Transcript Understanding the Physics of Plasma Display

Understanding the Physics of
Plasma Display Addressing
Vladimir Nagorny
(available at www.plasmadynamics.com)
Addressing Speed and Cost


Cost is the name of the game
today
Single-scan
Frame - 16.7ms ≥10 Subfields,
~1000 lines)

Reset ~ 1.5-3ms

Sustain > 5-6ms, or it is dim

Address ~ 7.7-10.2ms
0.75-1ms/line for single
scan HDTV addressing
Three periods - three components




Each period has its purpose, its specifics (priming, wall charges),
requirements (stability, efficiency, speed) and problems.
The purpose of the setup period is to place all cells into
identical and known state. Usually it is in the corner of the Vtcurve, where either PG or both PG and SG discharges are active
The purpose of address period is to do the opposite -to
discriminate between ON and OFF cells, so that one can start
sustaining of the ON and not the OFF cells
The purpose of sustaining period are to produce picture (light)
and provide priming
Starting point for addressing
Reset
Period

Vb
X

Y
VA
V AY
AY
V
XY
(sustain gap
discharge)

A
AX
V XY
XA
Assume that at the end of the reset period
both SG and PG or at least PG discharges are
active and ramp is stable. Cell is in the
corner of Vt-curve – breakdown boundary for
3 electrodes, which can be verified by
measuring the Vt-curve. After that every cell
is “locked”.
For addressing one unlocks line by line and
addresses it. Unlocking the line for
addressing brings every cell again in the
corner of the Vt-curve (or to the edge), and
extra voltage initiates the discharge.
Strong SG discharge transfers large charge
between X and Y electrodes, which then
assist sustaining.
Addressing vs. discharge “speed”

Voltage
Current
Density
Addressing “speed” is the
time T, required to reliably
address the line – maximum of
individual discharges times
T   stat   form   decay

statistical
formative
decay


 stat
 form
- statistical delay
- formative delay
 decay - plasma decay time
Where statistical effects come from?

Electron density is zero, they present in the volume only
occasionally, so
1. when the voltage is applied there may be no electron in the volume
2. when electron appear it may not start the discharge (#1)
V
1
2
3

Sources of electrons

Metastables (Volume source ##1, 2) – not important


~10-15ms after strong discharge they can only produce 1 e/ms per cell,
much less after the ramp setup – not enough for addressing
Exoemission (Surface source #3 - best kind) - the only source capable
of “working” for long time. Exoemission – almost uniform above
sustain electrodes, and absent elsewhere
Statistical effects (continue)
V
3'
3

When high voltage applied any
electron starting from the wall
(#3, #3’) will start the discharge
 the only way to decrease
statistical delay is to increase
exo-emission rate (by doping, …).
Right?
Statistical effects (continue)
V
3'

3
When high voltage applied any
electron starting from the wall
(#3, #3’) will start the discharge
 the only way to decrease
statistical delay is to increase
exo-emission rate (doping, …).
Right?

Wrong

It’s NOT the only way
(where the OTHER come
from?)
Statistical effects – OTHER ways

V
W all
W all
3'
3

Electrons diffuse and may end up on the
sidewalls. These losses affect both
statistical and formative delays. The
closer walls, the more electrons (and
ions) are lost, more need to start
discharge. 3 and 3’ are not equivalent
any more
stat =stat(stat 0, L1, L2, E) >>stat0
Statistical effects – OTHER ways

V
W all
W all
3'
3
V
Electrons diffuse and may end up on the
sidewalls. These losses affect both
statistical and formative delays. The
closer walls, the more electrons (and
ions) are lost, more need to start
discharge. 3 and 3’ are not equivalent
any more

stat =stat(stat 0, L1, L2, E) >>stat0

We can:


Increase E (Compensate losses by
increasing e/i production).
E -is
the strongest factor
Concentrate source in the center

All of it

W all
W all
3'
Move walls (Decrease losses)
3
Statistical effects – PDP cell
PDP has physical walls (barrier ribs) and
virtual walls (shown). Both affect
efficiency of exoelectrons to start the
discharge  increase statistical delay
(up to 10 times). Also, as discharge
initiated in zone A feeds zones B and
C it may change the mode, as the
region C is unstable
Va=80V (3D PIC/MC)
Emission from the edge does not produce
self-sustaining discharge - it decays.
Electron emitted in the center produced
~4.5 times stronger avalanche than
from the edge, and then number of
particles increases with characteristic
time ~ 43ns (this case) by itself.
Statistical effects – sliding discharge
g=0
E
In a strong electric field, electron diffusion across electric field
may actually result in the avalanche sliding along the surface. In
a weaker field avalanche ends up close to the point of the point
where electron trajectory crosses the surface
Statistical effects – PDP cell (cont.)
Virtual walls can be moved away by
a)
decreasing potential of the “left”
(bus) electrode and
b)
Increasing the voltage across PG
(between address and both sustain
electrodes) – strong effect.
Higher E (or Va)  wider zone A of
certain discharge initiation (if
ionization is strong, it
overcompensates losses; even
electron near the wall will start the
discharge)
c)
Decreasing aspect ratio PG/SG
Formative delay


