Note : Les descriptions sont présentées dans la langue officielle dans laquelle elles ont été soumises.
200604~
.. 1
PATENT
PD--188459
.
PRINTHEAD PERFORM~CE TUNING VIA
INK VISCOSI~Y ADJUS~ENT
~C~
This invention relates ~o ink-jet printers, specifi-
cally ~hermal ink-jet printers, and more particularly, to
'a structure for substantially improving the performance of
the nozzle~s) in a thermal ink-jet printhead. This im-
provement in st~bility and consistency of operation ex-
tends the firing frequency range of the nozzle and reduces
cross-talk, as well as desensitizes the exterior surfaces
of the jetting nozzles to th~ir w~ttability state.
The design principles des~ribed herein are not limit-
ed solely ~o th~rmal ink jet applications bu~ in fact are
. o~ value in the design of athermally excited inX-jet
prir.theads as well.
15This invention is of particular value in those cir-
cumstances of design in which other means of attaining the
above-mentioned operating benefits are not available
through geometry changes due to manu~acturing process
constraints.
B~ÇKGROUND ~RT
When designing printheads containing a plurality o~
inX-ejecting nozzles in a densely packed array, it is
necessary ~o p~ovide some means of isolating the dynamics
of any given nozzle from i~s neighbors, or else cross-talk
will occur between the noz21es as they ~ire droplets of
ink from elem~nts associated with the nozzles. This
cross-talk seriously degrades print quiality and hence any
providently designed ink-jet printhead must include some
Case 188459
2~)6~7
1 ~atures to accomplish decoupl:ing between the nozzles and
the common ink supply plenum so that the plenum does not
supply a cross-talk path between neighboring nozzles.
Furthex, when an ink-jet printhQad is called upon to
discharge ink drople~s a~ a very high rate, the motion o~
the meniscus present in each nozzle must be care~ully
controlled so as to prevent any oscillation or "ringing"
o~ the meniscus caused by re~ill dynamics from inter~ering
with the ejection of subsequently ~ired droplets. Ordi-
narily, the "settling time" required between firings setsa limit on the maximum repe~i~ion rate at which the
nozzle can operate. If an inX droplet is fired from a
Aozzle too soon after the previous firing, the ringing of
the meniscus modulates the quantity of ink in the second
droplet out. In the case where the meniscus has "over~
shot" its equilibrium position, a firing superimposed on
overshoot yields an unacceptably large ejected droplet.
The opposite is true i~ the firing is superimposed on an
undershoot condition: the ejected droplet is too small and
extremely ~ast (this is known as a spear drop). There-
~ore, in order to enhance the maximu~ printing rate of an
ink-jet printhsad, it is necessary to include in its
design some ~eans for reducing meniscus oscillation so as
to ~inimize the settling time between sequential firings
o~ any one nozzle.
In addition ~o cross-talk minimization, an important
ob~ectiv~ of prin~head design optimization is the control
o~ meniscus dynamics during refill. During the overshoot
ph~se of rsfill, the ~o~entum o~ the fluid which has
~lowed into the firing cha~ber carries the meniscus beyond
i~s equilibriu~ position. At that point where the compli~
ance o~ the meniscus has halted the fluid flow, the menis-
cus has bulged out of khe bore and appears briefly as a
spherical section or "igloo" o~ ink projecting out o~ the
nozzle. Within microseconds, it has retrac~ed itsel~ back
;~ into the nozzle bore under the in~luence of surface ten~
sion force~ which s~rive to ~inimize the surface area o~
the meniscus. Viscous losses which are caused by the
Case 188459
2~)01~147
; ..
1 ~tio~s o~ the fluid behind thl~ meniscus cause the seesaw
oscillation o~ the meniscus to decay with time and eventu;
ally halt.
