What's It Take to Get 16 Hz @ 116 dBA in a 13" Cube?

[toddalin]

I pulled the plate amp out of my Sunfire True Subwoofer Signature for some repair work. Seems that powerful subs have a habit of shaking things up pretty good inside which can be very detrimental to their electronic health.

The sub is rated to deliver greater than 116 dB peak SPL (including room gain) from 16 Hz to 100 Hz (-6 dB) at 1 meter with a THD of 1/10 of the fundamental between 18 Hz and 80 Hz. Sound & Vision Magazine tests of old found these claims to be surpassed.

Anyway, for those pressure levels you obviously need lots of power. The Sunfire plate amp is rated at 2,700 WRMS at 3.3 ohms, the dc resistance of the woofer.
http://www.largescaleonline.com/eimages/lsolpics/Team_Member_Pics/toddalin/Sub/Faceplate.jpg

So that obviously requires a massive power supply, or so you would think. But Bob has found a way around all that. The whole plate weighs less than 10 pounds and uses a small transformer and just a couple high voltage caps. Incidently, even after the amp had been sitting, I measured 146 volts on the power supply capacitor, so discharged it through a big resistor before handling and packing. :biting:
http://www.largescaleonline.com/eimages/lsolpics/Team_Member_Pics/toddalin/Sub/Guts1.jpg

But that kind of power does take some beefy semi-conductors to handle the switching duties.
http://www.largescaleonline.com/eimages/lsolpics/Team_Member_Pics/toddalin/Sub/Guts2.jpg

And you got to move a lot of air. The 13" cube uses a 12" long-throw woofer.
http://www.largescaleonline.com/eimages/lsolpics/Team_Member_Pics/toddalin/Sub/Woofer.jpg

With a big azz vented magnet structure.
http://www.largescaleonline.com/eimages/lsolpics/Team_Member_Pics/toddalin/Sub/Magnet.jpg

A passive radiator is also included.
http://www.largescaleonline.com/eimages/lsolpics/Team_Member_Pics/toddalin/Sub/Passive.jpg

[Fred Sanford]

We had them come in for repair ALL the time, far more often than the other brands we carried. I think I only ended up installing one on all of my jobs ('cause I'd seen so many come in from service calls), and it failed twice in the first few years. They work really well, when they work.

YMMV, I hope.

je

[Eaulive]

2700W RMS.

Ok, I see....

[toddalin]

We had them come in for repair ALL the time, far more often than the other brands we carried. I think I only ended up installing one on all of my jobs ('cause I'd seen so many come in from service calls), and it failed twice in the first few years. They work really well, when they work.

YMMV, I hope.

je


Thanks.

I am seriously considering just putting a "block-off plate" on the speaker cabinet and making a separate cabinet for the plate amp so that it can be placed remotely and not subject to the vibration and internal pressures generated by the speaker/passive radiator.

[JBL 4645]

Don’t mean dbC weighting. If it was dbA weighting you’d need a lot.


2700W RMS.

Ok, I see....
Are you been sarcastic? I know you’re thinking Volts!

What is the sensitivity of the sub bass driver?

[Allanvh5150]

2700 of what type of watts? Milli?

[Eaulive]

2700 of what type of watts? Milli?

:rotfl::rotfl::rotfl:

[robertbartsch]

Is this a Bob Carver design? I think he also designed the Flame Linear amps of yester-year.

I'm using a plate amp for one sub and, so far, it has held up well for more than one year.

[spkrman57]

Those outputs look like they would be hard pressed to put out 100 watts.

Must be a switching design using smoke and mirrors!

Has anyone actually measured the output from that plate amp to see if it even gets close to 1,300 watts.

Sorry to be a sceptic, but I don't see enough heavy duty parts to get that wattage.

Just my 2 cents worth of course.

Ron sends...

[Eaulive]

Those outputs look like they would be hard pressed to put out 100 watts.

Must be a switching design using smoke and mirrors!

Has anyone actually measured the output from that plate amp to see if it even gets close to 1,300 watts.

Sorry to be a sceptic, but I don't see enough heavy duty parts to get that wattage.

Just my 2 cents worth of course.

Ron sends...

If the line voltage is directly converted to a 160 VDC supply, the theorical power output at 3.3 ohms could well over 3000W, then again a 120V 30A circuit should be used and the power cord should be something else than the little 16 guage toaster cord.

Second I don't see enough capacitors to hold that DC voltage when peak currents over 40 amps are sollicited.

