Re: Noticed new vector monitor discussions, this is my input...

`10 amps is fine, but though 100 volts SOUNDS like a lot
(and is
fine for everything except the G08), to get the entire
picture you
have to see where the secondary breakdown limits are.

The 2N6259 used in the G08 had fantastic secondary breakdown

limits---it could handle full power with close to 100 volts
on it.

An example, the 2N3716 that's used on almost everything
except
the G08 is rated as 80 volts, 10 amps, and 150 watts. First
of all,
to handle 150 watts takes way more than mounting the part on
a
sheet metal chassis, but we'll let that slide for now.

Let's say we are driving the yoke at 3 amps. At 150 watts
the part
should be able to take 50 volts, right? WRONG! If you take
a
look at the data sheet at
 http://www.mot-sps.com/books/dl111/pdf/2n3715rev7.pdf
you will see it will take 50 volts for 1 MILLISECOND! At 40

volts it can run continuously (with a LOT of heatsinking).
In the
Wells-Gardner monitor you use 24 volt (or something like
that)
supplies so there's no problem, except for the inductance of
the
yoke giving you some real interesting voltage spikes.

But the G08 used 63 volt (or so, wasn't regulated)
supplies. The
transistors could easily see over 125 volts across them
during
switching. That's partly why I say Zanen has no business
selling
2N3716s for this monitor (80 volt part, remember?). The
other
part is what happens when you have your 3 amp yoke current.
IF the yoke current is being held steady (as it is in one
axis when
drawing a straight vertical or horizontal line with the
other axis) the
voltage drop across the yoke will be due to winding
resistance only.
Let's say that gives you a 10 volt drop in the yoke. 63-10
is 53
volts at 3 amps is 159 watts. The 2N3716 is good for that
for
about 700 MICROseconds. The 2N6259 could handle that for as

long as you could keep it cool. Electrohome used pretty
good
heatsinking, so it could handle that indefinitely. Where
they got in
trouble was 1) the 63 volts WASN'T 63 volts, was often more,

2) bad power supply connectors or other CPU lockups would
tell
the monitor to deflect the beam WAY off the tube face, and
3)
inductive kickbacks. Noone ever seemed to bother designing
a
board to clamp inductive kickbacks properly.

The 2N6259 was, if memory serves, good for over 10 amps, at
least 150 volts, and over 200 watts. Plus it didn't go into
secondary
breakdown until close to 100 volts! VERY few transistors
can do
that.

All right, I rambled again. Sorry for the parts that had
nothing to
do with anything.

James Nelson wrote:

> >How about boosting the voltage to the board?>The driver
> bias could be adjusted to compensate,>and getting hi-power
> transistors at 6 Amps and>200 Volts are possible now. I
> don't quite understand what the voltage increase is
> supposed to accomplish. You can try this right away by
> bypassing the voltage regulator transistors on your power
> supply. In fact, there is an evil tech document,
> http://www.gamearchive.com/video/manufacturer/atari/vector/monitors/wg_color/final_solution.txt,
> that someone actually reccomended this. (What were they
> thinking?) The power dumped across a transistor (or any
> device) is going to be precisely equal to the voltage
> across it times the current through it. The current, in a
> correctly working drive, at any given moment will be the
> same no matter what the voltage input is becuase the
> current is what is fed back. The conclusion to draw here
> is that you want the voltage to be as low as you can make
> it so that you are just above the threshold of degrading
> performance. Increasing the voltage would be desirable
> only if you couldn't get enought deflection or your draw
> speed was being limited (warped vector draws). In your
> case, where you are going to make a custom vector game
> from scratch, maybe higher voltage would be good, but I
> wouldn't do it ever unless the monitor couldn't keep up
> with the game.
> By the way, I think those existing transistors are 10 amp
> 100V devices. Good enough on paper.
Received on Fri Oct 22 12:34:00 1999