As such the time response oS a nozzle during refill
S can be approximated by a damped second-order harmonic
oscillator in which the ~ass of fluid entrained wîthin the
nozzle, firing chamber and refill port "bounces" on the
compliance of the meniscus while viscous dissipation
gradually damps out the oscilla~ion. ( I~ will be noted
10 that none o~ the parameters involved - mass, compliance or
resistance - are constants in this system; this is a
linear approxi~ation . )
' During ~he brief time that the meniscus has overshot
its equilibrium position and is bulging out of the nozzle,
it is possible for the fluid in the bulge to spill out
onto the material surrounding the lip of the nozzle. This
spillage becomes very liXely if the angle defined by the
tangent ~o the meniscus bulge at the lip of the ori~ice
~quals or exceeds the wettin~ angla criterion for the
material rrOm which the nozzle plate has been manufac-
tured. If this happens, the meniscus will break free from
~he ~ip of the nozzle and the fluid bul~e will then spread
out across the nozzle plate. As the meniscus retracts
back into the bore, it reattaches itself to the edge of
the nozzle and in so doing pulls most but not all of the
~luid ba~k down the bore with it. A small and ~ery shal-
low puddle o~ ink is typically "stranded" in the immediate
vicinity o~ the nozzle a~ter each fsrinq and refill cycle.
At low frequency operating conditions - typically less
than about 1,500 Hz - ample time exists between ~irings
~or essentially all o~ this stranded ink to be wicXed back
up by the nozzles. However, at high ~requency operation -
typically greater than about 1,500 H~ and above a new
accumulation of puddled ink occurs at the nozzle lip.
This accumulation can al o occur at low frequencies if (1)
; the surface tension of ~he inX is sufficiently low, (2)
the exterior surface of ~he orifice plate is sufficiently
wettable, or t3) the bac~ pressure (de~ined as the abso-
Case ~88459
~! ~.i . ' ' ' ' '
21)13~ 7
1 lute value of the s~atic negative gauge pressure in the
inX plenum) is su~ficiently high (at least about -6 inch~s
~2)
This accumulated ink has a deleterious e~ect upon
print quality by capturing and deflecting the ejected ink
droplet during the phase o~ droplet ejection when the tail
is abo~lt to detach itsel~ rom the meniscus and follow the
head of the dxoplet away from the nozzle. This causes
breakoff to occur not ~rom ~le xetracted meniscus but
instead from a random point around the wetted periphery of
the nozzle; the drop is pulled of ~-axis in the direction
of t~e puddle. This direction error is integrated over
~he flight time of the droplet to result in a dot place~
ment position error on the print medium. Since these
errors are random in magnitude and direction, the result
is an unpredictable and serious degradation of print
quality. In so~e cases, the ink accumulation is severe
enough to completely block droplet ejection ~rom the
nozzle .
HQnce, any providsntly designed ink jet printhead
~ust include some features to minimize meniscus overshoot
and ~inimize the time required for the meniscus oscilla~
tions to dacay away, so that the precedin~ scenario
(referred to as "nozzle wet-out~) is avoided. It should
be noted that wet-out can be caused or exacerbated by
spray that breaks o~f the`-tail of the drop and rains back
down on the nozzle plate. This is worst for low ~iscosi-
ty, high velocity drops.
Traditionally, wet-out is prevented by ~aintaining a
sta~ic negative pressure, also known as back pressure,
throughout the i~k supply syste~ so as ~o de~ine an equi-
librium position for the meniscus which lies inside the
noz~le bore. Another method involves the use o anti-
wetting coatings applied to the area surrounding the
nozzle lip, which prevent ~eniscus breakoff during over-
shoot. Yet another me~hod is to increase the amount of
viscous damping present in the ink supply system, thereby
holding overshoot below the value required to initiate
Case 188459
,' ~.' ! ~' ' ~' . `, ' . . .... ..
~201)~0~17
. s
1 wet-out. Still another me~hod is to provide a contact-
line barrier that preven~s the puddle fro~ advancing out
past a certain radius.
~nti-wetting coatings are of limited utility in
S preventing wet-out during overshoot since their lifetimes
are typically shorter than that of the printheads to which
~hey have been applied, causing wet-out to reappear prior
to the completion of the printhead's service life. Fur-
ther~ore, it is difficult to sufficiently immobilize these
coatings so that wiping detritus from nozzles does not
~orce the coating down into the bores, wreaking ha~oc
irreversibly upon the prin~head.
