3D printing using electron beam melting

[Jamie M]

Hi,

I saw this 3D printing technology and think it is the best one out there:

http://en.wikipedia.org/wiki/Electron_beam_melting

http://www.arcam.com/technology/ebm-process.aspx
(some cool videos)

It can print metal (ie titanium, stainless steel) 3D objects from
melting metal powder one thin layer at a time in a vacuum using an
electron beam.

How hard is it to make an electron beam like this? :) I guess it is
very similar to a CRT electron beam except for the power levels and/or
focus diameter of the beam? Also an electron beam lithography system
has some similarities:
http://en.wikipedia.org/wiki/Electron_beam_lithography

For hobbyist applications this is a good 3D printer for plastics:
http://www.makerbot.com/

cheers,
Jamie

[mike]

I built a 3-axis positioner with a dremel mounted on it. Did
some drilling and routing...
I too was fascinated by the makerbot and considered building the
extruder head and bolting it to my machine.
But I was put off by the $50/spool cost of the plastic filament.

[Martin Riddle]

"Jamie M" <jmo...@shaw.ca> wrote in message
news:jgsr7m$15g$1...@speranza.aioe.org...

Just read this article...
<http://futureoftech.msnbc.msn.com/_news/2012/02/06/10332210-3-d-printed-jaw-lets-83-year-old-breathe-chew-and-talk>
Wonder if it implies the same method.

Lots of possibilities.

Cheers

[The_Giant_Rat_of_Sumatra]

On Tue, 7 Feb 2012 23:16:50 -0500, "Martin Riddle"
<marti...@verizon.net> wrote:

>Wonder if it implies the same method.

You must be going senile too.

There is hard milling "proto" systems:

http://www.rolanddga.com/products/milling/

And there is 3D printing which is a layer by layer 'ozze' build up.

There was a method where a huge vat of liquid was hardened by the
intersection of two lasers, and extremely complex 3D structures could be
prototyped.

[The_Giant_Rat_of_Sumatra]

Make that "ooze".

[Jon Elson]

The_Giant_Rat_of_Sumatra wrote:


>
> There was a method where a huge vat of liquid was hardened by the
> intersection of two lasers, and extremely complex 3D structures could be
> prototyped.

There are guys doing this with computer projectors and a vat of
photo-initiated polymer. See :
http://tech.groups.yahoo.com/group/diy_3d_printing_and_fabrication/
You will have to join to see the pix, but they are awesome.
These guys are doing it with used DLP projectors and home-made
vats with leadscrew elevators.

Jon

[Bill Sloman]

On Feb 8, 4:47 am, Jamie M <jmor...@shaw.ca> wrote:
> Hi,
>
> I saw this 3D printing technology and think it is the best one out there:
>
> http://en.wikipedia.org/wiki/Electron_beam_melting
>
> http://www.arcam.com/technology/ebm-process.aspx
> (some cool videos)
>
> It can print metal (ie titanium, stainless steel) 3D objects from
> melting metal powder one thin layer at a time in a vacuum using an
> electron beam.
>
> How hard is it to make an electron beam like this? :)

It's a well understood technology.

> I guess it is
> very similar to a CRT electron beam except for the power levels and/or
> focus diameter of the beam?

As in an elephant being very similar to a mouse except for the weight
difference.

> Also an electron beam lithography system
> has some similarities:http://en.wikipedia.org/wiki/Electron_beam_lithography

The electron beam microfabricator that I worked on at Cambridge
Instruments had a 20uA primary beam, and could be focussed down to a
10nm diameter spot (though you couldn't get anything like 20uA into a
spot that small - the current density at the imaged spot can't be any
higher than it was a source).

The machines Jamie is talking about use milliamps of beam current, and
look a lot more like an electron beam welder.

http://en.wikipedia.org/wiki/Electron_beam_welding

They talk about splitting the beam into multiple spots, which gets
around the problem of not being able to get a lot of current (and
power) into a small spot.

--
Bill Sloman, Nijmegen

[Robert Macy]

On Feb 8, 5:39 am, Bill Sloman <bill.slo...@ieee.org> wrote:
> ...snip...

> The electron beam microfabricator that I worked on at Cambridge
> Instruments had a 20uA primary beam, and could be focussed down to a
> 10nm diameter spot (though you couldn't get anything like 20uA into a
> spot that small - the current density at the imaged spot can't be any
> higher than it was a source).

> ...snip...
> Bill Sloman, Nijmegen

Bill,

10nm beam width! Not sure of the correct term, what was the focus
depth? How much 'deflection' did you get from ambient magnetic
fields?

