Steel barrel, TO-220 heater, PLA experiments
mccoyn
last week. Warning: It is quite long.
I ran out of ABS about a week ago and so I'm switching to PLA to print
out the remaining Mendel parts. Since my brass barrel is clogged with
ABS I decided this would also be a good time to experiment with a
stainless steel barrel.
I disassembled my extruder last week. There was a small amount of ABS
feed still sticking out above the brass barrel. When I removed this
from the PTFE I noticed a bulge in the ABS that extended from the
brass barrel opening up about 3/8 inch. This occurs because the
material is liquid before entering the barrel and wears away at the
PTFE. The brass barrel does a good job of spreading heat out and I
doubt that it is possible to have the nozzle hot enough to extrude
while the PTFE end is below the melting point.
This is where the stainless steel barrel comes in. It has lower
thermal conductance and so can maintain a much steeper temperature
gradient. It is also stronger, which means the walls can be thinner
decreasing the thermal conductance even further.
A few people have experimented with steel barrels and the general
theme is to concentrate the heater near the nozzle end because it is
difficult to heat the nozzle end if it is far away from the heater.
The simplest method I saw was to get a power resistor in a TO-220
package. This is very small and has an integrated heatsink which I
drilled and tapped so that I could screw it onto my heater barrel. I
also drilled and tapped a small bit of Aluminum to use as a nut to
lock against the resistor holding it in place and improving the
thermal coupling. This all seemed to work well at first. After a few
hours though, the heater would sometimes fail to operate. I believe
the cause was an internal solder joint overheated and came loose.
I only had one of those resistors so I had to fall back on a nichrome
wire heater. I had set this up plenty of times on my brass barrel so
it was not difficult to build it up. I used the same piece of 8Ω
nichrome wire that I had on my brass barrel. It heated up much faster
since less heat was lost on the top half of the barrel. It also had
more overshoot and oscillation around the target temperature. My
thermistor is in the same spot as it was on the brass barrel, so the
cause of that is the lower thermal conductance between the heater and
thermistor. I am still using bang-bang heater control so PID control
may solve this problem.
With the heater apparently operating well I turned my attention to
extruding PLA. I found some recommended temperature settings and
tried them out. The extruder spat out some black liquid and smoke, a
sure sign that I was overheating the PLA. I lowered the temperature
until this no longer happened. I tried running the extruder for an
extended period of time and it slipped and stalled. Decreasing the
gap from 2.7mm to 2.5mm prevents the stalls, but now my extruder
skipped. I increased the current limit on the chopper driver, but it
still skipped. Increasing the temperature reduced the amount it
skipped and increasing the gap to 2.6mm reduced the skipping even
more.
This is a situation I recall being in while trying to get ABS to
extrude. There are many variables to adjust and while they are all
incorrect there is no way to tell what the correct value for any one
variable is. The variables in this case are the temperature, the gap
size, the stepper current and the flow rate. I know enough to not get
ahead of myself this time so I am not moving the head meaning the
feedrate and bed clearance are not significant factors like they were
when I was trying to get ABS working.
I eventually noticed that the black liquid does not come out right
away. Instead I extrude a little bit of PLA just fine and then it
gets really bad. First the liquid comes out and then the smoke pops
out. If I keep running I get clean PLA again and as long as the
extruder does not stop it stays clean. What I figure is happening is
that the section of the barrel that is in direct contact with the
heater is getting much hotter than the nozzle. When I stop it
overheats the PLA that stops in that section. Since starting and
stopping is common during a print I do not think this set up will work
at all.
Tomorrow I plan to build a new heater using parts I have lying
around. I'll cut a piece of aluminum to transfer heat. This will be
drilled and tapped to screw onto the steel barrel, very close to the
nozzle. I'll put the thermistor on the aluminum near the barrel so
that bang-bang control is stable. I'll file the edges so they won't
cut into nichrome wire and then wrap it with nichrome wire insulated
with kapton tape. Finally, I'll clamp some aluminum sheet around the
outside of the heater coil to slow down the break down of the kapton
tape and add stability. I've done this on my previous heater barrel
and it was the most reliable heater barrel build I've had.
