Ansys FAQ

This article contains answers for Ansys related FAQ. Please feel free to add questions
and answers here!
Contents
[hide]
1 CFX-
o 1.1 !ol"er related questions
1.1.1 #y steady state solution con"er$es for a while %ut sto&s
con"er$in$ %efore reachin$ 'y con"er$ence criteria
1.1.( )ow to a"oid *+++ - ,+++ - te'&eratures usin$ finite rate
che'istry 'odel
1.1.. Can CFX do a (/ si'ulation0
1.1.1 2s there any way of doin$ a (/ si'ulation in CFX0
1.1. 2 $et the error 3Floatin$ &oint e4ce&tion5 6"erflow3. 7hat
does this 'ean and what do 2 do a%out it0
1.1.* /urin$ the solution &rocess8 2 $et the warnin$ 'essa$e 3A
wall has %een &laced at &ortion9s: of an 6;T<=T3
o 1.( Post&rocessin$
1.(.1 7hy are 'y results inaccurate0
1.(.( Are 'y results &u%lisha%le0
1.(.. 7hy do 2 see a non->ero "elocity at the walls0
( F!2 9Fluid !tructure 2nteraction:
o (.1 ?eneral
(.1.1 7hich A@!A! &roducts are necessary to sol"e a F!2
si'ulation0
(.1.( 7hich Ansys licenses are required for F!2 in Ansys0
(.1.. 7hat Bind of cou&lin$ 'ethods are &ossi%le0
(.1.1 2s it &ossi%le to &erfor' a steady-state F!2 si'ulation0
(.1. 7hat is the $eneral &rocedure for a F!2 si'ulation in Ansys
11.+0
o (.( !ol"in$
(.(.1 )ow do 2 start an F!2 si'ulation in Ansys 11.+0
(.(.( 2s it necessary to define the Ansys 2nsallation Coot in 3/efine
Cun3 in Ansys 11.+0
(.(.. 2s it &ossi%le to sto& a runnin$ F!2 si'ulation in Ansys 11.+0
(.(.1 2s it &ossi%le to write %acB-u& files of a F!2 si'ulation in
Ansys 11.+0
(.(. The sol"er ter'inates with the error 'essa$e 3A ne$ati"e
"olu'e a&&eared3. 7hat went wron$0
o (.. Post&rocessin$
(...1 2f 2 o&en the F!2 si'ulation in CFX-Post 11.+ only the fluid
data are a"aila%le. 7here are the solid data0
. 2C=# CF/
o ..1 ?rid Cefine'ent
o ..( ?rid Cesolution
o ... ?rid Quality
o ..1 7here can 2 find a %asic 2C=# tutorial0
1 )ow should 2 asB 'y question on the CFX foru' to $et the %est &ossi%le
answer0
CFX-5
Solver related questions
My steady state solution converges for a while but stops converging before reaching
y convergence criteria
Failure to o%tain full con"er$ence is a co''on issue for steady state si'ulations. 2t often
occurs when doin$ 'esh refine'ent studies8 where coarse 'esh si'ulations con"er$e
fully and quicBly %ut as the 'esh is refined the si'ulation fails to achie"e the sa'e
residuals con"er$ence. Frequently the cause for this %eha"iour is %y the coarse 'esh not
resol"in$ s'all sheddin$ features8 %ut as the 'esh is refined s'aller flow features are
resol"ed and unless they are &ro&erly handled %y the si'ulation setu& it can lead to
con"er$ence &ro%le's.
To resol"e the issue worB throu$h the followin$ &oints5
1. Cead the CFX docu'entation. !&ecifically the section titled 3Ti&s for o%tainin$
con"er$ence38 and any %est &ractises $uide a&&lica%le to your si'ulation.