[Bill Sloman]

On Feb 8, 5:12 pm, Robert Macy <robert.a.m...@gmail.com> wrote:

> On Feb 8, 5:39 am,BillSloman<bill.slo...@ieee.org> wrote:
>
> > ...snip...
> > The electron beam microfabricator that I worked on at Cambridge
> > Instruments had a 20uA primary beam, and could be focussed down to a
> > 10nm diameter spot (though you couldn't get anything like 20uA into a
> > spot that small - the current density at the imaged spot can't be any
> > higher than it was a source).
> > ...snip...
>

> 10nm beam width! Not sure of the correct term, what was the focus
> depth? How much 'deflection' did you get from ambient magnetic
> fields?

Not the beam width, but the diameter of the focussed spot. I've no
real idea of the depth of focus, but it wasn't large - some microns.

The magnetic lenses were easy enough to focus and if you'd mapped the
surface you wanted to write you could vary the focus across the
written field.

The deflection from ambient magnetic fields depended on the field. One
of my bosses had a great diagnostic moment when riding a lift up to an
up-stairs electron beam microfabricator installation that had been
giving problems - patterns would spontaneously move by a few hundred
microns while being written.

He was an engineering history buff, and recognised the lift as an old
hydraulic lift, which consisted big lump of - magnetic - iron pushed
up and down in a shaft by water pressure. When the lift was up, and
the block of iron was close to the electron-beam microfabricator, the
magnetic field at the electron-beam microfabricator was different from
when the lift was on the ground floor.

Problem recognised was problem solved - the guys using the machine got
to keep the lift out of service when they were writing a pattern.

--
Bill Sloman, Nijmegen

[John Larkin]

I was at the Photonics West show. Lumera Laser makes picosecond-pulse
lasers that ablate most anything. They had some superb
machined/sculpted samples, super resolution, that were made by laser
ablation, one from tungsten carbide and one from mastodon ivory (which
is legal, whereas elephant ivory isn't.)




--

John Larkin, President
Highland Technology, Inc

jlarkin at highlandtechnology dot com
http://www.highlandtechnology.com

Precision electronic instrumentation
Picosecond-resolution Digital Delay and Pulse generators
Custom laser controllers
Photonics and fiberoptic TTL data links
VME thermocouple, LVDT, synchro acquisition and simulation

[Jamie M]

On 08/02/2012 4:39 AM, Bill Sloman wrote:
> On Feb 8, 4:47 am, Jamie M<jmor...@shaw.ca> wrote:
>> Hi,
>>
>> I saw this 3D printing technology and think it is the best one out there:
>>
>> http://en.wikipedia.org/wiki/Electron_beam_melting
>>
>> http://www.arcam.com/technology/ebm-process.aspx
>> (some cool videos)
>>
>> It can print metal (ie titanium, stainless steel) 3D objects from
>> melting metal powder one thin layer at a time in a vacuum using an
>> electron beam.
>>
>> How hard is it to make an electron beam like this? :)
>
> It's a well understood technology.
>
>> I guess it is
>> very similar to a CRT electron beam except for the power levels and/or
>> focus diameter of the beam?
>
> As in an elephant being very similar to a mouse except for the weight
> difference.
>
>> Also an electron beam lithography system
>> has some similarities:http://en.wikipedia.org/wiki/Electron_beam_lithography
>
> The electron beam microfabricator that I worked on at Cambridge
> Instruments had a 20uA primary beam, and could be focussed down to a
> 10nm diameter spot (though you couldn't get anything like 20uA into a
> spot that small - the current density at the imaged spot can't be any
> higher than it was a source).

Hi,

Do electron beams diverge as they travel due to the electron charge?
I guess if they do it also depends on the density of electrons in the
beam? And also the divergence angle would depend on the beam velocity.

cheers,
Jamie

[Bill Sloman]

> Do electron beams diverge as they travel due to the electron charge?
> I guess if they do it also depends on the density of electrons in the
> beam?  And also the divergence angle would depend on the beam velocity.

It's called "space charge", and was a known problem with thermionic
valves.

http://en.wikipedia.org/wiki/Space_charge

<snip>

--
Bill Sloman, Nijmegen

[Tim Williams]

"Jamie M" <jmo...@shaw.ca> wrote in message

news:jgvahf$vot$1...@speranza.aioe.org...

> Do electron beams diverge as they travel due to the electron charge?
> I guess if they do it also depends on the density of electrons in the
> beam? And also the divergence angle would depend on the beam velocity.

