Subject: Re: serial port control
To: Ted Lemon <>
From: Wolfgang Rupprecht <>
List: current-users
Date: 02/12/1998 15:20:37
Ted Lemon writes:
> Also, I don't think CRTs use switching supplies, do they?

I just pulled apart a Sun 20" (sony) monitor and had a peek.  There is
nothing that I can see thats big enough to be a ~200watt transformer.

I can't pull apart this Trinitron on the other computer now, Alison
will be home in a while and she would have my head if I was visibly
messing with her monitor. ;-)

I kind of suspect that it would be too challenging to shield a big
200watt transformer enough to not cause 60hz picture interference.
Remember the magnetic deflection circuit in the crt needs to have less
than 1 part in 1000 magnetic noise in order to do a non-jittery 1024
pixels across the screen.  This is best done by making small
transformers that one can stick a copper shorting strap around.  The
60hz noise from a big transformer would cost a bundle to sheild.

> Is it possible that a typical (cheap!) switching supply makes
> assumptions about its supply that aren't valid if the supply
> produces a square wave, and that this could cause the linear supply
> to behave badly even though an ideal linear supply might love a
> square-wave input?

I'm sure they wouldn't hesitate to make any assumptions that would
save money.  I don't see any that they could make.

> The reason I ask is that I have definitely heard of switching
> supplies being fried by square-wave UPS's.  Could be Urban Legend, I
> suppose.

I tend to believe stuff like this.  Usually the story is true, but for
a slightly different reason than the folks telling it think.  

I'm beginning to think that the real reason is that a square-wave
output UPS is an anything but a square wave, and is so dirty that
anything anything but an electric hairdryer is going to complain.

Just to bury some netbsd relevant info here (is anyone else still
paying attention? ;-)) I got a email note from APC saying they would
license the APC Smart-UPS protocol for a nominal $50 fee.  Heck, I'll
pay the $50 bucks if it will let me implement the full daemon and if I
could then donate the code to NetBSD.

Unfortunately I can't read the fine print of the document.  It came
across as a uunecoded file for what appears to be for some BillyGates
abomination.  Anyone care to have a crack at this?  The content-type
specs say:

    Content-Type: application/winword6; name="LINKAGRE.DOC"
    Content-Transfer-Encoding: base64

Wolfgang Rupprecht    <>
	  Never trust a program you don't have sources for.