(. The first thin$ to consider is whether your si'ulation is sufficiently con"er$ed
e"en thou$h your s&ecified criteria has not %een 'et. To checB whether your
si'ulation is sufficiently con"er$ed8 out&ut the &ara'eters i'&ortant to your
si'ulation to 'onitor &oints and dis&lay the' in the sol"er 'ana$er. A&&ro&riate
&ara'eters could %e lift8 dra$8 &ressure loss8 flow rate - what e"er is an i'&ortant
&ara'eter to your si'ulation. 2f these &ara'eters are not chan$in$ to an accuracy
tolerance suita%le for your si'ulation then your si'ulation is al'ost certainly 6-
as it is and no further worB needs %e done.
.. 2f i'&ortant &ara'eters ha"e not con"er$ed so a ti$hter con"er$ence is required8
the ne4t thin$ to try is tryin$ to $et the si'ulation as it currently is to con"er$e.
Ti&s here are5
o ;se a lar$er &hysical ti'e ste&. A ti'e ste& a&&ro4i'ately equal to the
a"era$e residence ti'e in the si'ulation do'ain is a $ood $uide for 'ost
si'ulations. 2f it is a recirculatin$ syste' without an inlet or outlet then
use the turn o"er ti'e of the lar$est flow feature. Aou can $et the
residence ti'e in CFX-Post %y &lacin$ a strea'line and looBin$ at the
3Ti'e3 "aria%le on it. The 'a4i'u' "alue of ti'e is the residence ti'e.
o ;se <ocal Ti'escale Factor. A factor of a%out is a $ood $uess to start
with. 2f this is successful you should run the final few iterations to
con"er$ence with a &hysical ti'escale 9not local ti'escale all the way to
con"er$ence:.
o For so'e si'ulations usin$ dou%le &recision can hel&8 &articularly if there
is a lar$e ran$e %etween 'a4i'u' and 'ini'u' "alues of di'ensions or
flow &ara'eters 9"elocity8 &ressure etc:. 2t can also useful for %uoyancy
dri"en si'ulations.
o 2f usin$ the hy%rid differencin$ sche'e you can consider reducin$ the
%lend factor. /onDt reduce it %elow +., without showin$ it does not har'
accuracy %y a sensiti"ity checB.
1. 2f you cannot $et it to fully con"er$e then you should looB at why it has not
con"er$ed and try to fi4 the &ro%le'. ChecB5
o The &hysics of the si'ulation is correct.
o /o a test run with the residuals included in the result file. 2t is liBely a
s'all re$ion of the flow has hi$h residuals while the rest is con"er$in$.
Consider why are the residuals hi$h in that re$ion - 2s it5
1. Poor quality 'esh - the fi4 is o%"iously do a %etter quality 'esh
(. A &hysical insta%ility8 such as "orte4 sheddin$ - the fi4 here is to
use a lar$er ti'escale8 a coarser 'esh in the "orte4 sheddin$
re$ion8 decrease the %lend factor 9if usin$ hy%rid differencin$: or
use a lower order tur%ulence 'odel. The first o&tion is &referred as
the latter o&tions can ha"e accuracy i'&lications.
. 2f you still cannot $et the si'ulation to con"er$e then try runnin$ it as a transient
si'ulation. Ada&ti"e ti'este&&in$ can %e useful here to quicBly find the
a&&ro&riate ti'escale. Transient si'ulations are 'uch slower than steady state
si'ulations so %e aware that you will need e4tra &atience.
*. 2f the transient si'ulation shows the results to not %e steady state then $i"e u& on
the steady state 'odel as the flow is transient and needs a transient solution to
&ro&erly ca&ture it.
!ow to avoid "### - $### % teperatures using finite rate cheistry odel
=dit the def file and add the followin$ to the edited ccl file. 9Aou can do it %y si'&ly
clicBin$ the =/2T %utton in cf4sol"er or usin$ cf4c'ds co''and: Add these lines to
the =XP=CT section5
EXPERT PARAMETERS:
stiff chemistry = t
END
2f you donDt ha"e this section8 you can create it and then add a line there 9see the 'anual:.