The output devices in class D don't dissipate much heat, it could very well be a floating bridge class D configuration but I'm still skeptical about the 2700Wavg (no, I will not say RMS)

I doubt any driver of that size could handle that much power and heat dissipation anyways.

[Titanium Dome]

In the long run, the convenience of a self-powered sub is a diminishing benefit. More and more I'm a believer in external amplification.

[toddalin]

In the long run, the convenience of a self-powered sub is a diminishing benefit. More and more I'm a believer in external amplification.


In speaking with Rita of Rita's Vintage Audio, and a former Carver employee who did the repairs, she noted that the problem with these is that they literally shake themselves to death.

To that end, I've made a 3/4" block-off plate for the cabinet and a separate enclosure with lots of ventilation for the plate amp. This should also let the amp run a little cooler.

[jcrobso]

The concept was to have a very smart power supply that could supply the current needed with out big transformers and caps.
I remember Bob's cube amp, 200WPC RMS (OK, I said RMS) from a 6" cube.
Looking the size of the plate it would be possible to get a lot wattage using a mag-field type of amp.

[Eaulive]

00WPC RMS (OK, I said RMS)

Shame on you :D

[Eaulive]

The concept was to have a very smart power supply that could supply the current needed with out big transformers and caps.

If you supply DC directly from the 60Hz AC line and you need current, you WILL nedd caps, a lot of them.
In simple terms, capacitors are needed to provide energy during the dips of the rectified AC waveform (120 dips per second), the more current you draw, the bigger reserve capacity you need.

In some cheapie power supplies the capacitors are cut short, so the voltage supply drops drastically during peak demands, so does the output power.

[jcrobso]

http://en.wikipedia.org/wiki/Bob_Carver

[Robh3606]

Second I don't see enough capacitors to hold that DC voltage when peak currents over 40 amps are sollicited.


The power supply is transformer based. Magnetic Field Amplifier

Read This: http://www.google.com/patents?id=5K4uAAAAEBAJ&printsec=abstract&zoom=4&source=gbs_overview_r&cad=0#v=onepage&q=&f=false

Rob:)

[jcrobso]

Take a look at the power supply in your PC, if you look a the current ratings and the size and weight of the supply you will no way can this thing do this.
But it does, works very well.;)

[Eaulive]

The power supply is transformer based. Magnetic Field Amplifier

Read This: http://www.google.com/patents?id=5K4uAAAAEBAJ&printsec=abstract&zoom=4&source=gbs_overview_r&cad=0#v=onepage&q=&f=false

Rob:)

Ok, so it's a two-tier output stage with a switching power supply. What I say still holds, if you want to get a stable DC source from a 60Hz AC line, the amount of capacitors is determined by the current draw. No way around it.

A computer power supply is only ±400W (3.5A RMS) and it has usually around 300 to 800µF of capacity at the rectifier.

So in order to produce 2700W, it has to draw 40A peak on the supply line, no way the capacitor in the picture can perform that task.
Hence there is some voltage sag that Bob is willing to accept :D

To confirm I just checked the schematic of the switching power supply of a PRO Crown amplifier rated 750W per channel @ 2ohms (1500W total)... it has 6 big 1800µF 200V capacitors at the rectifier, 10800µF total.
Each of these caps is about twice the size of the Carver.

[Allanvh5150]

Ok, so it's a two-tier output stage with a switching power supply. What I say still holds, if you want to get a stable DC source from a 60Hz AC line, the amount of capacitors is determined by the current draw. No way around it.

Not quite. Most of the Carver amplifiers I have worked on had 3 supply rails. If we look at the schematic for the Carver PM1.5 we find the low voltage rail of 36.6 volts wich is operating on one pair (per side) of MJ15024 transistors. As the power increases, the second rail kicks in at 76.6 volts coupled to another MJ15024. Finally at maximum power, the third rail kicks in at 124 volts with anothe pair of MJ15024. The power supply had 8800uf on the 36.6 v rail, 4400uf on the 76.6v rail and 9600uf on the 124v rail. The system works pretty well under the scope these kind of amplifiers were designed for. But in my experience, the powere delivered from these amplifiers is peak power. The amp delivers maximum power on the peaks but if they are fully loaded to full power they do not deliver their rated power. To replace a 500 watt power amplifier with a "switchmode" amplifier, it is normal practice to use double the power. This becomes very pronounced on signals with a very high average RMS level, like radio or rock bands using compression. Make sense?

Allan.

