[GreenKeys] GreenKeys Digest, Vol 140, Issue 12

[John Nagle]

      I've looked at that in LTSpice.  I have all Model 14 and 15
machines, none of the newer stuff, and it's all 60mA gear.
I did a simulated design for a 60mA 120V current loop supply that would 
run from the +5V of a USB port.  You really need 120V only for about the 
first 2ms after a space to mark transition.  The sustain voltage is much 
lower, only 3.3V (55 ohms times 0.060 amp).  So one way to do this is to 
have a low-current DC-DC supply generate 120V to charge up a capacitor 
to 120V during SPACE. Then, on a MARK transition, connect
the capacitor to the selector magnet.  This pulls in the magnet.
Use diodes to provide the small sustain current to hold the magnet
once the capacitor has discharged.  I didn't build it, though,
because it needed surface mount parts.    The power needed from
the USB port is about 250mA max, which is half of a standard
USB power budget.  That's how you'd do it with modern electronics
technology.  You'd probably want transformer and optoisolation
between the 120V side and the USB port, to prevent spikes from
damaging the computer hardware.  You get about 400V at a MARK
to SPACE transition when the selector magnet dumps, and that
has to be both snubbed and isolated.

     This is basically what the driver board in a Model 33 does,
although they weren't worried about making it a low-power device,
as one would today.

				John Nagle

> Date: Tue, 15 Sep 2015 13:08:46 -0500
> From: Jeffrey D Angus<jdangus at att.net>
> Subject: Re: [GreenKeys] Of current loops and voltages.. (M15 20mA /
> 	60mA)
>
...
> Back to getting coils to move mechanical parts fast.
> The whole point of the 120 VDC loop supply, with it's attendant 2K ohm
> or so total resistance, is to provide enough voltage compliance to force a
> fast rise time through the selector magnets.
>
> You can see how this works by playing with a circuit simulator like LTSpice
> and monitoring the current rise time through a 4 Hy inductor and varying
> the square wave source driving it from 12 vdc @ 200 ohms to 120 vdc @
> 2000 ohms. Look at the time delay required to rise to 80% or 48 Ma.
>
> Most telephone exchanges had no shortage of 48 vdc to run things. It
> should be obvious that they went through to additional expense of using
> 120 vdc for good reason.