Yes.

Incidentally, as you might notice from varying the intensity on a scope, the
spot size generally varies with intensity, because the grid cuts off all but
a spot of the cathode's surface. Larger spot = larger spot. Although a
larger spot delivers more total brightness (same voltage, higher beam
current), the intensity is about the same, because the current density of
the beam is about the same.

Exceptions include very nicely made tubes, where spot size is dominated by
other features, and intensity is particularly high thanks to the accelerator
voltage. Some Tek tubes used a mesh grid at the end of the electrode
structure to enhance deflection sensitivity, while allowing much higher
acceleration voltages; the grid causes some dispersion I think, sinking beam
current and causing the beam to be somewhat thicker, more or less
independent of intensity setting.

At very high current densities and velocities (relativistic I think), the
magnetic field generated by the beam current itself focuses it tighter, much
like a high intensity laser beam heats the air, making a light pipe that
keeps it from dispersing as quickly.

Tim

--
Deep Friar: a very philosophical monk.
Website: http://webpages.charter.net/dawill/tmoranwms

[Robert Baer]

For the first step,fake it by stealing an electron gun from an old CRT.
Pour air into it carefully first...

[Jamie]

Yes they do. Take it from a guy that has had his hands in 2..4 Mev
irradiation units. I can tell you all about focusing, beam steering,
x-rays etc.

Oh, did I forget the mention lots of ozone generation?

Jamie

[Jamie M]

Hi,

So for an EBM application, would this be a good type of setup for
the electron gun?

www.kimballphysics.com/electron_guns/catalog_PDFs/Gun_intro_prelim.pdf

Also when searching for "electron gun" on ebay, mostly these come up:
ebay search: "Airco Temescal Supersource Electron Beam Gun"

What exactly are they? They look like they are meant for metal
sputtering I think.

Also for focusing the beam and deflecting for EBM applications would
modules like these work (of the appropriate sizes), also would more
than one focusing solenoid be required?

deflection coil:
ebay: "Deflection Yoke Coil for Monochrome CRT"

focusing coil:
ebay search: "EM-PM FOCUS COILS RLF-028 & RLF-029"

also a thermionic source from an electron beam welder that might be
good for EBM:
ebay search: "Hairpin EB Welder Filaments HSD W-2 Leybold 12 Mil"

Sorry for not putting the ebay url's but I got an error from the url's
being too long.

cheers,
Jamie

[Jamie]

In our case, these would be the scan coils, they diverge the
beam on the 1 mil titanium sheet window that keeps the vacuum in
check but is thin enough for the electrons to pass/displace.

>
> focusing coil:
> ebay search: "EM-PM FOCUS COILS RLF-028 & RLF-029"

The focus coils sit above the scan horn and scan coil section.
They are responsible for confining the beam to a focal point on
the target, the titanium window in this case. THe focus is calibrated
to not produce the smallest point, otherwise, we'd be creating holes
and enjoy nice implosions from the near perfect vacuum.

> also a thermionic source from an electron beam welder that might be
> good for EBM:
> ebay search: "Hairpin EB Welder Filaments HSD W-2 Leybold 12 Mil"
>
> Sorry for not putting the ebay url's but I got an error from the url's
> being too long.
>
> cheers,
> Jamie

Our system uses 100-200 hz of electron beam scanning. Steering coils
are in place on some of them to avoid structures that tend to miss align
the beam that is not part of the scanning. Also, These are employed with
what is called wobble coils to oscillate the beam using 60 hz. This
prevents burning of the window material.

At the tips of each scan, on the outer edges of the window (horn),
Trapezoid peeks are added in the scan signal to help start the beam
back on its return path sooner. The theory here is, the beam spends more
time on the edges as it gets ready to scan on the return and thus may
heat that section of the window more. So, this wave form was
added to help this out however, from what I've seen over all, it does
not seem to do much for it. The signal at the peeks are getting lost in
the induction of the scan coils.

If you need more specifics, I can supply more but this was just a
general break down. As for cooling, we use chilled water that needs to
be circulated around the drift tube, aperture and scan horn sections.
With out this, thermo distortion in the structure would cause steering
problems and heat would cause vacuum leaks.

Jamie

[Bill Sloman]

On Feb 10, 12:46 am, Jamie

Somebody isn't doing his insulation right. If you expose 2-4MV
potential differences to air, you do get corona discharge and ozone
generation (which rots the lungs, and quite a few plastics) so
standard practice is to keep high voltage terminals well shrouded.

