[Prev][Next][Index][Thread]

Re: Musein' and mathin'



Subject: 
           Re: Musein' and mathin'
      Date: 
           Sat, 15 Mar 1997 14:28:18 -0800
      From: 
           Greg Leyh <lod-at-pacbell-dot-net>
        To: 
           Tesla List <tesla-at-pupman-dot-com>
References: 
           1


> Pspice, like anything else on the CPU, is not auto-erecting and talent is
> still required to formulate a good model followed by the grueling input
> machinations.
> 
> Richard Hull, TCBOR


Richard,

All too true. The use of PSPICE as the sole design tool in TC work
without an 
experimental background could lead to some very costly heartbreaks,
indeed.  
For the most part I use PSPICE as an accounting measure, in order to
make sure
preliminary component values are adequate, and to find gross design
blunders.

For instance, shortly after I had built my coil the RF choke on the DC
line
kept overheating; the first one (PVC) melted! The single-layer choke was
wound
on 20" of 4" dia PVC pipe, using #22 ga wire.  The RMS DC current was
only 
3.5A, and the Rchoke was 10 ohms -- about 100W of dissapation.  

What's wrong with that?! 

As it turns out, when I simulated the _entire_ primary circuit in PSPICE 
I discovered that the RMS current in the choke was actually 14A instead
of 3.5A, 
due to a large unforseen oscillation between Cpri and components of the
filter 
network.  That oscillation circuit path was then "de-Qed" using a diode
and a 
resistor, first in PSPICE and then in the actual circuit.  The choke
runs
only slightly above ambient temp. now.

IMO, theory is much more accurate at predicting failure than success. 
Virtually
all designs that are theoretically impossible will ultimately fail, so I
find
PSPICE more useful as a litmus test for new circuit designs. 

-GL