[Robh3606]

Ok, so it's a two-tier output stage with a switching power supply. What I say still holds, if you want to get a stable DC source from a 60Hz AC line, the amount of capacitors is determined by the current draw. No way around it.


The schematic clearly shows 3 taps at 25V, 50V and 75V. There is also a complete schematic of the power supply and one amplifier channel with values. There is about 10000uF total on all three taps on each rail.

Rob:)

[Eaulive]

The schematic clearly shows 3 taps at 25V, 50V and 75V. There is also a complete schematic of the power supply and one amplifier channel with values. There is about 10000uF total on all three taps on each rail.

Rob:)

I didn't go past the first schematic that show a two tier, and maybe the rest of the schematics make sense but it's not the amp we see in the first picture.

I used to work with Carver PRO amps in the 1980s, with Bose 301 and 801 (flameproof suit secure :p) They were amazingly powerfull for the size and weight, no doubt.

I'm discussing the amp that is in the first post here, nothing more! I doubt very much it can deliver 2700W avg.

[Eaulive]

The schematic clearly shows 3 taps at 25V, 50V and 75V. There is also a complete schematic of the power supply and one amplifier channel with values. There is about 10000uF total on all three taps on each rail.

Rob:)

What's the rated avg power output of this amp?

[jcrobso]

There are several versions. Years ago IBM stared switching to one kind of supply. The concept was to have a 60hz main supply and high current power amp that was driven by a 400hz oscillator. The output of the power amp was fed to full wave rectifier the output of the rectifier was DC at 800ppm. At 800ppm the amount of capacitance needed was relatively small compared to a 60Hz driven supply. There was a feed back circuit the monitored the voltage and if it started to sag it would increase the gain of the amp to compensate.
PC Power supplies are multi-voltage. I looked at one of mine.
400W, +12v@15a,-5v@5a,+5v@5a,=3.3v @30a and +5v@40a.
All of that from a small package via switching power supplies. If PC used a conventional PS it would fill the case.

[Allanvh5150]

Chances are this amp is class G or H, commutating power supply not switching.

Allan.

[jcrobso]

Chances are this amp is class G or H, commutating power supply not switching.

Allan.

But I think that the voltage rails are supplied by a variation of a switching supply.

[Eaulive]

There are several versions. Years ago IBM stared switching to one kind of supply. The concept was to have a 60hz main supply and high current power amp that was driven by a 400hz oscillator. The output of the power amp was fed to full wave rectifier the output of the rectifier was DC at 800ppm. At 800ppm the amount of capacitance needed was relatively small compared to a 60Hz driven supply. There was a feed back circuit the monitored the voltage and if it started to sag it would increase the gain of the amp to compensate.

I know this design, it is notoriously noisy.
No matter wich way you put it, if you switch a 60Hz sine wave at high frequency to be able to transform it with a smaller transformer, fine, but the resulting waveform still have the dips and peaks of the 60Hz base signal.

All this lack of capacitors is compensated by the ability of the PSU to increase its output during the dips and the small capacitor makes the rest, it makes for a smaller supply obviously.

I don't know how the amplifier in the first post works, but I will be very surprised to see this thing outputting 93VRMS in a 3.3ohms load.

I would accept to be proved wrong however ;)

[Eaulive]

But I think that the voltage rails are supplied by a variation of a switching supply.

It has to be, it's the only explanation for the small capacitor bank.

[Allanvh5150]

It doesnt have to be. With "class H" the power supply tracks the audio signal and usually stays a few volts above what is required for the power output at any given time. Carver calls it "tracking downcoverter" but this is just a fancy name. Because of this the power supply requirements are about 1/10th of what a normal amplifier would be. Some of these amplifiers do not even have heatsinks as the output stage is so efficient. With a highly efficient output stage and the fact that heat is down to a minimum, the power supply does not have to provide power to heat up the heatsinks. Also, this type of amplifier banks on the fact that most people do not run their amplifiers above 40% so the power supply can be smaller still. If you run a class G or H amp at full power for long periods the power supply will need to be large.

Allan.

[Eaulive]

It doesnt have to be. With "class H" the power supply tracks the audio signal and usually stays a few volts above what is required for the power output at any given time. Carver calls it "tracking downcoverter" but this is just a fancy name. Because of this the power supply requirements are about 1/10th of what a normal amplifier would be. Some of these amplifiers do not even have heatsinks as the output stage is so efficient. With a highly efficient output stage and the fact that heat is down to a minimum, the power supply does not have to provide power to heat up the heatsinks. Also, this type of amplifier banks on the fact that most people do not run their amplifiers above 40% so the power supply can be smaller still. If you run a class G or H amp at full power for long periods the power supply will need to be large.

