[VoIP] Asterisk and Payphones

[Mark Rudholm]

Martin Harriss wrote:
> Mark Rudholm wrote:
>> Well, I'm glad to hear there is significant interest in this.
>>
>> For the moment, I'm working on coin-relay control.  My first
>> draft of the schematic is available here:
>> http://rudholm.com/coin-control.pdf (use "File, Rotate" to
>> fix the orientation in Acrobat Reader).
> 
> Mark,
> 
> Looks like there's problem in your schematic.  If you activate the 
> return/collect line, you will turn on both the LCC120 SSR's.  This will 
> short the 130v line.

As I drew it, the only way to short the 130v supply line would
be if you energized one of the LCC120 SSR's and not the other,
but that shouldn't be possible since they're energized by the
same control line.

I'm not sure how clear it is in the PDF unless you zoom really
far in, but the LCC120 chip is two SPST switches in one package.
One switch is normally-open and one is normally-closed.  I gang
those two internal switches together (by shorting pins 6 and
7) to turn each LCC120 into an SPDT switch.  Then I gang the two
LCC120's together (by tying their control lines together) to make
a DPDT relay.  So as long as they both energize and de-energize
together, there should be no unwanted shorts.

Granted, there may be a few milliseconds of overlap when the
relays are changing states, but the data sheets show the transitions
as being about 5ms worst-case scenario, which is fine, especially
considering that R1 limits the load to 120mA even if something
downstream becomes a short.  That 1:1 isolation transformer is
130mA unit and the SSR's can all handle 170mA, so even a serious
mistake shouldn't set any fires :)

> Also, I think I see what you're trying to do with the strobe line.  Why 
> not just have the processor (Asterisk, or whatever) turn the line on and 
> off to do the strobing?  You may not even need that strobe once you get 
> the other part of the circuit sorted out.

Well, the reason for doing the timing in hardware is that I don't
trust the PC or the software to not set the strobe line high and
leave it there, which would leave my coin-relay energized all day,
possibly damaging it (I don't have the relevant BSP, but I doubt the
coin-relay would be happy with a 100% duty-cycle).  So yeah, you're
absolutely right, I could do the timing in software, but I trust
C2 more than untested PERL or shell script code :-)

As far as one control line goes, I thought about just sending the
control voltage pulse out whenever the return/collect line changed
states, but increases my part count, and it means that you couldn't
do "collect, collect, collect" (which ACTS does when you're on a
non-local call).  I suppose I could do something with a tri-state
output that floats the line when it's "neutral" but <shrug>, adding
that strobe line seemed easier and less uncertain.

Thanks very much for the feedback.

-Mark