It is often impossible in practice to draw down
overshoot via static negative bac~pressure, since this
backpressure acts to retard the refill of ink in the
nozzles between firings. Hence, sufficient backpressure
to prevent wet-out also compromises operating speed.
But a third op~ion, increasing the amoun~ of viscous
da~pin~ has been ~ound ~o be the most practical solution
to the wet-out problem. ~his is because it (1) lasts the
li~e of t~e pen, (2~ does not slow refill so much that the
firi.~g frequency limit is compromised and (3) damps rip-
ples and waves on the meniscus surface and hence makes the
process more stable.
Increasing the hydraulic resistance (via ink-channel
d~mensional changes) is not the most practical method of
increasing damping, at least in some situations, as print-
head geo~etries push the li~it of the smallest dimensions
attainable with a particular material and process. (There
! 30 is an ongoing need to scale down printhead geometries to
allow the firin~ of s~aller droplets, as is desirable when
printing very high quality tex~, high resolution graphics
or images containing gray levels or ~half~oning". At drop
volumes below 50 picoliters, nozzle wet-out due to insu~-
~icient damping becomes ~he dominant factor in degrading
print quality.) The feed channel dimensions of such
structures are already so minute that to include pinch
points (as lumped resistive ele~ents) in the feed channel
Case 188459
i ,., ~ .. . . . ... .
! ,' ~, . .. ` : .... . . . . ` `
~',; ,'' ~?, ' ,:: ,
r. ~,, .,, ,, ~ . . ,: ~
s 2001~
1 struct~re would exceed the aspect ratio li~its of the
resist ~ rom which the barrier structure containing
the feed channels is ~ormed. For DuPont ~Vacrel" film,
this aspect ratio is approximately 1:1. Henca, in Vacrel,
a feature desired to be free of Vacrel must be at least
0.001 inch wide if the basis thickness of the film is
0.001 inch, 0O002 inch wide for 0.002 inch thick film and
so on. ~ence, to be manu~acturable, some o~her means of
obtaining sufficient meniscus damping to prevent wet-out
10 must be included in the printhead structure.
Previous approaches to the problem of cross-talk,
or minimizing in~er-noz21e coupling, can be separated into
three classes: resistive, capaci~ive, and inertial. The
following is a brief discussion of each me~hod and a
15 critique of the typical embodiments of these methods.
Resistive decoupling (to hydraulically "decouple" the
nozzlss fro~ one another) uses the fluid friction present
in the ink feed channels as a means of dissipating the
~nergy content of the cross-talX surges, thereby prevent-
20 $ng the dyna~ics of any singl~ meniscl~s fro~ being strong-
ly fel~ by its nearest neighbors. In the prior art, this
is typically implemented by making the ink feed channels
lon~er or smaller in cross-section than the main supply
plenum. While these axe simpie solutions, ~hey have
25 s~veral drawbacks. First, such solutions rely upon fluid
motion to generate the pressure drops associated with the
anergy dlssipation; as such, they can only attenuate the
cross-talk surges, not completely block the~. Thus, some
cross-talk "leakagesn will always be presentO Second,,any
30 attempt to shut of~ cross-talk complétely by these methods
will necessarily restrict the refill rate of t~e nozzles,
thereby compro~ising the maximum rate at which the print-
head can print:. Third, the resistive decoupling tech-
niques as pract:iced in the prior art add to the inertia of
35 the fluid refill channel, which has serious implica~ions
~or the printhead performance (as will be explained at the
end of the inertial decoupling exposition which follows
shortly).
Case 188459
ZOOG04t7
,.;; 7
1 In capacitive decoupling, an extra hole is put in
the nozzle plate above that point where the inX feed
channel meets the ink supply plenum. Any pressure surges
in the ink feed channel are transformad into displacements
of the meniscus present i.n the extra hole (or "dummy
nozzle~). In this way, the ho:le acts as an isolator ~or
brief pressure pulses but does not interfere with re~ill
~low. The location, size and shape o~ the isolator hole
~ust b~ carePully chosen to derive the required degree of
10 decoupling without allowing the hole to e ject droplets o~
ink as i~ it were a nozzle. This ~ethod is extremePy
e~fective in preventing cross-talk (but can introduce
problems with nozzle meniscus dynamics, as will be dis-
cussed below).
