Protecting Compression Drivers

[PrestonTom]

I am wondering what some of the folks are doing to provide protection for the compression drivers on their Jubilees, especially the case where an electronic crossover is being used.

There are two possible problems.

The first is protection from simply over driving the K69 driver. Are you guys using a polyswitch or wiring some auto bulbs in series with the driver? I am never satisfied that a fuse offers real protection (but I might be wrong). I notice that Zenner diode pairs do not seem to be used by Klipsch anymore

The second is protection from DC. There is a possibility that the amp (after the electronic crossover) could pass DC (from a glitch somewhere). One strategy is to wire a capacitor in series with the driver. The chosen value could provide a corner frequency at least an octave below the CF of the crossover and still provide some protection from DC. Or perhaps you are using amps that are AC coupled.

How have you solved the problem or what do you recommend?

Good Luck,

-Tom

[mas]

Tom, the idea of using bulbs in conjunction with drivers was pioneered by Community.

Talking to Bruce Howse, i have come to discover that this process is definitely NOT as simple as it may seem. Complex impedance modeling and designing a circuit very specifically for EACH individual load is required, otherwise the bulb is just a very pretty cosmetic addition.

Not only that, but exactly how they are doing this is still a tightly guarded proprietary methodology.

There is definitely something to it. But (as I have discovered) implementing it in a truly meaningful and effective manner is JUST a bit more involved!

[PrestonTom]

MAS, As far as problem # 1 and the bulbs: I certainly do not know the history on this; however I believe that Peavey actually sells a kit that includes the bulb (and possibly a polyswitch). Although I only quickly glanced at it, I saw no mention that it was difficult to fit to specific set ups. The board actually looked pretty simple. It may be possible that it does not provide very good protection - I have no idea. I understand your point that a varying impedance could be problematic. Again, this is certainly not my idea. Others have used such circuits and I have no idea what their effectiveness was.

I imagine this must be a very common problem in sound reinforcement. I also imagine that providing protection has probably been solved a number of times and in a number of different ways. In my case I know from expereince 1) that DC has a way of creeping into a circuit - i.e., glitches do occur and 2) sometimes the levels to the drivers can get turned up too far - i.e., either friends drink too much or 2 year olds are allowed to roam.

Good Luck,

-Tom

[mas]

Tom, I fear you may be looking for a solution in an area where there are only better compromises without guaranty!

To paraphase a few points expressed well at http://www.churchsoundcheck.com/faq4.html (one of the best resources available) by some of the best in the business:.:May I suggest you take a look at the entire post.

Fast blow fuses large enough that they don't blow in everyday use, but small
enough to offer SOME protection. Just realize that it will not be perfect
protection! A fuse can also add some distortion to the sound - particularly if
it is close to blowing.

More speakers are blown from the use of too small amplifiers than too large.
If your amplifier never clips, then it is unlikely that you will damage speakers
in normal use. This assumes you are using speakers reasonably close to powerful
enough for your application. In systems I design I always include a limiter
before the power amp that is set to prevent clipping. Where budget allows I
always use an amplifier at least rated to deliver the manufacturers rated power
into a speaker - even when I know that much power will never be needed."

Bulbs have been used in lieu of fuses for quite some time as they usually don't blow, but they add significant distortion. Community has refined this. In pro systems, the normal protection method is via a limiter.

Also, the following comments of Andrew Smithies are excellent:

It seems
that all is forgotten regarding the ideal amplifier output impedances when we consider light bulbs for
speaker protection. First some comments on the idea amplifier for driving the ideal speaker.

The ideal
speaker is a device that accurately converts supplied voltage into speaker cone displacement which
translates directly in to sound pressure waves. The departure from the ideal is that the speaker cone
and coil assemblies have mass
and are mounted on an elastic mount. It is therefore a resonant system. Speaker manufacturers attempt to
remove this resonance by damping the cones movement and therefore making the speaker less efficient.
(There are some grand old
speakers such as those made by Langsen that are incredibly efficient. The
shame is that they have only
average amounts of frequency bandwidth.) Another way of helping the speaker problem is to rely on the
EMF generated in the speaker voice coil as the cone moves in the magnetic field. By making the
speaker a low impedance
device. i.e. a few ohms. And then by using an even lower impedance driver such as an idea amplifier,
the amplifier has some control of the velocity of the speaker cone. A quality amplifier will have an
output impedance of a few milliohms. (plus speaker leads) Ideally this would be zero.

