[GreenKeys] Model 28 signal lines

Wed Oct 16 22:48:10 EDT 2019

Greetings --

I've been looking at selector magnet current as a function of time the last
few days, mostly inspired by this article (
http://www.navy-radio.com/manuals/tty/570-224-400.pdf ).  And Micro-Cap
circuit simulation software is now free (
http://www.spectrum-soft.com/download/download.shtm ) which is a pretty
piece of serious software.

I've been intrigued by this topic because when I first started with the
Model 15 in about 1982 I used my own constant current driver fed from about
24 volts to run the TTY from a computer.  It seemed to work well (meaning
only it printed well).  I didn't do what I should have done and that was
put a scope on the loop to measure magnet current as a function of digital
signal input, to see how much of a delay I was introducing in getting the
magnets up to 60mA.

When I later tried that electronic constant current driver on a friend's
28, it didn't work at all.  When I mentioned that to Bill Henry of HAL he
told me that the selector magnets on the 28 had much more inductance than
the 15, the 28 magnets in standard configuration (parallel) being about 2H
total and that it really did need a high voltage loop in order to use a
large current limiting resistor, so that the circuit could have a large R
to L ratio (for fast rise time in the presence of a large inductance).  As
Jim pointed out, the 15 may have been more tolerant of timing issues for
other reasons than just magnet inductance.

I recently measured the inductance and resistance of my 28RO magnets (with
the armature in an unknown state - the position of the armature may affect
the inductance value a bit) as being 1.73 Henrys and 64 ohms.  Consider
those to have an accuracy of about 5% (just reading things off the digital
scope).

I'm not up to speed on Micro-Cap yet, so I'm stuck with hand calculations.
Tonight I came up with the following numbers.

For a conventional loop of 170VDC and 2800 ohms with selector magnets of
1.73 H and 64 ohms resistance (selector magnets in parallel) that gives a
long-term stead state current of about 60 mA.  After closing the loop
switch to allow current to flow, the current in the magnets starts at zero
of course and then rises with exponential decay and reaches 55 mA after
1.5ms (and is now very slowly rising, but it's almost at 60mA).
Theoretically, it never reaches the final value of 60 mA, but that's in the
realm of how many angels can dance on the head of a pin.  When the current
is within 10% of the final value I call it done.

Consider now an electronic driver loop circuit that consists of a 32 VDC
source, a 125 ma fast blow fuse (5 ohm), the resistance of the magnets (64
ohms), the resistance of a MOSFET power switch in the ON state (less than 1
ohm) and a 10 ohm current sense resistor.  The final long-term steady state
current would be no less than 32/(5+64+1+10) = 400mA.  Which is not good
for selector magnets, hence the old-school fast blow fuse if something goes
wrong.  But the plan is to use an electronic circuit to disable that MOSFET
when the current gets to 60mA and then a secondary MOSFET with a 470 ohm
resistor holds the loop at 60 mA long-term.  So, in considering when MOSFET
1 is first turns on, the current in the magnets is at 0 mA and starts to
rise with an exponential decay proportional to R/L, with R being only 80
ohms in this case.  Running the numbers using the values above shows that
the magnet current hits 60mA in 3.5 mS (again within 5%) and would keep
rising if MOSFET one is not disabled at that point.  So the low voltage low
resistance loop does introduce some additional delay compared with the HV
loop, 2 ms more, which isn't a whole lot.  I'm inclined to think that's
acceptable, but I'd need to test that to confirm.

The reason I like the 32VDC source is the availability of $15 HVAC
transformers that convert 120 to an isolated 24 VAC at about 40VA (meaning
about 1.5 A is available on the secondary).  Put that 24 VAC into a full
wave rectifier and one gets about 32VDC across the filter cap (with the
amount of ripple depending on current draw).

Another thing to consider is that once the armature is pulled in (I don't
know how long that takes after the mag current hits 60 mA) it may be
possible to reduce the magnet current to something like 36 mA (60% of 60)
because the holding current of most relay-like structures is much less than
the pull in current.  The only reason to look at this is to reduce heat
dissipation from the loop circuitry (meaning the 470 ohm resistor which in
the Marking state would normally dissipate a little less than 2 watts).
It's probably not worth the extra circuitry to go down that path to achieve
a slight reduction in heat disipation.

If anyone would like a copy of my calculations or a PDF of the circuit I
hope to start testing next month (will be placing a Mouser order this
weekend for parts) just let me know.  I'd be happy to provide them.

Please note that I could be all wet on this low voltage electronic magnet
driver, but it's the type of stuff I like to play with.

73, Paul Newland, ad7i

On Wed, Oct 16, 2019 at 9:32 PM Jim Haynes <jhhaynes at earthlink.net> wrote:

> On Wed, 16 Oct 2019, Paul Heller wrote:
> > Barry,
> > My only point, and others may disagree, is that 48v is not enough.
>
> Oh, I was going to say that too.  Although with perfect signals like
> you get off itty it might be OK.  Also there was a constant-current
> circuit in QST a long time ago and I believe the article said it was
> good to use with 48V.  I keep wishing someone who knows how would
> simulate that circuit and see what kind of rise and fall times can
> be achieved versus the simple power supply and resistor
> loop.______________________________________________________________
> GreenKeys mailing list
> Home: http://mailman.qth.net/mailman/listinfo/greenkeys
> Help: http://mailman.qth.net/mmfaq.htm
> Post: mailto:GreenKeys at mailman.qth.net
>
> 2002-to-present greenkeys archive:
> http://mailman.qth.net/pipermail/greenkeys/
> 1998-to-2001 <http://mailman.qth.net/pipermail/greenkeys/1998-to-2001>
> greenkeys archive: http://mailman.qth.net/archive/greenkeys/greenkeys.html
> Randy Guttery's 2001-to-2009 GreenKeys Search Tool:
> http://comcents.com/tty/greenkeyssearch.html
>
> This list hosted by: http://www.qsl.net
> Please help support this email list: http://www.qsl.net/donate.html
> Message delivered to ad7i at ad7i.net
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mailman.qth.net/pipermail/greenkeys/attachments/20191016/ae312ff1/attachment.html>