In inertial decoupling, the feed channels are made as
long and sl~nder as possible, thereby maxi~izing the
inertial aspect of the fluid entrained within them. The
inertia of the fluid "clamps" its ability to respond to
crosstalk surges in proportion to ~he suddenness o~ the
surge and thereby inhibi~s the trans~ission of cross-talk
pulses into or out of the ink feed channel. While this
deco~pling scheme is used in the prior art, it requires
considerable area within the print head to implement,
making a compact structure imposslbla. Furthermore, since
the resistive component of a pipe having a rectangular
cross-section scales directly.with length and inversely
with the third power of the s~aller of the two cross-
section dimensions, the flow resistance can grow to an
unacceptable level, conpromising refill speed. ~ore
30 importantly, however, are the dynam~'c effects caused by: :
the coupling o~ this inertance to the compliance o~ the
nozzle meniscus, as will be discussed below. -
With regard to the pro~lem o~ meniscus dynamics,
there are apparently no solutions offered in the prior
art. Apparently, this is a problem that has only recently
; surfaced as printhead designs have been pushed to accommo-
date higher and higher repe~ition rates. Clearly, any
method used to decouple the dynamics of neighboring noz-
Case 188459
. :.,. - . , -
,., ~ . . . .
..... . . .
:. .
: ~ ~ ., . " . . .
:: . : . . .
:~: . . . .
, .:: -:: . . .. . .
l zles will also aid in damping out meniscus oscillations,
at least from a superficial consideration. In practice,
problems are experienced when trying to use the decouplin~
means as the oscillatory damping mean~. These problems can
b~ traced to the synergistic effects between the nozzle
meniscus and the fluid entrained within the ink feed
channel, as outlined below.
If resistive decoupling is attempted by reducing the
width of ~he entire ink feed channel, the inertia of the
fluid entrained within ~he feed channel increases. When
this inertia is coupled to the co~pliance of the meniscus
in the nozzle, it results in a lower re onant frequency of
oscilla~ion of the meniscus, which requires a longer
settling time between firings of the nozzle. The inertial
effect and the resistive effect are tied together, with
the net effect being that settling time cannot be reduced.
Capacitive ~ecoupling has been proven effective at
droplet ejection frequencies below that corresponding to
the resona~t frequency of the nozzle meniscus coupled to
the feed channel inertia. Howev~r, its implementation at
*requencies near meniscus resonance is also complicated by
int~ractiv~ effects. Specifically, the isolator orifice
acts as a low impedance shunt path for high frequency
surges. ~ence, the high frequency impedance of an ink
feed chann~l terminated at its plenum end with an isolator
orifice will be lower than an equivalent channel without
an isolator. This means that during the bubble growth
phase, blow-back flow away from the nozzle is increased by
the isolator orifice. This robs kinetic energy from the
droplet emerging from the nozzle, wh~ch results in smaller
droplet size and lower droplet velocities and thus lower
ejection effic:iency. During the bubble collapse ph~se,
the isolator orifice meniscus pumps fluid ~low back into
the refill chal~er, which excites a resonant mode in which
the two ~enisci trade fluid between themselves via the ink
; feed channel. Since these two menisci are for most prac-
~ical designs similar in si2e, and since they are effec-
tively "in series", the equivalent co~pliance of the
Case 188459
;' i~;.~ ' '.'. '." ' ' ';! ' ' . ' . . . .
2qD06~7
,,`, g
l coupled system is roughly half of that with only one
orifi~e in it. The two-ori~ice system will thus re~onate
at a higher frequency, which is a benefit from a settling
time point of view, but the energy stored in the resonat-
ing syste~ still needs to b~ dissipated and th~reforeconstrictive damping will be necessary in such an imple-
mentation. While the ef~ects of these resonances is
poorly understood at this ti~e, the efficiency decrease
~ay be severe enough to prevent the printhead from work-
ing.