Can CFX do a &' siulation(
This is discussed in the CFX docu'entation8 %ut it has %een asBed so 'any ti'es on the
CF/-6nline Foru' it is worth re&eatin$. CFX cannot do a (/ si'ulation.
)s there any way of doing a &' siulation in CFX(
Aes. Fro' a (/ 'esh of the $eo'etry8 e4trude it one ele'ent in the nor'al direction.
For a &' planar siulation this would %e one ele'ent in the nor'al "ector direction8 for
a &' a*isyetric siulation this would %e swee&in$ a s'all an$le with one ele'ent.
For the &lanar 'esh the e4trusion should %e a&&ro4i'ately equal to the s'allest ele'ent
ed$e len$th in the 'odel8 for the a4isy''etric 'esh the swee& should %e a s'all an$le8
'a4i'u' E %ut s'aller if you want hi$h accuracy.
2n CFX-Pre you should set the to& and %otto' faces of the e4trusion as sy''etry &lanes.
2f you want to include swirl in the 'odel use &eriodic %oundaries. The re'ainin$
%oundaries should %e set as walls8 inlets8 o&enin$s and outlets to define the flow.
The CFX docu'entation discusses (/ si'ulations and it is reco''ended you read it
%efore &roceedin$.
) get the error +Floating point e*ception, -verflow+. /hat does this ean and what
do ) do about it(
This error 'eans the linear sol"er has di"er$ed %adly. Aou will need to i'&ro"e the
nu'erical sta%ility of the si'ulation for this si'ulation to con"er$e. There can %e 'any
causes for this di"er$ence8 %ut the 'ain ones are 9listed in a&&ro4i'ately the order they
should %e checBed in:5
)s the physics of the siulation set up correctly( )a"e you for$ot a 'aterial
&ro&erty0 6r ha"e you entered a 'aterial &ro&erty incorrectly0 )a"e you left out
a "ital &hysical 'odel0 /oes your C=< correctly e"aluate0
)s the esh of high enough quality( /ifferent si'ulations ha"e different 'esh
quality require'ents. !i'&le low Ceynolds nu'%er flow can handle lower quality
ele'ents than hi$h Ceynols nu'%er flows8 and so'e &hysical 'odels require hi$h
quality 'esh. #ulti&hase 'odellin$ usually requires a hi$her 'esh quality than
sin$le &hase8 and so'e 'ulti&hase 'odels are "ery strict - for instance accurate
surface tension 'odellin$ can require as&ect ratios of less than 1.8 %ut si'&le
flows can handle as&ect ratios 1+ or 1++ ti'es worse than this. Also note that hi$h
as&ect ratio ele'ents in &ris' layers does not necessarily cause &ro%le's as the
ele'ent is ali$ned with the flow direction.
)s a better initial condition required( The closer the initial condition is to the
real flow field the easier it will %e for the sol"er to con"er$e. Aou can achie"e this
i'&ro"ed initial condition %y5
o ;sin$ a &re"ious si'ulation which has con"er$ed at a condition close to
the &ro%le' si'ulationDs condition.
o ;sin$ u&windin$ for the ad"ection ter's. @ote this is not "ery accurate
and is not reco''ended for final results8 %ut it is a useful way of
$eneratin$ initial conditions which are close.
o ;sin$ a coarse 'esh and inter&olatin$ as an initial condition onto the final
'esh.
o /oin$ a si'ulation which does not include tricBy &hysics8 for instance
doin$ a sin$le &hase 'odel %efore introducin$ the 'ulti&hase &hysics.
)s double precision required( !o'e 'odels require additional nu'erical
accuracy to sol"e. =4a'&les include 'ulti&hase 9es&ecially surface tension: and
%uoyancy dri"en flows.
'oes the odel require sall tie steps to start( !o'e si'ulations require
s'all ti'e ste&s to start8 %ut once $oin$ the ti'e ste& can %e increased.
