Chronology Current Month Current Thread Current Date
[Year List] [Month List (current year)] [Date Index] [Thread Index] [Thread Prev] [Thread Next] [Date Prev] [Date Next]

Re: transformers



After mentioning DC-DC converters I ended up being in class or in meetings
all day and was unable to explain what I meant by that.

First, a couple diversions to explain what I was not talking about. There
is indeed a thing called a "switching regulator" as William Beaty mentioned.
However, I don't think his circuit description described a switching supply.
Rather, I think the proper name for the regulator Beaty described (that uses
the collapse of the magnetic field in an inductor) is "flyback supply."
Flyback supplies are common in TVs and other CRT circuitry to develop the
high-voltage for accelerating the electrons.

Switching supplies are different and are common in computer power supplies.
The typical "switching regulator" steps down voltage from one DC level to a
lower one. It works by having an unregulated DC voltage considerably higher
than the final regulated voltage desired, and current goes to the load
through a series transistor to regulate the voltage. That's not unusual.
But when a series transistor regulates voltage from a higher DC level to a
lower DC level it often runs in a "partially on" state. This means it has
both current through it and potential difference across it. That means it
dissipates considerable heat. In a switching supply the series transistor
is either totally off (big potential difference but no current) or totally
on (big current but small potential difference) so it never generates much
heat. To get the proper output voltage the duty cycle is varied. If the
load is drawing significant current the transistor spends more time in the
on state. If the load requires less current the transistor can spend more
time in the off state. So it switches between on and off at a high
frequency (perhaps 20kHZ) and regulates by varying the duty cycle.

So, what did I mean by transformerless DC-DC converter?

Remember, integrated circuits are inexpensive and we can dream up some
pretty complicated circuits that can be mass produced. Suppose you had
sufficient transistors acting as switches and oscillators to do the
following... (1) Charge two capacitors in parallel from a 5V supply. (2)
Disconnect these from the supply and reconnect them in series... voila...
10V. Hook these to another capacitor to charge it up to some voltage. (3)
Go back to step one, and repeat this over and over. The final capacitor
will charge to near 10V. This is sometimes called a "flying capacitor"
circuit.

You can also generate reverse voltage this way. Have a 5-volt supply and
need +and- 5V? Charge the capacitor... dump it to another capacitor, but
reverse the polarity... then charge it back up... etc. You can design
fairly simple circuitry to charge multiple capacitors in parallel then
connect them in series, and back again. You can get just about any voltage
(plus or minus) by doing this.

I can see why one might want to call this a "switching supply" because we do
a lot of switching to toggle capacitors from being charged in one
configuration then discharged in a different configuration. But these are
not called switching supplies because that name was already used by the
different circuit explained above.

Larry Smith mentioned the Meade Power supply. It truly does "step up" the
12V from a car battery to 18V to run the telescope without using a
transformer. I have one sitting on my desk right now with the cover off. I
didn't open it out of curiosity; I opened it because we fried it a couple
months ago. Although it is pretty blackened inside, and I decided it is not
repairable, it looks like a simple flyback design... IC, inductor, diode,
couple capacitors.


Michael D. Edmiston, Ph.D. Phone/voice-mail: 419-358-3270
Professor of Chemistry & Physics FAX: 419-358-3323
Chairman, Science Department E-Mail edmiston@bluffton.edu
Bluffton College
280 West College Avenue
Bluffton, OH 45817