It is clear tha~ wha~ is needed is a method of prin~-
ing that accomplishes bo~h (l) isolation of any given
nozzle from its neighbors and (2) reduced oscillation of
the meniscus during refill (to minimize interference with
the ejection of subsequently fired droplets. This method
must do the above while not introducing any adverse side
ef~ects.
~I~CLQSU~_QF TR~ I~VE~TION
In accordance with the present invention, the viscos-
ity ~f the jetted fluid is adjusted to control the quanti-
ty of damping presen~ in the fluid supply channels or
refill ports of ~he ink jet printhead. Since any ~iscosi-
ty increase acts to increase viscous damping presentthroughout the ink supply circuit, the feed channel
dimensions in the supply circuit may be increased in order
to pr~vent excessive pressure drops within the supply
circuit. From a processing and manu~acturin~ standpoin~,
enlargement or these features is simple, in contrast to
the much more ~ icult problem of ma~ing the sa~e fea-
tures s~aller, as would be required to enhance damping via
the traditional techniques discussed above in the 8ack-
ground Art.
There are ~wo examples of how this principle may be
. used to enhance the operation of ink-jet printheads. In
the ~irst example, the directionality proble~ arising from
nozzle wet-ou~, referred to as "streaking", is eliminated
Case 188459
.:.: . " . - , . . : . . .
04~
1 via an ink viscosity increase from an original value of 5
cp to an adjusted value o~ 7.5 cp.
In the second example, the issue o~ insufficient
damping at the manufacturable limit o~ the barrier struc-
ture is addressed. In this case, the channel architec-
tures are enlarged to accommodate the thicker ink. The
original ink had a viscosity of 1.2 cp. The intermediate
ink viscosity was 5 cp. The t.hickest ink had a viscosity
o~ p.
This invention involves adjustment of ink viscosity
as a means of enhancing printhea~ performance in situa-
tions where hydraulic tuning is impossible or impractica~,
i'.e., head architectures which are already at the limits
of manu~acturability and/or which are no longer available
for changes due to other design constraints. Adjustment
of inX viscosity allows an otherwise impossible range of
tradeoffs between nozzle refill and meniscus settling time
to be made i~ such printheadsO The operating speed im-
provements which this ~ec.hnique permit are quite large: a
three- to five-fold increase in operatin~ speed over
current state o~ the art.
~ While the drop stability in the pen is improved with
an increase in the viscosity o~ the ink, such increase in
viscosity does not generate crusting, clo~ging an~ prob-
l~ms in print quality over the entire environ~ental oper-
ating range ~typically given as 30-C/70% RH to 15-C/20
RH, where RH is relative humidi~y).
This invention allows s~all drop-volume printers to
operate witho~t the ordinarily-encoun~ered stability
problems at droplet ejection ra~ed a~ove 10 kHz. I~ also
allows the dyna~ics of ink droplet formation to be decou-
pled from the wettability of the orifice plate, which
prevents uneven frequency response, trajectory errors ~nd
air ingestion. It allows these printheads to avoid such
characteristics even in those situations where similar
tuning effor~s (via inclusion of lumped resistive ele
ments, ~or instance) are prohibited.
Case 188459
~I 2q~0160~'~
11
FIG. 1 is a perspective view of a nozzle plate and
no2zles therein, depicting an emerging droplet of ink from
5 a nozzle and a puddle of ink associated therewith;
FIG. 2 is a cross-section taken aiong th~ line 2-2 of
FIG. 1, showing the s~ruct~re of one par~icular drop
generator: and
FIG. 3 is a view similar 1:o that of FIG. 2, showing a
10 portion of an ink-filled drop generator with a bulging
meniscus and a drop of ink wetting the nozzle plate at
characteristic wetting angle ew.
~E~OD~S ~OR CAR~UT T}~NVENT~QN
Referring now to the drawinqs wherein liXe numerals
of reference designate like elements ~hroughou~, a portion
o~ a printhead is depicted in FIG. l. In particular is
seen a noæzle plate lO in which are recessed a plurality
of nozzles 12 in individual recesses 13. Ink 14 is fired
~rom resistors through the nozzles in a particular ar-
rang~ment toward a print medium (e.g., paper) to form
alphanumeric characters and graphics.