'uring the solution process0 ) get the warning essage +A wall has been placed at
portion1s2 of an -34564+
This 'essa$e a&&ears whene"er %acBflow would occur at a &ressure outlet %oundary
condition. This ty&e of %oundary condition cannot handle %acBflow into the
co'&utational do'ain.
2f the warnin$ 'essa$e disa&&ears %efore o%tainin$ the final solution8 the results are still
"alid. 2f the warnin$ &ersists8 you ha"e three o&tions5
1. ;se an o&enin$ ty&e %oundary condition liBe the warnin$ 'essa$e su$$ests.
)owe"er8 you ha"e to s&ecify the direction of the %acBflow at an o&enin$. ;nless
you ha"e infor'ation a%out the direction of the %acBflow8 the second o&tion
in"ol"es less 'odelin$ uncertainties.
(. #o"e the outlet further downstrea' to a &osition where no %acBflow occurs.
.. 2f the outlet and the wall is not si$nificantly affectin$ your results you can i$nore
it and acce&t the solution with the %acB flow warnin$.
7ostprocessing
/hy are y results inaccurate(
1. Firstly8 is the si'ulation correctly set u&0 /oes your 'odel include all rele"ant
&hysics0
(. )as your solution con"er$ed to a reasona%le "alue 9see a%o"e:0 For a si'&le
analysis8 an C#! residual of 1=- should %e sufficient. -ee& lowerin$ your
residual "alue until the solution no lon$er con"er$es 'onotonically.
.. Perfor' a sensiti"ity analysis on the rele"ant features of your si'ulation. 2n its
si'&lest for'8 a sensiti"ity analysis in"ol"es alterin$ one &ara'eter 9e.$.8 $rid
refine'ent: while Bee&in$ the other &ara'eters constant and o%ser"in$ the effect
this &ara'eter has u&on the si'ulation accuracy 9e.$.8 %y 'onitorin$ sBin
friction:. The &ur&ose of a sensiti"ity analysis is to o%tain a si'ulation that is %oth
accurate and cost effecti"e. For Fust a%out e"ery si'ulation a sensiti"ity analysis
on 'esh si>e and con"er$ence tolerance is required. /e&endin$ on what you are
'odellin$ other sensiti"ity analyses 'ay %e required8 co"erin$ thin$s such as5
o /o'ain si>e 9how close can the %oundaries %e to the re$ion of interest0:
o ?rid density
o ?rid quality
o ?rid ty&e 9e.$.8 structured "ersus unstructured:8
o !&atial discreti>ation 9or ad"ection: sche'e
o Te'&oral discretisation sche'e
o Tur%ulence 'odel8
o Tur%ulence nu'erics8
o Tur%ulence intensity8
o Goundary condition ty&e8
o Any e'&irical 'odels used8 such as heatH'ass transfer coefficients8 dra$
laws8 free surface 'odel8
o Two- "ersus three-di'ensional si'ulations8
o and ti'este& si>e.
1. Perfor' an error analysis5
1. Calculate your discreti>ation errors. A calculation should %e &erfor'ed on
each quantitati"e &ara'eter of interest 9e.$.8 lift:. The Iournal of Fluids
=n$ineerin$ &ro"ides an e4&lanation of &redictin$ discreti>ation errors
here5 Calculatin$ /iscreti>ation =rrors. 2n ter's of errors8 discreti>ation
errors are the $reatest concern to a CF/ 'odeler. /iscreti>ation errors are
those that occur as a result of 'odelin$ the $o"ernin$ flow equations as
al$e%raic e4&ressions in a discrete s&ace-ti'e do'ain. /iscreti>ation
errors should a&&roach >ero as the $rid is resol"ed 9i.e.8 as 'ore $rid
&oints are $enerated and the solution %eco'es $rid-inde&endent:.