Allan.

Agreed, class D are even more efficient , close to 97, 98%. So assuming a perfect amplifier 100% efficient and a perfect power supply 100% efficient where all the energy is properly used and no heat is generated, 2700W at the speaker output implies 2700W of AC consumption, meaning close to 23A at 120V.

I don't see the tiny circuitry/power cord/fuse to be able to deal with this kind of current, again assuming a total 100% efficiency.

I just went to the sunfire website... power output: 2700W, RMS of course:rotfl: and a power consumption of 600W typical.

I rest my case.


P.S. I know that most amplifiers rated xyz Wavg always indicate a lower power consumption in W, simply because the audio program is not a pure sine wave at rated power, but they will never claim a supposed output power that is almost 5 times the power consumption. Usually it's about 50 to 60% more, not 450% :blink:

[Allanvh5150]

Nope. Lets look at it another way. I will talk about 1000 watts. Say the amplifier is class H and can deliver 1000 watts on peaks, for example on kick drum, the pulse only lasts for 50 - 100 milliseconds. The power supply only has to provide power for this very short amount of time. Many class H amplifiers are spec'd at 10% THD, this doesnt really matter for sub bass, but will allow the power output to be about 30% higher than the power before clip. This brings our true output power down to 750 watts. Also, as I have said, most people only run their amps to about 40% and below so the power supply then only has to supply 300 watts. There is also very little heat generated in the output stage so the power supply does not have to waste time providing that either.
Generally, the power supply for class H is about 10% of the capacity of a true constant power RMS amplifier. A normal class A/B amplifier of 1000 watts may consume 1600 watts or more. A class G amplifier will only consume 100 watts or so. You need to remember that if you try to get the rated power from a class H amplifier for more than a few points of a second, the limiting circuitry in the amp will drastically lower the out put power.
It is quite unfair that class G and H amplifiers can specify their output power as RMS because it is peak power in reality. A normal way to test the output power of a class H amplifier is to put a dead short on the output and send a 50ms pulse through the amp and measure the power. Totaly bogus if you ask me but it give the manufacturers the numbers they want although they will never tell you how the power is measured. Real sledgehammer amplifiers will run all day with a continuous sine wave at full power and not bat and eyelid. The 1000 watts may be up to 5000 watts when tested the same way. The true RMS power of our 1000 watt class H amp may only be 150 watts or so but it is how the amplifier can delliver peak power that counts. Or so it seems.

Allan.

[toddalin]

User Manual text states:

Notes:
Amplifier Output:
2,700 W rms @ 3.3 ohm impedience

Line Power Consumption:
800 Watts average 16Hz - 100 Hz at absolute continuous maximum output
8 Watts at idle
2,700 Watt peak time limited basis.

THD:
Less than 1/10 of the fundamental between 18 and 80 Hz (This includes the amp and speaker.)

See the link to the User's Manual for more info.
http://sunfire.com/manuals/True%20Su...e%20Manual.pdf (http://sunfire.com/manuals/True%20Subwoofer%20Signature%20Manual.pdf)

[toddalin]

Also may be of interest in the User's Manual:

"The power amplifier within your Sunfire Subwoofer is capable of delivering over 2,700 watts into a 3.3 ohm resistor (the voice coil resistance of the subwoofer's driver). When the same full output is applied to the driver, however, the enormous back-electromotive force generated as a consequence of its large motion and giant magnet causes the current flow to be much less than if it were in a 3.3 ohm resistor. It is this signular property of the driver that allows the woofer to be approximately ten times more efficient than a woofer this size would normally be. Hence, the actual average input power is substantially less than 2,700 watts, and is approximately 380 watts for most musical material on the loudest passages.

A compressor circuit kicks in automatically if the input signal reaches a level that would overload the driver. This maintains a ceiling on the output without clipping.

If the input signal is driven even further, a 'soft clipping' circuit is enaled. This allows the woofer to put more sound into the room to satiate the power hungry user, but without distortion or damage to the woofer. Thus, for explosive scenes in movies, this produces extremely high sound pressure levels in your room without the woofer banging against its mechanical stops."

Regardless of the amplifier wattage, when Sound and Vision tested the unit, they found it to be the flattest sub they ever tested and easily out performed its 16 Hz, 116 dB claim.