FIG. 2 d~picts a portion of a feed cha~ber 16 in
which is located a resistor 18; there is one resistor
associated with each nozzle 12. Ink is ~ed into the feed
chambers from a plenum tnot shown). Upon receiving a
puls~ o~ energy from an external source, ~he resistor 18
is heated to a level sufficient to expel a droplet of ink
14 toward the print medium. Pollowing ejection of the ink
droplet 14, additional ink fills the chamber 16 in prepa-
ration for another ~iring.
The nozzle ~2 has a nozzle diameter d; ~ach resistor
covers a square area with side dimension s; the channel
~5 width is given by w. The thicXness of the nozzle plate lO
is tp, while the thickness of barrier layer 20 is tb. In
a preferred example, the printhead employs a barrier layer
20 comprising ~lacrel 55 ~ thick and a nozzle plate lO
Case l88459
~ 7
1 comprising gold-plated nickel 63 ~m thick. The nozzles 12
are 47 ~3 ~m diameter, with resistors 64 ~m x 64 ~m, and
channel width 84 ~m wide.
As indicated in the Background Art section, during
the overshoot phase, a puddle 22 of ink may forD adjacent
th~ nozzle 12~ rf not wicXed back into ~he chamber, such
a puddle may have a deleteriou~; ef~ect upon print quality
by interfering with the droplet 14 of ink as it is ejected
from the noz21e 12.
During the refill process, the meniscus overshoots
its equilibrium position, is slowed, stopped, and eventu-
ally reversed by the surface tension of the meniscus. The
~aximum overshoot occurs when the meniscus is stopped. Xn
PIG. 3, e corresponds to the maximum overshoot of the
meniscus. The angla e is defined by a tangent to the
meniscus sur~ace at the nozzle perimeter and a line drawn
parallel to ~he top plate surface. To avoid spillage onto
t~e top pla~e, e should ~e less than ew, the characteris-
tic wetting angle for the ink and top plate materials.
As used herein, a s~able drop gen~rator is one ~hat
makes drops with consistent trajectories, volumes, speeds,
and ~reak-up patterns. In accordance with the invention,
this stability becomes more likely as the viscosity is
increased. This is becaus~ it is the damping effect of
2~ viscosity that will balance and control the inertial and
surface forces that drive the refill and ejection process-
~s. Unstable drop qenerators with low viscosity are
characterized by chaotic meniscus movement, large meniscus
overshoots, erra~ic spray patterns, and puddles 22.
This stability can be ~easured by looking at the
accuracy and consistency of dot placemen and size.
Stability was measured by looking a~ line spacing on
paper. Tha od~-numbered nozzles in the pen were fired
across th~ page, forming a set of parallel lines. Then,
an identical pattern was made with the even-numbered
nozzles on a di~ferent part of the page. A ~ision system
then exami~ed the patterns, measuring line spacing uni-
formity and line width uniformity.
Case 1~8459
$
4~
13
1 ~hese measure~ents of line spacing and width were
then combined into an overall print quality number", with
4 being a perfect grade. Test:ing at 30-C (the worst-case
operating temperature for print quali~y) revealed that 40%
H20/60% DEG ink (DEG is diethylene glycol) had a print
quality number of 3.2, which was a ~ull two points better
than the l~ss viscous 50% H20/50% DEG ~print quality
number of 1.2). ~lso, consi~ering dot placement only,
measuremen~s of cross-scan directionality using the same
plot showed that going from an ink viscosity of S to 7.5
cp (50/50 water/DEG to 40/60 water/iEG) decrea~ed the
variation in anqular misdirection by 43~:
_==========_=====================______=__======__========
TABLE I.
Printing Results (30 C/70%XH)
InX ambient viscosity cross-scan 3-sigma
( %H20/DEG ) ( Cp ) ( degreQs )
50/50 5O0 1.20
40J60 7.5 0.69
Co~putex ~odeling of the ink flow in the printhead
confirmed the improved stability obtained with the higher
~iscosity ink modeling o~ a pen e~ploying 59 ~ ~ac.el and
nozzles 12 having a diameter of 43 ~m and showed ~hat a
change in vehicle composition from 50/50 water/DEG to
40/60 water/DEG should result in the following changes in
. 30 refill time, overshoot~ and damping:Y
Case 138459
20~604~
14 :~
TA~LE II.
