Subject: RE: Bumpy Surface on SLA parts
Victor Devadas[SMTP:victor@sirim1.sirim.my] wrote:
I am an SLA user, for the past 1 year. The resin used is the SL5170. =
................bottom surface in contact with the supports appear to =
be bumpy. I have reduced the hatch spacing (Box) to .05mm .............
************************
Victor,
I would also check the items which Kamesh at Prototype Express mentioned in
an earlier post today. I am not quite sure what you mean by "bumpy". If
the bumps are occuring between your supports, in a grid like fashion, there
are several other factors you might want to consider.
The type of bumps I am thinking of are caused by fluid forces from the resin
on a thin layer of cured material during the deep-dip process (soon to be
overcome by Zephyr...). Variables which apply here are: skin fills (which
Kamesh mentioned), resin viscosity, dip velocity, dip depth, support spacing,
layer thickness and overcure, and some other minor ones. If the type of
bumps you are experiencing are between the supports (my assumption), then a
quick test would be to slow down the dip velocity during recoat. A negative
to this at that build time will increase slightly, but it really wouldn't
hurt anything just to try it on one build and see if there is a difference.
If you do see a difference, the hard part is root causing what changed in
the first place. Assuming none of the process variables had been changed,
and the bumps suddenly started to appear, I would check the viscosity.
While we have all kinds of fancy rheometers in our R&D labs here, you
probably don't have access to this equipment. If you don't, you could pick
up a viscosity cup in a paint store which is just a cup with a tapered or
spherical bottom with a hole in it. The principle is that thicker materials
will take longer to drain through the hole than thinner ones. The cups for
paint are designed so you just dip them in the paint bucket, lift it up, and
time how long it takes to drain back into the bucket. They are calibrated
for a certain time, and if it takes too long, then you add more thinner. OK,
OK, enough regression for now!! Either find one or make one and do the
following test.........
Time how long it takes for a cup of virgin resin to drain through your
viscosity cup. Time how long it takes for the material in your vat to drain
through. The fresh material material will drain faster (let me know if it
doesn't!!). How much of a difference should one worry about? Hard to say.
You might want to try stirring the vat (there is an automatic routing for
doing this), and measuring it again. How much worse does it get? You are
looking for significant differences that would cause your bump problem. If
they are close to the same, not only have you eliminated viscosity change as
a highly probable cause for your bumpy parts, but you also know that the vat
is fairly healthy (viscosity can be used as an indicator).
What else to try? Hmmm I don't think making your hatch spacing smaller
would help, in fact I think it could make things significantly worse! When
a thin layer of material shrinks (5170 still has some shrink), it will tend
to bunch up, causing bumps. This is the primary reason that the ACES build
style doesn't work so well with materials that have a lot of shrink. When
the first layer starts to draw, bend down in front of the machine and look
through at the surface of the resin very closely. If you stand just right,
you can see tiny differences in the surface with the reflections. Try to
see if the bumps occur while drawing, or during recoat.
Also try to dissect the bumps, are they just one layer thick? This would
provide a lot of information as well. If you are still stuck, check out the
laser power relative to sensor readings. If you don't have a good power
meter or don't care to mess with the K-factors, just compare readings from
sensor 1 with sensor 2 using the beam power utility. You will always see
some difference, but if it is more than a few milliwatts, I would correct
this. Adjust the K-factors such that each sensor reads within less than a
milliwatt of the same power. If you have a power meter, adjust the
k-factors so each sensor reads the actual power.
By the way, the k-factors are in the MACHINE.PRM file in the 3dsys directory.
I think it is the 1 mil hole (not sure now), but should be somewhere
around 0.01. This is a multiplication factor that the machine uses. It
senses some value from the sensor, let's call it Mach_Read. The software
multiplies the Mach_Read by the k-factor to output a power. A little
algebra will speed the adjustment process, rather than hit or miss. This is
a bit of a simplification, because the k-factor for sensor 1 actually
affects the reading of sensor 2. I haven't thought much of it, but one time
I tried changing the k-factor by a large number, and noticed that the other
sensor reading was affected. Anyway.
Mach_Read * k-factor_old = Sens_Read
look for k-factor_old in the machine.prm file, and get Sens_Read using beam
power in the utility menu. Solve for Mach_Read. Now use this and the
actual power obtained with the meter to solve for a new k-factor.
Mach_Read * k-factor_new = Act_Power
Plug the new k-factor in place of the old in the machine.prm file. You may
have to do this twice, but it is really quite easy. Now you have both
sensors reading the same as your power meter. If you don't have a power
meter, it doesn't really matter. What??!! That's right. The laser power
reading of your system relative to the "real" power value in milliwatts
doesn't matter one bit as far as your ability to build good parts. IF you
perform WINDOWPANES, that is!!!!! This is because the calibration
procedure is a closed loop system, where changes from absolute power will be
compensated for by changes in your measured Ec and Dp values. Heck, even 3D
Systems and Ciba Geigy have some of this error in their published Ec and Dp
values. I do happen to know that 3D uses a nice thermopile power meter back
in the labs (has it been calibrated recently guys!) Where it does matter is
when you want to talk to the rest of the world about your laser power, ec,
or dp values. This whole power issue is really about standards, and I don't
want to get into this today, that's for sure. Plus, it won't help your bump
problem any!
I think I'll sign off for now, by the time you finish reading this, your
problem will probably have fixed itself.
E. Derek Smith
Motorola Radio Products Group
Ft. Lauderdale, Florida
(954) 723-4790
eds014@email.mot.com
This archive was generated by hypermail 2.1.2 : Tue Jun 05 2001 - 22:37:26 EEST