Re: [Amps] Magnetic shielding 2

[Borislav Trifonov <bdt@shaw.ca>]

Thanks for the replies.

The steel looks like grain oriented silicon steel, which I thought would 
be around 15; with the formula Will provided, I'm calculating about 
19...  I checked the secondary waveform on my scope through a voltage 
divider and I have attached a sketch of the trace (no load except the 
divider, which was several MOhm).  Pretty bad.  I only added 15 extra 
turns to the primary, when I should have added 45.  Worse, to fit more 
turns I'd have to rewind the primary with a smaller gauge (and that also 
is going to need 700 turns more on the secondary to recover lost 
voltage).  Given that I used nail polish to put the laminations back 
together (got no shellac or epoxy), I'll have to soak the whole thing in 
solvent to take apart again.

It is definitely the leakage due to saturation that's causing the case 
hum.  The hum increases as I lower the transformer into the chassis 
regardless of physical contact.  Even if I bring a scredriver near the 
core, I can feel the 60 Hz in my hand.

Now, besides the noise issue, how is performance affected?  My 
application feeds the output into a bridge rectifier, then a large CLCRC 
filter.  Will I be getting significant power reductions when driving 
this load?  Will the leakage amount change when connected to such a load?

If performance doesn't suffer too much, perhaps I could just reinforce 
the chassis and pot the transformer.  I'm thinking that potting would 
make the pot vibrate instead of the chassis, and then if I fill it with 
epoxy, that should physically reinforce the pot from vibrating.  Does 
that sound like a good solution?  I guess I could find a bigger 
transformer, but I've not found anything large with 2 kV secondaries 
that is not very expensive.  Not to mention that I'll have to get a 
bigger case, and throw out the $60 I spent on this one, and hours of 
drilling...
_______________________________________________
Amps mailing list
Amps@contesting.com
http://lists.contesting.com/mailman/listinfo/amps