Now to the
problem of speaker protection. I've described the speaker above. If lots of volts are pumped into a
speaker (e.g. an amplifier when an output device goes short circuit) there WILL be lots of current
IMMEDIATELY. The first 0.1
milliseconds will go directly to the tweeter through your faithful crossover
unit. If your unlucky the
voice coil on the tweeter will either evaporate or depart from is mounting! Within a millisecond
the amps of current will have been removed from the remains of the tweeter and will be applied
to the mid and then bass
speakers. After several milliseconds of this the fast blow cartridge type
fuse will have melted the
wire. But wait! there's more! As these fuses rely on an air gap there will be a spark
maintained across the first bit of the fuse wire that melts away. Since the amp has failed with DC on
the output. Theamplifier
output voltage maintains the arc. Depending on the crossover design when the fuse finally blows the
input to the crossover network will go to the opposite polarity of the DC delivered by the amp. All
energy stored in the crossover's
inductors will be delivered to the tweeter and mid range speakers. This ought to finish them off!

This is
very sad. This will be more spectacular the bigger the amp is relative to the speaker size. Bear in mind
though that a bigger amp is less likely to overload and therefore die. But don't be to sure of yourself
if your system has a
very big amp compared to the speakers. Overload is not the only mechanism that
can make an amp go ballistic.
It is however, surprising how much a speaker can survive.

OK then. So
a fuse is no good. What about the light bulb you ask? A light bulb is a simple piece of resistive wire
with a positive temperature coefficient. It goes up in resistance with an increase in temperature. A
filament operates at white hot temperatures. 3000 odd degrees Celsius. As a
speaker protector during normal
operation it remains cold and has a low impedance. During an overload there
is lots of current. The
filament gets hot its resistance goes up (by a factor of 10) and
gracefully reduces the power to the speakers preventing any nasty back EMFs.

It is not
(a good) idea though for two reasons.

1) The
filament has thermal mass. It will do absolutely nothing to stop the first
0.1 milliseconds of power sent
to the tweeter.

2) The bulb
has some resistance even when cold. The reduces the ability of the amplifier to control the speaker
cone velocity. (This is referred to the amplifier's 'damping factor'. Many speakers cope
ok with a low damping factor but I would like to hear comments on what difference can be heard on
Bose and other speaker systems that use precorrection on the amplifier input. (PS. (Most) Valve amps (McIntosh being an exception) have a low
damping factor anyway
because they use small amounts on negative feedback. One main reason for this is
that valve amps are hard to apply negative feedback to due to limitations
in the transformers required
in a valve amp)

The only
real way to protect a speaker system is to monitor the amount of energy applied to each driver in the
speaker system and ensure this is not exceeded. (A speaker system will have a tweeter that can handle
only a fraction of the system
rating) In order to protect the tweeter the protector must remove the applied
signal very fast. Most big
relays take 10 ms or so to operate. This is far to slow for a tweeter. Some relays are fast
but I challenge anyone to find one big enough to operate in less than a millisecond. It is ok for a
bass speaker though. This leaves only electronic devices to do the current breaking. Back to back
mosfets or bipolar
transistors would do but they too could die when needed most, have some internal resistance and are
nonlinear. A complete protector would also take care of any energy stored in the cross
over network.

My thoughts
for protection is that you cannot have everything. Expect to pay either in dollars, sound quality or
potential for smoke (more $).

To make
something fully bullet proof an individual power sensor / trip and isolator circuit for each driver
that is powered separately from the rest of the PA is about the only solution I can see working. This
would be quite complex. In
all situations I suggest that a selected, trained group of people do all the
operating. This requires much
dedication. Also choose a system with plenty of headroom. Especially in the speaker department. This
way the voice coils will have
more thermal mass and strength than the fuses.

Peak
limiting is a more expensive choice, and must be
set low enough so damage does not result, but at the
same time no too low that the dynamics in your presentation are lost. The most expensive solution is found with some processed speaker systems (eg. Meyer, Apogee) which have processors that actually sense the
voltage present on the speaker lines and then apply
the proper amount of limiting necessary for the
particular speaker model.

Bottomline, proper design and responsible operations is more effective than depending upon any magical device. But if you are really concerned, a limiter prior to the amp is always a NICE option!

[ZAKO]

Marantz used light bulbs as protection in the Mod.15 amp. 25yrs ago...

[sfogg]

Tom,

Amps have DC protection circuits (relays open if DC is detected) and they are AC coupled.

As far as overdriving the compression driver I just don't worry about it. It is rated for 200w peak power and 50 watts continuous with a 500hz crossover. The amp driving it has 9w available. Even full on clipping with all its HF harmonics just doesn't have enough power to do much.

The systems gain is setup so you really can't clip it anyway. And when the kid(s) get old enough to have to worry about it the pre-amp has a max volume function where I can limit how loud the system can be turned up.

Shawn

[PrestonTom]

Shawn, you are probably right.

The amps I will be using are Adcom 545ii & 555ii. Their power bandwidth indicates that they are probably AC coupled (although I do not know if it has DC protection, or if it would be fast enough). The amps are after a 4 channel "passive pre-amp" (simply: dual, stereo 10k pots). Before the preamp is the Behringer DCX (electronic crossover). Its output might actually be AC coupled also.

