In message <3.0.5.16.19990505152140.28bf6522@vaxxine.com>, Phil T.
(VE3OZZ) <phil@vaxxine.com> writes
>
>
>Well, you're one up on me at this point Steve. I haven't measured some of
>the parameters to the degree you have, nor do I have SPICE for modelling.
>My comments are based on somewhat theoretical calulations from a couple of
>the few available (but well known in their field) books on transformer
>design. One is by Reuben Lee and the other is by W. Flanagan. Your model
>for the doubler "seems" to predict peak secondary currents that "look" too
>high. Even the peak current for the FWB is alot higher than I'd expect,
>except you've measured real values that agree fairly well. I'll play
>around abit to see if I can verify one way or 'tother.
>
>Thanks...Phil
I'll try to find time to re-measure too. The Spice numbers stack up with
data from the 5th ed. RSGB handbook from 20 yrs ago, but that's far from
a reliable guaantee!
Double checking the bridge system on paper with slight simplifications:
>From the peak of the ac input, the cap discharges until diode conduction
starts again. For ease assume that the discharge is linear so CxV = Ixt,
where V is the voltage drop as the cap. discharges. Then V=Ixt/C so
for 0.5A, 30uF and .01secs (50Hz mains and full wave) the drop is 167V.
For a 1500V secondary, peak voltage is 2121V so the capacitor voltage
droops to 1954V until diode conduction starts again. The ac waveform
hits 1954V at 67 degrees of the sinewave (0 degrees = 0V) so diode
conduction occurs for 23 degrees from 67 to 90 degrees. This is 1.28ms,
or 12.8% of the ac waveform period - call it 1/8th for ease. The AVERAGE
current during this diode conduction period has to be 8x the dc current
and it's not a rectangular current pulse so the peak current will be
higher. In practice the peak is lower and the conduction period is a bit
longer because (I guess) of the time delay/low pass effect of the
transformer source resistance and the smoothing cap. [After writing the
message I had a thought and set the source resistance to 1 ohm in the
Spice version, the peak current goes to 16x dc current and the shape of
the current pulse changes to something close to a right angle triangle,
which gives an average of 8x. Yes, really, not fixed!]
If I get round to setting it up on the bench, I'll get rms numbers too
as my 'scope calculates them and saves me the pain of doing it.
If my numbers are correct, it helps explain why you need to shell out on
top grade caps for doubler systems.
Steve
PS I don't know where you to go to get it, but I'm using PSpice which is
freeware.
PPS I've no axe to grind here - According to Spice using the figures I
guessed at, the doubler does not deliver significantly lower ripple and
puts everything but the transformer insulation under greater strain.
--
FAQ on WWW: http://www.contesting.com/ampfaq.html
Submissions: amps@contesting.com
Administrative requests: amps-REQUEST@contesting.com
Problems: owner-amps@contesting.com
Search: http://www.contesting.com/km9p/search.htm
|