Recently bought some big PA Mosfet amps off a mate of mine.

[JimC]

pure DC is being applied to the voice coil, usually at several amps if it's a high-power amplifier, it doesn't take long for the voice coil to burn-out and become open-circuited.

I think you'll find the tales about DC are pretty much a myth. This is an AC signal like any other. It's not DC at all, even if the signal, by the time it got through the wiring, any crossover and anything else, were still a pure square wave. What it is is a very high energy signal, and any sufficiently high energy signal will burn out the coil. Most likely what causes the confusion is that a nominal 1KW amp will deliver 1KW as an undistorted sine wave, but getting on for double that if you actually drove it into the unlistenable level of distortion a pure square wave would imply. Real DC does have an extra risk, since the voice coil isn't moving, which deprives it of cooling, but an AC square wave has the cone moving just as much as any other signal, and provides the same air movement and cooling.

For all the myths around square waves, we should remember they're one of the main building blocks of additive synthesis, and there's no risk at all in reproducing them provided that the drivers aren't overloaded.

[Marcel]

Oh wow, calling on detailed memory from 32y ago....

The Jands 920 was a brute at 450W RMS per channel into 8 ohms with 0.1% THD. It will do 4 ohms but not that well. Due analogue (non-switchmode) power supply constraints of the 920 and transistor based output finals design the best I measured back then at 4 ohms was 550W with 2% THD.

Modern switchmode amps with MOSFET output stages will do 4 or 2 ohms with relative ease with more or less doubling of output power with each halving of load impedance and virtually nil change in THD. The biggest difference and swaying factor for the committee in the selection of the 920 for Expo 88 was the soft clipping in the 920 at max output as opposed to the very square hard clipping of the Yamaha, Crown and Perreaux competition despite their better THD figures at typically better than 0.01%

All the PA's that used 920's back then in the late '80's were configured with 8 ohm or 16 ohm combined speaker loads. It was rare for any 920 system to use 4 ohm combined loads. And a fast way to blow 920 amps was to mistakenly wire up 2 ohm loads as the amp truly doesn't like feeding 2 ohms or less. A daily check of each drivers functionality at Expo 88 was done by pulling the XLR connectors on the patch panel behind each amp. Incorrect re-insertion of the XLR could cause a momentary short across the amp outputs instigating instant failure of the amp final output stage. I repaired dozens of failed channels across the duration of the 6 month Expo 88. The indication of a channel failed in this way is the "DC" alarm lamp being lit.

Through a neat set of circumstances I got a 920 cheap. Had it in my home stereo for a number of years beside my CD player and dual 15 band EQ. Burnt out quite a few 12" drivers at various parties until my partner convinced me to sell it and buy a new TV.

Hope this helps...

[Simon Barden]

I think you'll find the tales about DC are pretty much a myth. This is an AC signal like any other. It's not DC at all, even if the signal, by the time it got through the wiring, any crossover and anything else, were still a pure square wave. What it is is a very high energy signal, and any sufficiently high energy signal will burn out the coil. Most likely what causes the confusion is that a nominal 1KW amp will deliver 1KW as an undistorted sine wave, but getting on for double that if you actually drove it into the unlistenable level of distortion a pure square wave would imply. Real DC does have an extra risk, since the voice coil isn't moving, which deprives it of cooling, but an AC square wave has the cone moving just as much as any other signal, and provides the same air movement and cooling.

For all the myths around square waves, we should remember they're one of the main building blocks of additive synthesis, and there's no risk at all in reproducing them provided that the drivers aren't overloaded.

Some qualifications needed there, Jim. A square wave is a ‘high energy signal’ in a relative sense to say a sine wave or a triangle wave of the same amplitude, but a 1mW square wave is far less powerful and has far less energy than a 10W sine wave. A constant square wave contains the most % energy you can transmit for a given signal amplitude.

At very low frequencies, a square wave will approximate to DC being passed through a speaker coil e.g. at 1Hz, you’ll get alternating patterns of 1 second of positive DC signal and then 1 second of negative DC signal. The speaker coil will move to a position, stay in that position with the DC current passing through it, and then move to the next position and stay in that position with the DC current passing through it.

