Protecting my bench test-speakers

[diydidi]

So the other day a friend brings his bass guitar around for some re-wiring. After all was done and looking good, I plug the guitar into my bench power amp and through to my speakers. It works awesome, nice growl from the bass, until I pull the pull-switch on the guitar to activate a different EQ setting. The damn thing oscillates and sends a 150Hz square wave through my speakers. It wasn't even 5 seconds. My speaker coils all fried.
What do you guys do for protection of your lab speakers? I'm open for ideas.
I will be buying new speakers soon, and don't want them to run the same fate..

[km6xz]

At high SPL, speakers become more of the sound than the amp so for testing purposes I use an passive attenuator in the form of the regular shop dummy loads arranged in a series of 200 watt 2 ohm resistors with taps for putting them in series parallel to get 2 ohm load at 800 watts, 4 ohm, 8 and 16, and use a blocking cap in series with the speaker lead, which in turn connects to a tap through a series resistor. So at 800 watts, into the dummy load, the speaker sees 20 watts which is plenty to hear characteristics of the amplifier and not so much the speaker. It is protected from DC offset. This was a simple system, my old shop had an active load bank.

[The Dude]

You can be as careful as possible, and this will still happen on occasion (hopefully rarely!). Often you'll have a piece fixed and halfway through the burn in......bang! I have a variac with variable current limiting that has paid for itself in "re-repairs". It will normally trip before further damage is done to the repair piece or my bench speakers. Other than that, my best advice is- don't use expensive speakers for testing gear.

[J M Fahey]

So the other day a friend brings his bass guitar around for some re-wiring. After all was done and looking good, I plug the guitar into my bench power amp and through to my speakers. It works awesome, nice growl from the bass, until I pull the pull-switch on the guitar to activate a different EQ setting. The damn thing oscillates and sends a 150Hz square wave through my speakers. It wasn't even 5 seconds. My speaker coils all fried.
What do you guys do for protection of your lab speakers? I'm open for ideas.
I will be buying new speakers soon, and don't want them to run the same fate..

Well, if your bench amplifier destroyed your bench speakers in 5 seconds, then your bench system was very poorly matched.
The speaker must stand *any* frequency and waveform the amp throws at it.
So if you use an "X" watts amp, the speaker must stand at least 1.5"X" watts, as in any MI amplifier.
Using a very underpowered speaker and trusting "I won't play loud" is dangerous, as the practical example above shows.
I usually have one of my 80W 12" speakers in a 1 cu ft box by my bench, and have no problems at all.
Can be used for quick tests with, say, a Plexi 100W on 10 .
In fact, after 4 or 5 minutes full tilt I start to perceive hot epoxy smell from the VC and stop the test.
But .... 5 seconds? Ouch !!

[Enzo]

The secret to testing a 500 watt amp is not to hope you can connect it to 50 watt speakers an "only turn it up part way". Your test speakers and/or test loads must be able to handle the entire power of the unit under test. So I might use 500 watt speakers or I might use Stan's dummy load with monitor speaker or maybe a straight dummy load and a scope. But regardless, whichever I chose, it must handle the entire 500 watts of the example.

[km6xz]

That is a good point, connecting speakers is the last step in confirming that it is stable with an inductive dynamic load. That is after the unit under test is considered done with repairs. The loads and scope will tell you more of what you want to know than a speaker will up to that point. With a solid state amp, I normally use a 1k resistor for a load, feeding the scope and HP3580A spectrum analyzer tell me a lot about the amp's performance before ever connecting a dummy load.
Tube amps need a load so they are run on a load always. In my new small shop I do not even have large speakers, just some small 30 watt cabinets to be used as a monitor at low volumes. I have found that loud sounds in the shop distract attention from the task at hand which is to measure and then hear the characteristics of the amplifier, which really can't be done at high volumes. We are a lot less sensitive to the low amplitude nuances that create the character of an amp. Listening to a Marshall Major at full roar tells you very little about the amp, more about the speaker. But is the load is handling the full power and a small sample is being tapped to drive a low power speaker driven to a fraction of its rating, the harmonic content is very clearly heard without the non-linear characteristics of a large speaker driven very close to its destruction, over shadowing the amp sound.
I find that a scope and spectrum analyzer is a great way to telling at a glance what the amp is doing. If you have a spare computer with a sound card, and a few minutes building a suitable interface that limits input to the card to about 1 volt, you can have a very high resolution audio spectrum analyzer. There are a number of free programs that perform these tasks plus tracking sweep generate, square wave, pseudo random noise generate(White and Pink) and a very low distortion sine generator.
One example is: Visual Analyser details
I have used it for setting up P.A systems using a calibrated mic but you can use any mic by having it construct a compensation table with its built-in calibrator. I have an old spare laptop with it installed and it make a nice highly portable test set.
You can capture spectral displays of known good examples of amps that you run into so compare to units under test at some standard test conditions.

[diydidi]

Thanx guys far all the replies.
So do you guys thing I need only one speaker at hand?
My gear that I have are as follows. A stereo @x800W RMS amp, a 1000W dummy load with 2x 8ohm inputs and one input connected to an attenuator rotary switch for monitor speaker option.
does this mean I need a 1,5 * 800w speaker??

[km6xz]

Well, it all depends on what you use monitors for. Like stage monitors, they are used for a variety of things but all that is really needed is auditory cuing for distortion, and tone both of which are more easily done at low levels. Stage monitors have grown so loud and full spectrum that the original purpose of pitch and timing cuing has been obliterated, and are essentially slightly smaller versions of the already too loud and too bassy house system which hides that important information. If you are working on stereo systems and want to hear a unit under test, I suppose you might want 2, 5.1 or 7.1 speakers but I wouldn't.
In my old shop I had some high end B+W speakers for my bench but with 24 other benches in the shop, the noise could be hard to ignore so most of my monitoring was at very low levels or with headphones.
So it all depends on your use of the monitors, as a tool for evaluation of sonic deviation or a reproduction of the home system of the client. You really are not helped by having so much power that is capable of ruining any set of speakers if the wrong signal is sent to them. After all, the wrong signal is something that a defective unit often does send.

[J M Fahey]

does this mean I need a 1,5 * 800w speaker??

If you connect it direct, as you seem to imply by the "less than 5 seconds" bit, yes.

I'd rather not use an 800W "Bench amp" but, hey, my workbench covers less than an acre.

Or, to solve your problem in a more practical way; you need a bench amp not louder than 0.8X what your bench speaker takes safely.