Cambridge Instruments electron microscopes didn't use more than 100kV,
and off-the-shelf high voltage connectors seemed to prevent corona
discharge and any stink of ozone. When I was a graduate student in
Melbourne, Alex Strojnik (the father) spent a year (around 1968) there
building a 600kV scanning transmission electron microscope, where the
high voltage parts were all contained in a tank of transfomer oil.

He talked about a French megavolt scanning transmission electron
microscope which relied on gas insulation and consequently had to be
huge - he called it a "Russian solution" which was apparently the
Croatian idiom for anything half-baked and clumsy.

--
Bill Sloman, Nijmegen

[Jamie]

Ozone is unavoidable in this process.. I don't think you have a
realistic view of what we have. But then again, I shouldn't
expect any more from you..

Does that smart ass reply remind you of some one ?


Jamie

[Jamie M]

Hi,

I guess it is necessary to have the focus and deflection coils aligned
very accurately coaxially to the electron beam to allow for a fine focus
and proper XY sweeps. For an application with extreme accuracy like an
electron microscope, what is the method for aligning the electron beam
to these coils? I think for CRT's they can use permanent magnets to
move the beam to the center of the coils.

cheers,
Jamie



>
> --
> Bill Sloman, Nijmegen

[Bill Sloman]

The electron-beam microfabricator used double deflection in two
successive sets of X and Y saddle coils to route the beam through the
same point at different angles - first you diverted the beam away from
the column axis through the deflection angle you wanted, then you bent
the beam back through twice the angle to get it going through the
column axis at the angle you wanted.

The Cambridge Instruments EBMF 10.5 which I worked on was a bit of a
dinosaur in covering a large - 4mm by 4mm - field (which is how we got
the job of making the holograms for Australia's plastic bank notes).
The beam steering electronics worked to 18-bits - writing 10-bit sub-
fields at the rate of one pixel per 100nsec but using a wider, slower
DAC to step between sub-fields, where we had to wait a millisecond or
two for everything to settle down. There were all kinds of high order
corrections for the (small) non-linearity of the deflection system -
it was a delightful example of fast precision analog electronics,
though it was a bit obvious that it had evolved rather than having
been created with the capability of doing everything that it could do
in the 10.5.

In particular, the corrections were applied serially, rather than in
parallel - there was a problem with the delay through the op amps and
DACs creating the correction terms. The system - never built - for
which I was in charge of the hardware development, was going to do
that in parallel and our best analog engineer designed a phase-linear
low-pass filter to delay the main signal enough to allow us to sum in
corrections terms generated in parallel.

--
Bill Sloman, Nijmegen

[Bill Sloman]

On Feb 10, 11:30 pm, Jamie

> Ozone is unavoidable in this process.. I don't think you have a
> realistic view of what we have. But then again, I shouldn't
> expect any more from you..
>
>    Does that smart ass reply remind you of some one ?

Ozone is obviously avoidable - no oxygen, no ozone, Sulphur
hexafluoride comes to mind.

http://en.wikipedia.org/wiki/Sulfur_hexafluoride

Avoiding it may be expensive for all sorts of reasons, none of which
you've seen fit to tell us about.

--
Bill Sloman, Nijmegen

[Tim Williams]

"Bill Sloman" <bill....@ieee.org> wrote in message
news:079535b5-fed9-4add...@f30g2000yqh.googlegroups.com...

> Ozone is obviously avoidable - no oxygen, no ozone, Sulphur
> hexafluoride comes to mind.
>
> http://en.wikipedia.org/wiki/Sulfur_hexafluoride

But then you get F2 and lower SFx's from the same ionization processes,
which are just as bad, if not worse.

[Jamie]

Sure, spread SF6 all over the place, being heavy like it is, it wouldn't
do the operators of the machinery very good.

No Bill, you don't understand. Sulfur Hexafluoride is used in the
vessel as an insulator and dielectric for the dyno stacks. And ozone is
lighter which wouldn't do us very good now would it?

Plus, how do you think the process is going to work with all that
insulated gas encompassing the intended target?

Guess you haven't seen what x-rays do to some gases either.

Close the E-book. The night time reading hour is over, time to return
to the bed. Be quite now, you may awake the guy in the bed next to yours.