The thing that’s different between a 1Hz square wave and a 100Hz square wave is the amount of movement of the speaker and its coil between those two positions. A speaker relies on movement of the cone and voice coil to provide most of its heat dissipation through the air. If it’s not moving much, then it can’t dissipate much heat and a 100W rated speaker may fail at 3 or 4 watts once enough heat builds up. It’s a similar thing with very high frequency signals – although the speaker is in theory moving very quickly and so should get well ventilated, a point beyond which the mass and inertia of a speaker simply can’t respond fast enough to the signal to move much at all. So, the cone/coil hardly moves whilst a large amount of energy can be passing through it.

The other thing DC does to a speaker is provide an offset to the normal rest position of the voice coil within the magnet, either moving it forwards or backwards in proportion to the level of the DC signal. Let’s say a voice coil normally has a maximum travel of ±4mm before the voice coil starts to move outside the magnet’s main field and so starts to lose control of its position. A DC signal may move the voice coil forwards by 2mm, so it doesn’t take as much power for it to reach that 4mm of forward travel, after which the voice coil isn’t held centrally by the magnetic files but can be pulled to one side, rubbing against the magnet and wearing away insulation or even catching on the edge of the magnet. Both situations spell a short and unhappy future life for the driver.

So mixing a very low frequency square- or near-square wave in with a normal music/guitar/bass signal isn’t a good idea, even though the speaker is still moving a lot and being cooled.

In fact, any very low frequency wave alone being passed through a speaker isn’t good for it, it’s just that a square wave is the worst sort of wave to do that with.

It’s why most amps have DC blocking capacitors in the signal path, and many bass amps have high-pass filtering to cut off sub-sonic frequencies (you are far less likely to generate any with a guitar).

But, messing about with synths is the most likely way to get very low frequency square waves in the real world, with low frequency oscillators capable of say 0.1Hz to 20Hz affecting output levels Some can be routed to audio outputs, and if not actually modulating an audible wave, the results will be too low for you to hear, so you won’t know what’s going on until the blue smoke of death starts to appear.

[DrNomis_44]

Just did a quick calculation and, if you have 20 X 450 Watt Power Amps all running together, that gives you a total combined power-output of 9,000 Watts.....yikes!!!!!

[Simon Barden]

9kW is nothing these days. But with 2 channels per amp, that set up was probably 18kW

My small Yamaha PA system - 2 x DXR10s and a DXS12 sub was rated at around 3kW continuous and 5kW peak.

But it's the speaker efficiency that really determines the overall system loudness.

[DrNomis_44]

9kW is nothing these days. But with 2 channels per amp, that set up was probably 18kW

My small Yamaha PA system - 2 x DXR10s and a DXS12 sub was rated at around 3kW continuous and 5kW peak.

But it's the speaker efficiency that really determines the overall system loudness.

Yep, basically how many dB per Watts SPL the speaker would put-out, taking into consideration how the load-impedance varies with frequency, speaker cab resonances, and the frequency response of the speakers, you're quite right, I forgot that each of the 20 450W amps were stereo so 450W per stereo channel.


So, a speaker that has an efficiency of say 105dB/W is going to be more efficient, and sound louder, than a similarly specced speaker with an efficiency rating of say 75dB/W, because it is more efficient at converting the electrical power (energy) output of the amp into acoustic power (energy).

[Marcel]

Yep, basically how many dB per Watts SPL the speaker would put-out, taking into consideration how the load-impedance varies with frequency, speaker cab resonances, and the frequency response of the speakers, you're quite right, I forgot that each of the 20 450W amps were stereo so 450W per stereo channel.


So, a speaker that has an efficiency of say 105dB/W is going to be more efficient, and sound louder, than a similarly specced speaker with an efficiency rating of say 75dB/W, because it is more efficient at converting the electrical power (energy) output of the amp into acoustic power (energy).

Actually for the River Stage there were 23 Jands 920's per side.... 20 on the FOH each side (18kW + 18kW), and 3 more per side (2.7kW +2.7kW) on delayed stacks some 75m from the stage. (18 +18 + 2.7 + 2.7 = system total of 41400kW true RMS at onset of sine wave clipping)
As for speaker cabs do a search for "Concord PA cabinets" & "3way cabs" circa late '80's. There were 16 3 way cabs hung on motorized hoists in a 4x4 array per side on the stage with 2 of the 3 way cabs in each delay stack, and 4 sub cabs per side of stage for below 80Hz. ... I think the crossover points for the stereo 4 way system were 80Hz, 750Hz and 2.6kHz or somewhere near those frequencies.