)owe"er8 discreti>ation errors also de&end on $rid quality 9e.$.8 as&ect
ratio8 ortho$onality8 sBew:. A well-refined $rid is not necessarily a hi$h
quality $rid and8 thus8 $rid resolution and quality sensiti"ity analyses are
reco''ended.
2. Physical a&&ro4i'ation errors can cause si$nificant accuracy issues. This
ty&e of error is caused %y &hysically a&&ro4i'atin$ so'ethin$ to si'&lify
the si'ulation 9e.$.8 'odelin$ a real fluid as an ideal $as:.
3. Co'&uter round-off errors are caused when a co'&uter stores floatin$
&oint data. /e&endin$ on the accuracy your co'&uter stores the data at8
so'e round-off 'ay occur. This ty&e of error is usually ne$li$i%le
co'&ared to the others and can %e easily decreased %y runnin$ the
si'ulation in /ou%le Precision 'ode.
4. 2terati"e con"er$ence errors are due to the resolution of the CA@!
equations. !ince the @a"ier-!toBes equations are not sol"ed directly8 the
solution iterates until a s&ecified con"er$ence criterion is reached 9e.$.8 a
residual "alue of a certain si>e is reached:. This ty&e of error can %e
decreased %y allowin$ con"er$ence to an acce&ta%le le"el 9see a%o"e:.
5. ;sa$e errors are caused %y user s&ecifications. For e4a'&le8 the user 'ay
run a tur%ulent si'ulation as la'inar or an unsteady si'ulation as steady.
These errors can %e 'onitored "ia a con"er$ence study.
(. And finally - consider the results are you co'&arin$ a$ainst. Are they accurate0
)ow do you Bnow0 Are you 'odellin$ =XACT<A the sa'e conditions as the
results you are co'&arin$ a$ainst0
Are y results publishable(
The Iournal of Fluids =n$ineerin$ &ro"ides a $ood $uide of reco''endations re$ardin$
&u%lishin$ your data. Iournal of Fluids =n$ineerin$ @u'erical Accuracy
/hy do ) see a non-8ero velocity at the walls(
CFX uses control "olu'es centred on the nodes8 so the nodes on the walls $enerate
control "olu'es whose centroids is off the walls. That is why the wall control "olu'es
ha"e non->ero "elocities.
To 'aBe it dis&lay as e4&ected in &ost &rocessin$8 CFX has added the 3hy%rid3 "aria%le
o&tion for "elocities8 where it forces the "elocity to >ero 9or 'ore $enerally the wall
"elocity: at the wall. The 3conser"ati"e3 "aria%les show the actual control "olu'e "alues8
and will not in $eneral8 %e >ero at the wall.
FS) 1Fluid Structure )nteraction2
9eneral
/hich A:S;S products are necessary to solve a FS) siulation(
Ge$innin$ with A@!A! 11.+8 Ansys 7orB%ench with !i'ulation and CFX are required.
Prior to A@!A! 11.+ you needed to use CFX 9standalone or in worB%ench: and the
A@!A! Pre&, ?;2.
/hich Ansys licenses are required for FS) in Ansys(
7orBin$ with Ansys 1+.+ or older8 a #uli&hysics and a CFX license are required.
Ansys 11.+ has 'ulti&le o&tions. 2t is &ossi%le to run an F!2 si'ulation with a sin$le
license now.
/hat <ind of coupling ethods are possible(
6ne-way or two-way F!2 cou&lin$.
)s it possible to perfor a steady-state FS) siulation(
Aes8 Ansys 11.+ ena%les steady state and transient F!2 si'ulations as well.
/hat is the general procedure for a FS) siulation in Ansys ==.#(
For a two-way si'ulation5
1. /efine the !olid setu& in Ansys !i'ulation 9Ansys 7orB%ench:. This includes an
Fluid-!olid-2nterface.
(. 7rite an Ansys 2n&ut File 9.in&: of this setu&.
.. /efine the Fluid setu& in Ansys CFX-Pre. A linB to the Ansys 2n&ut File is required.