The other option is to clean the ABS out of my brass barrel and build
a more standard barrel. There are two ways to clean the ABS out. I
can dissolve it in acetone, but I don't know where to get that
locally. I can also drill it out. To hold it parallel to my drill
press I will drill a hole in a block of wood, tap it and screw in the
barrel. The block of wood gets clamped to my drill press in the exact
same position it was in when the hole was drilled so the bit should be
centered. The tapped hole won't go all the way through the wood block
so if the bit catches on the barrel it will tighten and then bind up.
I really would like a drill bit that is softer than brass so that I
don't damage the inside.
Tim Schmidt
Also, acetone is available at any drug store or supermarket. Look for
"nail polish remover". Some brands are acetone-free, but many still
contain acetone.
--tim
Keith Mc
Nick wrote:
> [PTFE wearing near barrel interface, probably due to brass barrel
> conducting heat too well, making the ABS liquid while still touching
> the hot end of the PTFE, wearing it down...]
Considering the barrel temps being attempted, the use of PTFE on the
top has always been a concern of mine all along, since it is close to
its disassociation temp.
It may be now worth trying out a ceramic tube in place of the PTFE.
We can probably get some fabbed for us, to spec. There is a Potter's Guild
makerspace in Ann Arbor. It is on the south end of the strip mall on Broadway,
right where it hits Maiden Lane. Given dimensional specs, we may be able
to get one of the potters there to whip up a few custom short tubes for us
cheaply, for testing.
In fact, if you can provide them a PAIR of barrels and a tightly fitting stainless
steel rod as an insert, it MAY work to form and fire one directly wrapped
on the top of the two barrels (spaced out by an inch or so by the rod),
to form a "threaded ceramic coupler". Fire it, then once cool pull out the rod
and unscrew the two barrels from the ends.
If it can be formulated and formed properly to survive the metal shrinkage
intact, you now have a custom tubular coupler with the proper ID,
rod dimension hole in the middle, and internally threaded ends.
It uses a second barrel as the "Upper Mounting Screw".
This should make a FINE coupler, and solve both the insulating problem
and the PTFE thermal wear problem. Does this image make sense?
> This is where the stainless steel barrel comes in. [lower thermal conductance, etc.]
I think addressing the coupler instead of the barrel may be more productive
in the long run. First off, Brass is easy, but Stainless Steel is a BEAR to machine,
especially if thin. Second, over time you may STILL end up with a thermal
interface problem with the PTFE.
If we can find a way to SIMPLY fab a higher temp isolation barrel mounting
material (eg fired ceramic, cast glass, refractory cement, hyper temp casting
material, "fireproof glue", or whatever) I think it could solve a LOT of problems
in one fell swoop.
> [thermal gradient problem]
> The simplest method I saw was to get a power resistor in a TO-220 package.
> [...]
> After a few hours though, the heater would sometimes fail to operate.
> I believe the cause was an internal solder joint overheated and came loose.
You're probably right. One of my concerns with the Power Resistor method is
that the whole point of the heat sink is to cool it enough in operation to allow
them to USE things like solder in its construction. We OTOH are TRYING to make
it get HOT, defeating the whole purpose of its heat sink. We are probably
therefore running it at temps higher then they intended, with that construction tech.
IF you wish to use power resistors, it may worth a bit of correspondence with the
local Rep firms for the brands you wish to use. Just call them, tell them what we
are attempting, and have their company FAE (Field App Engineer) investigate and
recommend the best part for the job that WILL allow operation at the desired temps
without failing. This can be done with a single phone call, and THEY will do the
legwork for you! <grin> (I used to be an industrial distributor FAE... This is
exactly what the Manufacturer Rep Firms are FOR... :-)
> I also drilled and tapped a small bit of Aluminum to use as a nut to
> lock against the resistor holding it in place and improving the thermal coupling.