Formative delay time is defined by the slowest - linear stage of the
discharge. Both applied voltage and gap size strongly affect form
D(a, L)= g (eaL - 1) –1 - the most important parameter, it varies. D > 0
discharge grows, D < 0 – decays; D > 1 – charge accumulates in the volume
1
 form   ,
Weak discharge (D <<1)
  (U  Ubr ) / L2
No field distortion, plasma
decays as fast as it grows
mi U
vi
~ D~ 2 D
L
L
Strong discharge (D >>1), linear phase
  U a ( E / p) / L
Completely compensates
field, plasma decays very slow
a, g – first and second Townsend coefficients,
U – applied voltage, L – gap size, E – electric field, p – gas pressure
Formative delay (3D cell)

form is affected by wall losses (a  a*, E,..), just like stat  same
recipe (larger E, better configuration)

Nonlinear (from D(j)) and related 3D effects
V
br
V
D(j)
j
The critical current at which
D starts to increase depends
on gap size j~1/L2
Current density
The range of voltages with linear
behavior is wider for shorter gap,
but discharge is much faster
Discharge development scenarios
Slow scenario (common) – losses to Y
significant
1)
2)
3)
Both XA and XY grow as et
XY reaches nonlinear stage
Discharge transfers to XY; further
development is independent on VA
Discharge development scenarios
Slow scenario (common) – losses to Y
significant
1) Both XA and XY grow as et
2) XY reaches nonlinear stage
3) Discharge transfers to XY; further
development is independent on VA
Fast scenario – losses to Y low or
suppressed
1) XA grows as e’t, ’>
2) XA reaches nonlinear stage,
strong XA discharge
3) Field bifurcates and discharge
switches to XA-Y; further
development is independent on VA
Discharge development scenarios
Slow scenario (common) – losses to Y
significant
1) Both XA and XY grow as et
2) XY reaches nonlinear stage
3) Discharge transfers to XY; further
development is independent on VA
Fast scenario – losses to Y low or
suppressed
1) XA grows as e’t, ’>
2) XA reaches nonlinear stage,
strong XA discharge
3) Field bifurcates and discharge
switches to XA-Y; further
development is independent on VA
One can suppress losses by decreasing Y potential, increasing A or choosing
geometry, in which X and Y electrically not connected SG>PG (Figure)
Plasma decay

Plasma decays primarily due to recombination:
Xe  Xe  M ( Ne, Xe)  Xe2  M
Xe2  e  Xe2*  2 Xe  

It is slow process, and there is not much one can do when the
plasma density is high. However, this is not really a problem since
the memory charge will then be determined by the final voltages of
sustain electrodes, and ON and OFF cell will still be significantly
different.
Control knobs of the addressing speed





Exoemission rate – affects the statistical phase (priming)
Plate gap voltage (VAX) and size affect both statistical and formative delays.
Higher voltage and shorter plate gap (t ~L-2) – better discharge time
Configuration of electric field and voltage of the Y-electrode (VY ) –
whether it suppresses or induce the electron transport to Y electrode
affects addressing. Stronger communication between X and Y electrodes –
worse addressing time
One may increase the voltage VA, by applying unselectively VOFF to all cells
of the line (between A and both X and Y), and using addressing drivers with
only selected cells. If communication between X and Y is suppressed, VOFF
and VA may be quite large.
Depending on geometry, one may use additional voltage applied to Y
electrode, to either assist in formation the SG (XY) channel or suppress XY
communication (and premature XY discharge)
Dual Addressing
Vaddr=VON - VOFF
In addition to VOFF bias, one may
decrease the voltage between
sustain electrodes (VXlock) in order to
better separate SG and PG
discharges (obstruct the spread of
the discharge across SG) and
maximize VOFF.
[12]
V AY
VA
[Sakita, et al.]
AY
AX
XY
V XY
XA
Addressing: Short PG, no XY communication
1. End of the Ramp Setup
3. After OFF discharge (VOFF=50V)
2. Address voltage (60V) is applied
4. Beginning of the ON discharge
VON=60V, t=335ns
Parameters of the
OFF discharge
Parameter /Va
30V
40V
50V
60V
DVgap, V
62
70
102
Break (ON)
1/2 , ns
270
155
118
xx
T(5000), ns
895
810
747
xx
Ni,max(105)
6.55
15
27.9
xx
Ttot, ns
770
680
472
xx
Summary



Understanding the mechanism of discharge development helps to
identify the nature of the problems and ways of solving them. Most
problems of address discharge are related to a low exoemission rate
and may be even more with inefficient utilization of those electrons.
Using simple modifications of addressing scheme in order to a) increase
addressing electric field by using two levels of addressing voltage (VOFF,
VON = VOFF + VS) and suppress “communication” between regions (VXlock)
and b) choosing smaller the plate, gap can easily cut addressing time in
half.
Exoemission rate of 20-40e/ms/cell and ther efficient use is sufficient
for addressing the line in about 0.5ms