[Allanvh5150]

User Manual text states:

Notes:
Amplifier Output:
2,700 W rms @ 3.3 ohm impedience

Line Power Consumption:
800 Watts average 16Hz - 100 Hz at absolute continuous maximum output
8 Watts at idle
2,700 Watt peak time limited basis.

THD:
Less than 1/10 of the fundamental between 18 and 80 Hz (This includes the amp and speaker.)

See the link to the User's Manual for more info.
http://sunfire.com/manuals/True%20Su...e%20Manual.pdf (http://sunfire.com/manuals/True%20Subwoofer%20Signature%20Manual.pdf)

I was only guessing the values but it seems I was pretty close to bang on!

Allan.

[Eaulive]

Nope. Lets look at it another way. I will talk about 1000 watts. Say the amplifier is class H and can deliver 1000 watts on peaks, for example on kick drum, the pulse only lasts for 50 - 100 milliseconds. The power supply only has to provide power for this very short amount of time. Many class H amplifiers are spec'd at 10% THD, this doesnt really matter for sub bass, but will allow the power output to be about 30% higher than the power before clip. This brings our true output power down to 750 watts. Also, as I have said, most people only run their amps to about 40% and below so the power supply then only has to supply 300 watts. There is also very little heat generated in the output stage so the power supply does not have to waste time providing that either.
Generally, the power supply for class H is about 10% of the capacity of a true constant power RMS amplifier. A normal class A/B amplifier of 1000 watts may consume 1600 watts or more. A class G amplifier will only consume 100 watts or so. You need to remember that if you try to get the rated power from a class H amplifier for more than a few points of a second, the limiting circuitry in the amp will drastically lower the out put power.
It is quite unfair that class G and H amplifiers can specify their output power as RMS because it is peak power in reality. A normal way to test the output power of a class H amplifier is to put a dead short on the output and send a 50ms pulse through the amp and measure the power. Totaly bogus if you ask me but it give the manufacturers the numbers they want although they will never tell you how the power is measured. Real sledgehammer amplifiers will run all day with a continuous sine wave at full power and not bat and eyelid. The 1000 watts may be up to 5000 watts when tested the same way. The true RMS power of our 1000 watt class H amp may only be 150 watts or so but it is how the amplifier can delliver peak power that counts. Or so it seems.

Allan.

Ok, I agree completely with what you say. It's just very misleading to spec an amplifier based on the peak instantaneous burst power.
If I take a common SR 500wpc @ 4Ω amplifier anI measure it's instantaneous burst current at, say 1Ω or less, I will get a power figure in the high 4 digits also.

Appeal to the masses. :barf:

[Eaulive]

Also may be of interest in the User's Manual:

"The power amplifier within your Sunfire Subwoofer is capable of delivering over 2,700 watts into a 3.3 ohm resistor (the voice coil resistance of the subwoofer's driver). When the same full output is applied to the driver, however, the enormous back-electromotive force generated as a consequence of its large motion and giant magnet causes the current flow to be much less than if it were in a 3.3 ohm resistor. It is this signular property of the driver that allows the woofer to be approximately ten times more efficient than a woofer this size would normally be. Hence, the actual average input power is substantially less than 2,700 watts, and is approximately 380 watts for most musical material on the loudest passages.

Another way to say "our speaker has a DC resistance of 3.3 ohms but since the operating frequency range is close to its fs, obviously the current is less so we fooled you in saying that our amplifier delivers 2700W, it delivers the voltage, yes, but fortunately (for its sake) it will never have to deliver the current" :rotfl::rotfl::rotfl:

I'm starting to understand :D

Singular property of the woofer my donkey.

[Allanvh5150]

It delivers peak current of what is needed to deliver the peak power. Think "PEAK".

Allan.

[Eaulive]

It delivers peak current of what is needed to deliver the peak power. Think "PEAK".

Allan.

You're gonna think that I'm fussy, and your right. "Peak" in this case is a misnomer, it should be "burst".
Peak is the highest voltage of a waveform, any waveform. Time is not a part of the equation, you can read a continuous peak voltage or current. Burst however is more appropriate to define the current debated situation, some manufacturers state the "burst power output" of an amplifier and they also indicate the lenght and maybe even the frequency of the burst. Yorville does for example: http://www.yorkville.com/products.asp?type=30&cat=8&id=20#overview

They state the continuous avg power and the burst avg power.

From the manual:
The burst measurements use a 10ms burst at 1kHz with 1/8 sec (125ms) pause between bursts