I'm unsure of your geometry of the aluminum. If it is like a strap or a Nut,
it may couple heat in too small of an area mid-barrel, making a Hot Spot
followed by a cooling zone before extrusion. (That DOES sound like a formula
for roasting the extruded material, for you'd have to overheat the material
to compensate for the cooling effect of the unheated nozzle.)
> [I had a Hot Spot problem mid barrel with PLA, making black smoke, etc,
> but it wasn't there all the time while extruding...]
You know, we really have FIVE separate Control Laws in play here, not ONE:
A) WARMUP (initial startup of the system - can be omitted with a good IDLE)
B) IDLE (could includes WARMUP)
C) EXTRUDE
D) START (The IDLE->EXTRUDE transition), and
E) PAUSE (the EXTRUDE->IDLE transition)
This may be a BIG part of the problem as well. Each of the last four cases
require different PID parameters. The last two are complex cases as
well. In control theory, they are Stochastic, Time-Variant problems.
(Like a rocket whose mass is shrinking and the air it thinning, their
control is actually more complex than either cases B or C).
I think we need to research ways to guarantee
1) the VERY hottest spot IS the nozzle tip,
2) the barrel-heater system's thermal mass, heater gain, and thermal gradient are
all properly designed, and
3) we have the proper Control Laws for each operating state.
I'm unsure we've properly answered these questions yet.
One thing I think would be worth exploring is a TWO STAGE heater.
Wrap a MASSIVE barrel with the NiCr normally. Create a SECOND heater,
that works with JUST the tip. Each heater has its own local thermistor.
We first bring the material temp CLOSE to extrude temp along the barrel,
but a few degrees below it. If the barrel is long, AND high mass, this
deals with PreHeating the material properly whether it is moving or not.
Now run a SECOND control loop at the very tip itself with a VERY high
gain, low thermal mass heater, to "cross the threshold" for extrusion.
This could GREATLY simplify creating and tuning the control laws, for you now
have very little total energy differences between EXTRUDE and IDLE, etc.
The barrel does most of the grunt work, but the tip heater is highly
adaptable, as it only has to modulate the material's temp by a few degrees.
A high enough gain tip heater and a fast control loop (low thermal mass)
should be able to help make up for the slight temp diffs when transitioning
between IDLE and EXTRUDE. It also makes up for the slight thermal drift
due to the new barrel's higher thermal mass.
Thoughts?
- Keith Mc.
Phil Frost
Any hardware store, home depot or similar will have real acetone for a
much better price. It's probably in the paint department, next to the
paint thinner and all sorts of other fun solvents. Kleen-strip is the
most common brand; It usually comes in a blue metal can.
Tim Schmidt
Yes. Excellent suggestion.
--tim
Keith Mc
Yes, be careful of some store's nail polish removers. Some are
totally acetone free now, and many have "contaminants" (eg fragrances).
I once researched what stores carry "pure chemicals". Many Paint and
Cosmetic/HBA departments have quite a selection of chemicals.
You can try Lowes, Home Depot, KMart, Meijers, and most hardware stores.
I found that KMart offers 100% Actetone in their "generic brand"
cosmetic area in a clear bottle, fairly cheaply.
BTW... I also found that you can get acrylic powder and hardener
in just about any large cosmetics area (eg Meijers), in the form of
"artificial nails from scratch" kits.
I've been thinking that given a powder and a hardener dispensing
RR head (or a hardner dispenser and a powder bed), I think there may
be a decent potential to make an RR for solid acrylics with that.
- Keith Mc.
mccoyn
http://www.codeerrors.com/blog/a/117/heater-barrel-failures
I found acetone at Meijer in the paint section. After soaking for 20
minutes the ABS as soft enough to scrape out. It worked well.