On Thu, 7 Oct 2004, Brad Parker wrote:
> Tom Jennings wrote:
<snip>
If it leaves the x & y in state where the
drivers are turned on fully,
the transistors get hot.
What I meant was, if you find "Z" off (eg. nothing to be
displayed) then you can put the beam in the center of the screen
and save the transistors, within your adapter.
There is a problem with simply using the Z axis to switch the inputs to the
deflection amps and center the beam.
If drawing a vector starts far from the 0,0 origin, the X and Y deflection
signals
will be ramping up to position the beam prior to the Z axis signal
unblanking the
beam. So in this case the deflection amps would be presented with a very
fast
'jump' while the Z axis signal had unblanked the beam.
This is not a good thing.
I looked at beefing up the deflection amps to allow keeping the beam at 0,0,
but
this is totally impractical. The power supply for the monitor is not able
to supply
enough power, and the deflection yolk would burn up quickly.
This is a very bad thing.
Fortunetly I think I have a workable plan:
The HP 1350 has two signals on the AUX connector that may save my bacon.
The first is an active low signal telling external equipment that the vector
generator is busy. I plan on monitoring this signal to detect the end of a
display update cycle.
Once a display update cycle is completes, I'll activate some analog switches
that will ground the X and Y axis inputs to the vector monitor, and also
assert a signal that holds off vector generation until a timer expires. At
that time I'll switch the HP 1350 analog vector data back to the monitor,
then drop the vector generation hold signal and wait for end of display
update once again.
So a small microcontroller will control the refresh rate, and give the
monitor plenty of time for the deflection amps to cool down between display
refresh cycles.
I had already thought of a simple state machine that would do something
similar to the Z axis signal in case deflection power to the monitor was
lost while a Z axis signal was still present. This will cause major screen
burn on this type of monitor because the high voltage supply remains high
for some time after power-off, but the spot-killer circuit drops out as soon
as deflection power is lost. In my application the monitor is not in the
same rack as the HP 1350, so a power sequence problem is quite possible.
So now I think I'll move all these functiuons into a small microcontroller,
and kill both problems with one solution.
The only down side to this is that this approach will only work with the HP
1350, my Imlac PDS-1 does not have these external handshake signals. But
thats ok, I have a new-old-stock Tektronics 611 storage tube I can use for
the Imlac. Imlacs could be ordered from the factory with Tek 611's in place
of the stock vector display, and its a 'period' device, unlike the WG
19V2000.