As far as the gain, I am a stickler for using all the bits in the Behringer's DAC and the Adcoms are capable of 100 & 200 watts (yes, I know this is gross overkill for high efficiency speakers ...). That is why I was worried about an "accident" where the volume knob was turned too far.

So you are suggesting that I don't even bother with the capacitor in series with the driver. This makes for a simple circuit.

-Tom

[WMcD]

I have a few observations without any real answers.

1) I don't know that there would be a problem with the DSP to amp set up at home. Certainly movie theaters are running similar equipment and throwing a lot more power at the driver. Somewhat similarly, HT systems have a lot of DSP and we have not heard of problems with less robust drivers which should burn out sooner.

2) I do agree with the general thought that accidents happen. I blew out a 200 watt driver in a bass horn by playing the Telarc "warning" T-Rex track with a 20 watt amp. I suspect the amp went nuts, or hit the driver at resonance. I also blew out a tweeter with a test tone when I failed to pay attention. They were expensive lessons. Despite, this, I've not gone into protection circuits.

3) Our moderator here, long ago, pointed out an issue with burned out drivers. A second order crossover forms an LC series circuit across the amp when the driver blows. Therefore you get a short to ground across the amp at some frequency and the amp might blow. Something to consider. If you put a fuse on the speaker side of a passive crossover, you get the same. So be careful.

4) People, probably correctly, point out that fuses can't act quickly enough.

5) The back to back Zeners across the driver do act as a "clamp". That is to say they clamp voltage but also start acting as a short circuit. Jim Hunter commented that this was not a nice thing to do to the amp.

6) I wonder if 4) and 5) would not be a good solution. If the Zener's clamp, then they'd blow a fuse. I'm not quite sure what could happen if we assume there is a lot of back EMF from the driver and it is looking into the very non-linear back to back Zeners.

7) I don't really have sympathy for the problem. Many of the situations needing protection are "near train wreck" circumstances. People want to flirt with danger of "actual train wreck" and then want protection from the folly. Granted there are situations of a tyke or party animal setting the volume to 10 and then hitting the ON switch. It happens. .

Smile,

Gil

[sfogg]

Tom,

"As far as the gain, I am a stickler for using all the bits in the Behringer's DAC and the Adcoms are capable of 100 & 200 watts (yes, I know this is gross overkill for high efficiency speakers ...). That is why I was worried about an "accident" where the volume knob was turned too far."

Are you doing volume ahead or behind the Behringer?

If you do it ahead (with attenuation behind) you can basically set up the system such that you won't be able to drive the amps to anywhere near full output. That gives added safety.

"So you are suggesting that I don't even bother with the capacitor in series with the driver. This makes for a simple circuit."

There is another thing to consider... does the Adcom's have turn on/off transients. I use a pair of 535IIs for my surrounds and I think they might put out a little bit of DC when they shut off. That wouldn't be good for a compression driver so a capacitor to block DC might be in order.

With my amps I don't worry about it. Not much power to begin with and they have the relay switched outputs that have the double benefit of having DC protection as well as disconnecting the outputs on turn on/off to avoid any transients.

Shawn

[DrWho]

Hey Shawn, if you still have your Behringer open, could you provide the part numbers for the DACs and the opamps being used on the input/output? I wonder if there isn't a drop-in replacement that operates over a smaller voltage range...and then you can simplify the whole optimizing the digital section of the Behringer. Or maybe there are easier ways to do this, I dunno. It just seems like such a simple mod if you can do it without changing any traces.

As far as back to back zeners....I blew up an amp in the lab not thinking about what happens when they go into protect mode...talk about an expensive mistake [A] If you do use zeners, just make sure you have some kind of load in the circuit for the amp to drive.

At work we have the lightbulb protectors on a pair of McCauley mains, but man they really distort the signal when they're lighting up. I'm not sure if these speakers were custom built or not - they're using ERSE crossovers inside. The tungsten filaments are only used on the HF section of the network.

http://www.erse.cc/xover/Results.asp?part=2wayc&f=350&t=399

I wouldn't recommend them in a normal hifi setting because you're really never going to be pushing the steady state of the system that loud (or at least shouldn't be....)

I really think the capacitor idea is a good idea though. While you can have electronics upstream filtering the signal, there is nothing preventing noise from being introduced between the active crossover and the amplifier. Twice now I've seen tweeter diaphragms launched because the input cables to the HF amp went bad and got bumped in such a way as to produce a loud 60Hz. Most tweeters don't like 60Hz played at the full output of the amp. Caps are also nice because you don't have to worry about any potential DC offsets either. I have never noticed any kind of degradation when using a cap in series with the tweeter, but then most of my experiences with that have been in prosound settings and I wouldn't really expect it to be audible in that setting if it did pose a problem.