% Chan~e Relative to 50/50 at 30 C
.... ..
Ink Temp. Refill Overshoot Da~ping
¦ 40/60 H20/DEG 60-C ~2.3% -25.1% +36.9%
. 40/60 H20/DEG 30 C +18.2 49.3 +67.4
In another experimental comparison, 50/50 in~ evi~
dence~ a spear drop onset at 3,500 Hz. (Spear drops are
headless, very ~ast, and us~lally misdirected: they appear :.
above certain critical frequencies.) 40/60 ink evidenced
a spear drop onset at frequencies of about 4,800 Hz, while -
30/70 ink evidenced a spear drop onset at frequencies
greater than 5,500 Hz. :.
In yet another experimental comparison, various
compositions of ink were ~ired ~rom pens with the follow~
ing results:
TABLE III.
Properties for Various Viscosi~ies of H20/DEG ¦:
% H2Vmin/Vss Directionalityviscosity
. cp (25 C)
0.729 1.5 3.ag
0.833 4 5.48
0.860 7.5 8.61
>0.95 9 Y 13.~1
:,
20% N~P 0.901 7.5 7.94 :
Notes: ( 1 ) Vmin/Vss is the ratio between the min-
imum velocity in a frequency resonance plot and
; the st:eady state velocity. A value close to 1.0 :.
indicates stability over the frequency range.
Case 1884S9
~0(~6~4~7
~:: 15
( 2 ) Direc~ionality i5 on a scale o~ 0 to 9,
where 0 i~; bad and 9 is good.
( 3 ) 20% NMP = 409~ H20, 40% DEG, 209~ N-
methyl pyrrolidone.
Print quality was d~ermined for a variety o~ compo-
sitions, using the pre~err~d printhead con~iguration given
above. The r~sults are listed below, with average print
10 quali~y given ~or ~he indicated vehicle composition. The
higher the value, the better the print quali~y. Each pen
ha~ three groups of ten nozzles each; each such group i~s
~alled a primitive. In the test, a visual deter~ination
W2S made for each half-primi~ive (the odd or even
nozzles), an~ su~med for all six half-primitives to arrive
a~ the average PQ rating. The rating is based on 0 = very
poor, 1 = poor, 2 = fair, and 3 = good; values of 2 and
abo~e there, 12.0 and above) are deemed acceptable.
7~-~--~=~7~D=== ~
TABLE IV.
Print Quality vs. % H20/DEG
H20/DEG Av~. PQ Rating
70/3~ 0 . 8
60/40 8 . 0
50/50 ~.1 . 9
40/60 16 . 5
30/70 18 . 0
3 0 I y
5% ~JMP 14 0
10~ IP 13.8
15~ IP 16.0 ;~
35 Note: ~ NMP plus equal portions of ~2 and DEG ~ :~
::,
~.
Case 188459 ~ ~
~ 21D(~60~
i-
16
1 From the foregoing, it is evident that an increase in
v~scosity of ink i~proves the print guali~y considerably.
~owever, there i5 an upper limit on the viscosity of ink
that may be e~ployed, since higher ,viscosity inks taXe
longer to dry and increase the refill time. Indeed, in
conjunction with some print ~eclia (e.g., Mylar transparen- :
cies)~ the upper limit is severely constrained. For
exa~ple, 30/70 H20/DEG is useful wi~h paper, but cannot be
used with Mylar transparencies.
Although diethylene glycol was used to increase the
viscosity of the in~s in the ~oregoing examples, it Wi7 1
¦ be readily clear to those skille~ in this art that the
~eachin~s of this invention are applicable to any of the
water-miscible glycols typically used in ink-jet printing.
Thus, in addition to diethylene glycol, ethylene glycol
and propylene glycol are but a few examples of ~he many
I glycols that are used in ink-jet printing, and an increase
in the glycol conten~ rela~ive ~o water will accomplish
the same purpose, with the sa~e end result as indicated
a~ove.
,.. , ., ;'.
Case 188459
