B7971 x8 Display - The Movie
Cogwheel Circuitworks
http://cogwheelcircuitworks.com/projects/b7971x8-display-demo/
a sharp observer noted the following:
> > In a message dated 5/10/2011 11:40:57 A.M. Pacific Daylight Time,
> >
> > webmas...@jb-electronics.de writes:
> >
> > But I am curious: Did anyone else notice the weird spots showing up
> > between different fading sequences? It almost looks as if your driver
> > does not work properly sometimes.
> > Jens
> >
This is the famous ghosting issue with multiplexed designs; If there is a
voltage potential over 165V between any two elements in the tube, you
might see
this effect even if the current is very tiny.
This is addressed with "Pull Middle" circuits which are just pairs of
resistors
tied to HV and ground. If HV is say, 220V, the voltage between the two
resistors should be 110V, well below ionization of big tubes voltage and
therefore no
ghosting *should* occur. This is then supplied to cathodes and anodes
when they
are off or floated.
If you look at my schematic page 4, you will see these pull-middle circuits.
There is however, a note to omit R26 which is one half of the Anode driver's
pull-middle circuit.
http://cogwheelcircuitworks.com/wp-content/downloads/NixieDrvrIN17x7-RevA1-Rel1_Schematic.pdf
This was done because the resistor was overheating. There was no issue with
ghosting on smaller tubes; But with the B7971s the HVPS needed to be run at
higher currents. This meant the unloaded voltage of the HVPS would get high
enough that leakage current through the HV513s FETs was still sufficient to
reverse light the tube (Cathode @ higher voltage than anode). This was
partially mitigated by turning off the HVPS just for the period that a
tube was
unlit in software. Still, you can see some artifacts during fadeup/fadedown.
The video was taken without R26 in place, but I have installed a larger
wattage
resistor in its place and the ghosting has virtually been eliminated.
That said, it is a nice artifact.
There's other little imperfections you can see; like tubes getting briefly
switched on when they are supposed to. The effect is a single tube
flikering on
and off during transitions.
I had actually thought of adding a "dying neon tube" effect where one or
more
tubes would get deliberately flickered like a dying neon sign. Didn't get
around to it. Maybe someone else who likes to code can do that.
..c