As a side note....just days before the 6 month "show" started they had everything working including the 512 channel DMX controlled lights array.... PA was up making musical noise, and the lighting dude raised all his lighting channels and then the main dimmer on his console....
Next minit.... BZZZZ!!... the ground briefly shook at 50Hz, and all the lights suddenly went into uncontrollable auto disco mode with the PA making weird buzzing noises....within seconds everyone was scrambling to turn things off....
It tool a team of electricians over an hour to diagnose that there was nil Neutral connection from the river stage dedicated 250kVA supply transformer.... They were all baffled as to how a new power transformer of that size could literally burn out its Neutral link.... I laughed, and reminded them all how an SCR dimmer works.... and that there was over 200kW of incandescent lighting on that stage. 3 phase all on no problem, 3 phase all off no problem, 3 phase at 66% SCR dimming and damn huge Neutral problem. (and yeah there is another story on how I knew these things...lol)
A 750kVA transformer was fitted next day.... the next 6 months went just fine....

[DrNomis_44]

Actually for the River Stage there were 23 Jands 920's per side.... 20 on the FOH each side (18kW + 18kW), and 3 more per side (2.7kW +2.7kW) on delayed stacks some 75m from the stage. (18 +18 + 2.7 + 2.7 = system total of 41400kW true RMS at onset of sine wave clipping)
As for speaker cabs do a search for "Concord PA cabinets" & "3way cabs" circa late '80's. There were 16 3 way cabs hung on motorized hoists in a 4x4 array per side on the stage with 2 of the 3 way cabs in each delay stack, and 4 sub cabs per side of stage for below 80Hz. ... I think the crossover points for the stereo 4 way system were 80Hz, 750Hz and 2.6kHz or somewhere near those frequencies.

As a side note....just days before the 6 month "show" started they had everything working including the 512 channel DMX controlled lights array.... PA was up making musical noise, and the lighting dude raised all his lighting channels and then the main dimmer on his console....
Next minit.... BZZZZ!!... the ground briefly shook at 50Hz, and all the lights suddenly went into uncontrollable auto disco mode with the PA making weird buzzing noises....within seconds everyone was scrambling to turn things off....
It tool a team of electricians over an hour to diagnose that there was nil Neutral connection from the river stage dedicated 250kVA supply transformer.... They were all baffled as to how a new power transformer of that size could literally burn out its Neutral link.... I laughed, and reminded them all how an SCR dimmer works.... and that there was over 200kW of incandescent lighting on that stage. 3 phase all on no problem, 3 phase all off no problem, 3 phase at 66% SCR dimming and damn huge Neutral problem. (and yeah there is another story on how I knew these things...lol)
A 750kVA transformer was fitted next day.... the next 6 months went just fine....

Sounds like it must have been a fun time getting that all sorted-out, I can just imagine how loud that 50Hz buzz was at the time, and 50Hz must have been the natural resonant frequency of those light fittings too, or close to it anyway....lol.


I seem to remember studying SCRs while doing some electronics courses at uni more than ten years ago, when they are used to control AC power, the waveform isn't exactly sinusoidal all the time, it's a spiky looking waveform, and I seem to remember that snubber-circuits are sometimes added to reduce high-frequency harmonics, or, I could be totally wrong, it has been quite a while.

Still, great story you have there mate.


Note for everyone else, SCR is short for Silicon Controlled Rectifier, it's a special type of power diode used to control AC power, most often used in light dimmers, although nowadays you do see some stage lights that use Leds (Light Emitting Diodes) rather than incandescent lights, I think Led stage lights are powered by DC if I'm not mistaken.


I'm really amazed at how far P.A. sound technology has evolved since it's early days, who knows what the future has in store for us.

[Marcel]

A bit of searching and I was reminded that the speaker cabs at the entertainment venues at Expo 88 were predominately Concord TMS-3...

For a interesting read of Aussie PA and lighting history have a read of this.... it brought back plenty of memories for me of gear and people ....
http://www.australianroadcrew.com.au...production.pdf

[Simon Barden]

Interesting stuff, Marcel.