From: info@turneraudio.com.au   
      
   Phil mentioed......A test CD will give you that plus standard conditions for   
   burst frequency, length and repetition rate.    
   None of which matters one a HOOT to a hobbyist audiophile.    
   Cos he can use music programme and a scope to see what is really going on.    
   ....  Phil   
      
   Yes, music can show wat's sort of happening a bit, but then music's dynamic   
   range varies hugely between say busy music by AC-DC /noisy punk band, OR   
   something bt Motzart, or Beethoven.    
      
   I've sometimes used pink noise to test audio tube amps. The bandwidth of this   
   noise as a test signal is important, and if you use DC to 25kHz with a tube   
   amp, the very low F content causes mahem due to OPT core saturation because of   
   F below 20Hz - unless    
   the amp has sufficient bandwidth limiting with HPF C&R and input and between   
   stages. The maximum power output of the amp with a sine wave is handy to know,   
   and useful for engineers, but for hi-fi amps its an almost useless figure   
   because the max PO can    
   be deemed to be reached when the highest peaks in music are repeatedly just   
   beginning to cause clipping. So The dynamic range within the pink noise   
   becomes relevant, and that range is never infinite. But let us sippose we   
   limit the pink noise bandwidth    
   from 20Hz to 25kHz using simple R&C filtering, and then include a pole at 10Hz   
   at the tube amp input, then the occasional LF signal at max amplitude of   
   unfiltered noise will probably not cause much shit to hit fan in OPT. And the   
   shit to fan noise can be    
   heard as an irregular "knocking noise" coming from the OPT. Tubes don't like   
   it because during the "knocks", they are momentarily saturated, and with grid   
   current and there is charging up of coupling caps, so intermittent class A   
   operation and paralysis.    
      
   But one can get a fair idea of maximum PO possible with pinknoise using a CRO   
   by getting some slight clipping to occur, and setting the trace amplitude   
   while clipping to the full height of graticules. A clean sine save of 1kHz   
   from a sig gene can be used    
   to find out what Vrms voltage is needed to get the trace swing, and so then   
   you can determine the sine wave needed for clipping of an intermittent signal   
   with similar character to music. In a class AB amp, usually yhr pink noise   
   clipping voltage is more    
   than if you have a CW, understandable because of PSU rail sag.    
      
   Most levels used by most audiophiles most of the time do not make the rail   
   voltages shift at all, even with very low bias current in tubes and with load   
   matching that gives a small amount of class A PO as a % of the total AB PO   
   possible.   
      
   If one wanted a burst maker I guess one might use a 2 transistor multivibrator   
   that makes LF square wave control a couple of solid state switches. One might   
   then be able to vary the square wave timing, ie burst time, which seems   
   important if 50Hz is the    
   CW frequency, and one wondered what the performance is like with so much R&C   
   coupling and OPT coupling present.    
      
   I've got a home brew pinknoise maker, and its all I seem to need. Another way   
   is to have a simple mute switch to reduce input -20dB, and just flick it   
   on-off. This is rough, but you can get an idea between what happens with CW or   
   intermittent bursts    
   because rail sag and recovery take time. If fact there should be considerable   
   time because of large size rail caps. The behaviour also needs to be known for   
   both clipping and with say +20dB input overload, and the recovery from this.   
   Nothing should smoke.   
       
   Patrick Turner.   
      
   --- SoupGate-Win32 v1.05   
    * Origin: you cannot sedate... all the things you hate (1:229/2)