[Islander]

As well as drunks and children, people who normally have a clue can make a mistake. It could be something as simple as hooking up a CD player directly to a power amp. That way, there is no volume control and the amp puts out full power until it's turned off or disconnected. I made this mistake when testing my new amp with a portable CD player. I had no idea the line out socket bypassed the volume control. Luckily, the first song had a spoken word intro, so I was able to disconnect before it got too loud.

When I took the amp to the dealer to check it for buzzing when hooked up to their equipment, the salesman did the same thing, hooking the amp directly to a CD player instead of to a receiver. It put out its full power for 5 or 10 seconds before he realized the problem and shut off the CD player. It was fairly loud.

Maybe I just got lucky, but there was no harm done in either case, to the amp or to the speakers. Also, there was no buzzing with those low-sensitivity Totem Wind speakers, although it can be heard faintly with the La Scalas.

[sfogg]

Mike,

"Hey Shawn, if you still have your Behringer open, could you provide the part numbers for the DACs and the opamps being used on the input/output?"

It isn't the DAC that creates the larger output voltage but the output line stage itself.

The DACs are AKM units. I forget the exact part number but can check it tonight.... I think it is 3x AKM4393. The DACs actually have a straight typical 2v output on them. Some people modify the Behringers by taking that signal and running it through a passive low pass filter and just take that straight out of the unit bypassing the rest of the analog stage. That avoids some of the mute circuitry though which might result in turn on/off transients. There are also other output stages that can be dropped in place of the stock setup. Or a couple of resistor takes care of the high signal levels just fine too.

"It just seems like such a simple mod if you can do it without changing any traces."

To do it without changing any traces find which opamp(s) are adding all the gain to the analog section and change their gain. That would just involve swapping out some components for different values.

What you need to do is build S/PDIF inputs on the amps you are working on (support 96kHz sampling rate) and add some form of volume control directly in the amps modulators. Then just mod the Behringer for 3x S/PDIF outputs with one of my boards and you are golden.

"I really think the capacitor idea is a good idea though."

In the past I have done this too. Just make sure it is far away from the crossover point so it doesn't effect that. Also take into account the 90 degree phase shift this is going to add.

Shawn

[DrWho]

To do it without changing any traces find which opamp(s) are adding all the gain to the analog section and change their gain. That would just involve swapping out some components for different values.

That's pretty much what I was thinking. Changing the opamp might have some other benefits since you're not going to be using it's full potential either, and most of them seem to be using the same pinouts.

What you need to do is build S/PDIF inputs on the amps you are working on (support 96kHz sampling rate) and add some form of volume control directly in the amps modulators. Then just mod the Behringer for 3x S/PDIF outputs with one of my boards and you are golden.

lol - not a bad idea though. If only there were a clean way to adjust the volume of each amp so that everything stayed synced up perfectly...and then you could go a step further and put the amps right by the speakers and take advantage of shorter speaker wire runs and all that shnazz. Sounds like a good project for next semester.

"I really think the capacitor idea is a good idea though."

In the past I have done this too. Just make sure it is far away from the crossover point so it doesn't effect that. Also take into account the 90 degree phase shift this is going to add.

I'm pretty sure you don't get any phase shift in the passband - you only get 90 degrees in the stop band. Or am I missing something here?

[sfogg]

"Changing the opamp might have some other benefits since you're not going to be using it's full potential either, and most of them seem to be using the same pinouts."

Yes, many have the same pinouts. Swapping them is easy enough if you have SMT soldering equipment. If you go that route though be sure whatever you put in behaves in the circuit. Sometimes people do things like opamp swapouts with much faster parts and they end up not being stable in the circuit due to lousy power supplies or whatever.

"If only there were a clean way to adjust the volume of each amp so that everything stayed synced up perfectly..."

Build it.

Check out how Tact audio controls volume (at least partially) on their digital amps. I'm pretty sure they modulate the power supply as part of the volume control so you can adjust output levels without simply 'dropping bits.'

If I find the time (not terribly likely right now) I may mod one of the Panny digital receivers for 3x S/PDIF inputs and use its own volume control to handle level adjustments.

Shawn

[DrWho]

Modulating the power supply won't work on our amps...it's kinda freaky really, we've got the power supply connected to an autoformer because all of our caps will blow the fuses if we do an instant turn-on (and we haven't implementing the soft start circuitry yet). Anyways, you can turn the dial on the autoformer between whatever voltages you want as long as the 80% of the resultant DC rail voltage is higher than the output of the amp.

The research we've done so far indicates that most people are using optocouplers for the volume control, which apparently seems to work well once the everything is calibrated. It's not really a scalable concept though.

[Guest srobak]

American Acoustics Labs used lightbulbs in their crossover assemblies for protection. Seemed to work pretty nicely on the DS-1212's I had way back in the day.