1. 7rite an Ansys /efinition File 9.def: of this setu& fro' CFX Pre
. !tart the cou&led F!2 run with the CFX !ol"er #an$er
*. Post&rocess %oth the CF/ and solids results in CFX-Post.
Solving
!ow do ) start an FS) siulation in Ansys ==.#(
!tart the CFX !ol"er #ana$er and load the .def File. Ansys will load the .in& file
auto'atically.
)s it necessary to define the Ansys )nsallation >oot in +'efine >un+ in Ansys ==.#(
6n a 7indows PC it is not8 on ;ni4 it is necessary.
)s it possible to stop a running FS) siulation in Ansys ==.#(
Aes8 hit the sto& %utton. 2t worBs.
)s it possible to write bac<-up files of a FS) siulation in Ansys ==.#(
Aes. Aou can write %acBu& files fro' CFX in the usual way.
4he solver terinates with the error essage +A negative volue appeared+. /hat
went wrong(
This error often a&&ears with F!2 si'ulations. @or'ally it co'es to$ether with a lar$e
deflection si'ulation of the solid &art. #ostly the CF/-'esh defor'ation is too %i$ and
ne$ati"e ele'ents a&&ear. Possi%le solutions 'i$ht %e
a %etter fluid 'esh
'eshsiffness 1Hwall distance
s'aller ti'este&s
7ostprocessing
)f ) open the FS) siulation in CFX-7ost ==.# only the fluid data are available.
/here are the solid data(
CFX-Post only o&ens the .res file %y default. 2f you want to &ost&rocess fluid and solid
data to$ether you ha"e to load the solid data additionally5 File -J <oad Cesult -J <oad the
Ansys .rst or si'ilar data an acti"ate the radio %utton 3add data3. @ow you can
&ost&rocess all data in CFX-Post.
)C6M CF'
9rid >efineent
A well-refined $rid is essential for &ro&erly ca&turin$ the "iscous flow effects. For
tur%ulence 'odelin$8 a %oundary layer thicBness of yK
'a4
L 1 is reco''ended to ca&ture
the "iscous su%layer &ro&erly. )owe"er8 'any tur%ulence 'odels now include wall
scales that allow for yK "alues lar$er than one8 there%y creatin$ a cost-effecti"e and
ro%ust $rid. As Ceynolds nu'%er increases8 $rid refine'ent 'ust8 too. ?rids used in
la'inar flows can %e considered refined once they are resol"ed 9see %elow:.
9rid >esolution
A resol"ed $rid is one in which the solution &ara'eters of interest 9e.$.8 sBin friction: no
lon$er chan$e with further refine'ent. Pro&er $rid resolution is required for si'ulation
accuracy.
9rid Quality
The third &ara'eter to successful $rid de"elo&'ent is $rid quality8 a factor that is often
o"erlooBed. ?rid quality &ara'eters include 9%ut are not li'ited to:5
As&ect ratio
6rtho$onality
!Bew
/eter'inant of the Iaco%ian 'atri4
=4&ansion factor
Ty&e of $rid
o Ad"anta$es of a structured $rid
Mery efficient nu'erical &rocedures can %e i'&le'ented on
structured $rids.
#ulti-%locB $rids allow for custo'i>ation of "arious re$ions.
!tructured %oundary layers and other anisotro&ic re$ions li'it
discreti>ation errors.
!tructured $rids allow for decreased stora$e and less co'&licated
data structures.
o Ad"anta$es of an unstructured $rid
;nstructured $rids are hi$hly ro%ust and si'&ler to $enerate.
There is a lacB of restriction on where nodes are &laced8 increasin$
the auto'ation of $eneration.
;nstructured $rids allow for su%di"ision of cells in areas requirin$
increased accuracy.
