestingly a rotary servo motor at the top (blue, mostly hidden - looks like a shop tool motor) with a shaft that goes to that proprietary belt mechanism, which converts the rotary motion to linear motion - pushing two shafts that move two "traditional" 15' cone units in a linear way.

[Cutaway]

@Geolemon -
That is some great information and will keep me busy in the "rabbit hole" for a long time. Even though i am familiar with a few of your examples (the popular ones) i guess i hadnt really considered those innovative and in reality they really are. I have tended to shy away from brands like JL, JBL and RF for really dumb reasons.

Your knowledge share is really appreciated.

[Jdunk54nl]

Question that someone might be able to answer on the constant power amps,
On a 4 channel amp, if channels 3 and 4 are 2ohms and 1 and 2 are 4 ohms, when the amp detects 2 ohms on channels 3 and 4, does it then only adjust those channels for 2 ohms or does it also adjust channels 1 and 2?

Just curious about that. I am assuming it only limits channels 3 and 4.

[Ge0]

What you’re describing to me Ge0 sounds like a class H amplifier design which is a Class AB with variable power supply from my understanding. Arc Audio currently uses a variation of this technology in some of their amplifiers.

It's not class H, it's Magnetic Field baby!!!

Seriously though, amp classifications are so misused by marketing departments these days I just ignore them. What's class GD or class HD? Then there was the Tri-path class T. Ughh...

Ge0

[JCsAudio]

It's not class H, it's Magnetic Field baby!!!

Seriously though, amp classifications are so misused by marketing departments these days I just ignore them. What's class GD or class HD? Then there was the Tri-path class T. Ughh...

Ge0

Tell me about it. In the end though I think most of these fall into a few categories and the manufacturers are just using marketing tactics to make it seem like their product is different or better. We’ve got class A, class B, Class AB, Class D (T is basically chip based D) and Class G and H and a combination of those two.