H744 regulator whistle - test-drb@ccmp.vtda.org

13 Jan 2002

On Jan 13, 20:27, Tony Duell wrote:

...
  Yes, it's certainly a switching regulator, and the
inductor (the large
 block with screw-post terminals) is needed for it to work. Don't try
 shorting it out -- I am not sure what happens if you do, but I would hope
 the crowbar would fire as soon as the chopper turned on and supplied
 30V-ish to the output (with no inductor, the output voltage would rise
 essentially intantaneously).

 These supplies will whistle if :

 The capacitors -- particularly the output one -- have high ESR. Check or
 replace them. 
Done.

...
  The load is just 'wrong' (these, AFAIK, are
not constant-frequency
 supplies). Try adding or removing load (!). 
Tried that.  At very light loads, the whistle almost disappears.  It gets
louder and the pitch changes as the load increases.  The most load I've
tried is a full backplane, though I'm not exactly sure what the current
drawn would be.

...
  The Inductor potting compount is breaking up. 
Swapping the inductors between 4 x H744 and 2 x H745 makes no apparent
difference, and I expect it's unlikely they've all suffered the same fate.
 They came from two different machines.

...
  Loose fixing screws can cause odd noises -- try
tightening everything in
 the PSU chassis. 
One of the first things I checked :-)

I suspect, as Allison wrote, that it's just the normal noise from that type
of regulator, and I'm just a bit oversensitive to it :-(

Thanks for the various suggestions, everyone.  I'll try putting covers on a
couple of them (one already has one, and some of the other regulators do as
well) and mounting them on a sheet of damping material, and learn to live
with the residual noise.

-- 
Pete						Peter Turnbull
						Network Manager
						University of York