[GreenKeys] Power supply for Eric's demodulator

[epvgk at limpoc.com]

On Tue, Jan 04, 2011 at 09:34:23AM -0800, John Nagle wrote:
> On 1/4/2011 8:32 AM, greenkeys-request at mailman.qth.net wrote:
> > Subject: [GreenKeys] Power supply for Eric's demodulator
> 
> Anybody interested in building a more modern power supply?
> Take a look at
> 
> 	http://www.aetherltd.com/public/flashsupply6.asc
> 
> This is an LTSpice file of a circuit simulation.  You
> need LTspice
> 
> http://www.linear.com/designtools/software/
> 
> to run it.
> 
> This is a proof of concept simulation for a modern 60mA
> supply for a Model 15.  If built, it should be about a
> 1" square board running off 5V at 250mA.  I was looking into
> integrating this with a USB interface, and powering it from
> the USB port.
> 
> It may seem difficult to get 120V at 60mA from 5V at 250mA,
> but it's not.  The circuit assumes the selector magnet is 55 ohms,
> 4 henries.  To get 60mA through that, 120V is required at start,
> but once the inductance has been overcome, which takes about 3ms,
> the sustain voltage is only 3.3V.  (E=I*R, for 60mA * 55 ohms).
> 
> The circuit uses an LT-3484-1 part, which is a switching power
> supply controller for photoflash applications.  During SPACE
> periods, using a tiny surface-mount pulse transformer, it
> charges up a 1uF cap to 120V.  On a SPACE to MARK transition,
> the capacitor is dumped into the selector magnet, until it
> discharges and the fixed 3.3V supply (the LT-3080) takes over to
> sustain 60mA.  The waveforms look about the same as they do
> with a classic 120V fixed supply through a 2K 10W resistor.
> 
> In the usual Teletype power supplies, with a 2K 10W ballast
> resistor, 95% of the energy is going into heating the ballast
> resistor.  With a pulse design, far less energy is needed.
> 
> It requires surface mount parts and is an all-tiny-part design;
> the transformer is 6mm x 6mm x 3mm high.
> I did this as an exercise in low-power design, just to see
> if it was possible to power a 60mA current loop device from
> a USB port.  I haven't actually built one.  This requires
> making up a custom board; layout for switchers is critical.
> But it's clearly possible; the parts are being used at far below
> the ratings needed to fire a photoflash tube.
> 
> Personally, I've been using 120VDC 200mA open-frame
> linear regulated supplies, which I got from HSC Surplus in Santa
> Clara, CA.
> 
> 				John Nagle
> 
> 

Hi John, 

I ran across your idea for a power supply that only needed to supply a 
high voltage pulse initially to overcome the coil's inductance - I like
the idea, it seems much more sensible than dumping 5+ watts into a hot
resistor all the time. Even though the resistor is rated for it, having 
it that hot makes me nervous.  I wish I had time to try this one, maybe
some day... Have you attempted a layout for it yet? 

eric