2C=# CF/ includes a quality calculator for a priori $rid accuracy. CFX-Post also allows
for so'e a posteriori 'esh quality calculations. For e4ternal aerodyna'ic a&&lications8 a
C-$rid with a 'ulti-%locB structured 'esh is reco''ended with as&ect ratios near unity8
hi$h ortho$onality on walls8 'ini'al sBew8 and a "ery low e4&ansion factor 91.+1-1.+8 if
&ossi%le: in the %oundary layer re$ions.
/here can ) find a basic )C6M tutorial(
There are so'e tutorials a"aila%le on the A@!A! Custo'er Portal. A@!A! also has a
AouTu%e &a$e 9A@!A! AouTu%e &a$e:8 which includes a tutorial on creatin$ a
structured C-$rid around an airfoil.
!ow should ) as< y question on the CFX foru to get
the best possible answer(
!o'e $ood articles on asBin$ effecti"e questions on other online foru's are here and
here. These descri%e the $eneral &rinci&les of writin$ a $ood foru' question.
The 'ost i'&ortant &oint to understand a%out the CFX foru' is that the quality of the
answer to your question will de&end entirely on the quality of your question!
#ost of the questions &osed on the CFX foru' are so &oorly &osed that itDs i'&ossi%le to
understand what the &oster is actually asBin$. 2f you want effecti"e hel& with your
&ro%le' adhere to the followin$ $uidelines5
1. #aBe sure your question is as clear8 concise and as intelli$i%le as &ossi%le. ;se
&unctuation. 6ther foru' readers are not $oin$ to s&end ti'e tryin$ to deci&her a
$ar%led question.
(. ?i"e a clear $eneral descri&tion of what class of &ro%le' andHor a&&lication you
are worBin$ on G=F6C= you start asBin$ s&ecific questions. This will aid other
foru' readers to %etter understand your s&ecific questions.
.. /escri%e &recisely what you ha"e done yourself to try and sol"e your &ro%le'8
$i"in$ e4a'&les.
1. /e&endin$ on your &ro%le' you should always include the followin$5
o A co&y of your co''and file as a file attach'ent. #any si'&le &ro%le's
can %e s&otted in a co''and file %y an e4&ert user.
o 2f you are asBin$ a 'esh quality related question then include so'e
sa'&le i'a$es of your 'esh8 includin$ the %oundary layer.
o 2f you are asBin$ a user Fortran andHor user C=< question then include a
co&y of your e4istin$ code file as an attach'ent.
/ontDs5
/ont asB o"er %road questions e.$. 3)ow do 2 si'ulate a 1-stroBe en$ine3.
@o%ody is $oin$ to ty&e out fifty &a$es of $uidance. They ha"e %etter thin$s to
do.
/onDt shorten the words. 2tDs a we% foru'8 not a !#! te4t 'essa$e. 2tDs "ery hard
to read it. 6r C ; ( la>y0 Peo&le 'ay tend to %e la>y to deci&her and answer.
=4a'&le of a well &osed question5
F2X#=
=4a'&les of questions that are unanswera%le5
F2X#=
)ow to u&load i'a$es5
Create the i'a$e on your local 'achine e.$. sBewN'esh.F&$
;&load the i'a$e to 2'a$eshacB
Co&y the linB 3Thu'%nail for we%sites3
Paste the linB into your &ost e.$. <a href="http://img91.imageshack.us/my.php?
image=pipeym8.jpg" target="_blak"!<img
src="http://img91.imageshack.us/img91/"#81/pipeym8.th.jpg" bor$er="%"
alt="&ree 'mage (ostig at ))).'mage*hack.us" /!</a!
)ow to share non-i'a$e files as attach'ents5
2f you ha"e 'ore than one file8 >i& the' into a sin$le file.
;&load the file to Ca&idshare ?i$asi>e
!croll down to 32 donDt want a collectorDs account ri$ht now. Iust $i"e 'e the
download-linB.3 and clicB it.
!croll down until you see the linB e.$.
http://rapi$share.$e/files/++8%8#,#/au$io.log.html