Jamie

[Jamie]

Tim Williams wrote:

> "Bill Sloman" <bill....@ieee.org> wrote in message
> news:079535b5-fed9-4add...@f30g2000yqh.googlegroups.com...
>
>> Ozone is obviously avoidable - no oxygen, no ozone, Sulphur
>> hexafluoride comes to mind.
>>
>> http://en.wikipedia.org/wiki/Sulfur_hexafluoride
>
>
> But then you get F2 and lower SFx's from the same ionization processes,
> which are just as bad, if not worse.
>
> Tim
>

If you could ever see the cavity walls of what the X-rays and ozone do
with moister alone. Nice skin etching material.. Makes nasty acidic
slugs that will give you skin problems very quickly if not careful.

One of the reasons to not inhale large amounts of ozone is due to the
reaction in your lungs. It mixes with the moister in the tissue there by
created a acid that can burn your lungs. The X-rays in the process just
adds to the problem while working around the residue that has been
created by it.

Jamie

[Jamie M]

Hi,

Thanks, that sounds pretty interesting! :) I was wondering about
correcting non-linearities of the deflection and focus coils. I guess
there are several important variables including XY desired position of
the beam, focus diameter and beam intensity. If beam intensity and
focus diamter are constants that makes it a lot easier to calibrate the
XY positions probably.

[Jamie M]

Hi,

I will take as much info as you are willing or allowed to share about
the system, since I don't know much about electron guns its nice to
learn from practical examples! For a small scale electron beam melting
application, how much power do you think is required for the thermionic
source as well as the electron accelerator source? Also the focus and
scan coils probably use very little power I am guessing? Basically I
am wondering how much overall power is required to do x watts of
heating/melting. With a single focus coil and a single deflection coil,
would it be possible to generate an electron beam with a 1mm or
smaller beamwidth that is capable of melting metals for EBM? Also is
the beam focal point set before the part being melted so that electrons
cross at a point in space like an optical lens making an image on a
scree does? Also for the main structure of the electron gun is this
correct:

top: thermionic source produces free electrons with zero net velocity

cathode: small diameter metal circle positioned below thermionic source
(not sure about this one)

anode plate with central small hole: electrons accelerate from cathode
to anode and some go through the hole forming an electron beam while the
rest are absorbed back into the metal anode

focus coil: electron beam from anode hole is focused by a solenoid coil
in the same way as a convex lens

XY deflection coil: electron beam angle is changed to aim at desired XY
coordinate

Also would it be possibly better to use electrostatic focus/deflection
for an EBM application? I have heard that electromagnetic is more
linear and more accurate.

Thanks for any additional info, tips etc! I am dreaming of making a
bench top electron beam melter 3D printer now :)

cheers,
Jamie



>
> Jamie
>

[Bill Sloman]

On Feb 11, 2:12 am, "Tim Williams" <tmoran...@gmail.com> wrote:
> "Bill Sloman" <bill.slo...@ieee.org> wrote in message

>
> news:079535b5-fed9-4add...@f30g2000yqh.googlegroups.com...
>
> > Ozone is obviously avoidable - no oxygen, no ozone, Sulphur
> > hexafluoride comes to mind.
>
> >http://en.wikipedia.org/wiki/Sulfur_hexafluoride
>
> But then you get F2 and lower SFx's from the same ionization processes,
> which are just as bad, if not worse.

Probably not. SF5 can self-combine to S2F10, but that's pretty
reactive - as is fluorine - and you will get back to SF6 pretty
quickly

http://en.wikipedia.org/wiki/Disulfur_decafluoride

Ozone can self-combine to reform oxygen

2 O3 -> 2O2

but you've got to get the orientation exactly right, and two-atom
molecules don't have all that many quantum states into which to dump
the excess energy released in the reaction. SFn contains rather more
atoms and has a lot more vibrational modes, so it tends react faster.

--
Bill Sloman, Nijmegen

[Jamie]

http://www.iba-industrial.com/accelerators/dynamitron-operating-principles

Read that if you have not already seen this material.

The power supply is a cold cathode tube oscillator of a armstrong type
of configuration. It operates at 100khz into what is called a pie
transformer. This transformer is made of litz wire. The DC supply is
variable 0..15kVDC using a 3 phase SCR firing bridge.

Although the tube is capable of 350kwatts, we generally don't exceed
250kW.

The pie transformer is inside a vessel that is pressurized around 90
Lbs of SF6 gas, which is a insulating gas and also makes up as part of
the dialectic for the corona rings. These rings sit very close to the
wall of the vessel and serve as the capacitors in a cockcroft walton
full wave rectifier.

At the end of the stack, there still is a ripple in the rectifiers due
to the nature in how it works, this ripple is them used as a 100 kHz
source into a step down RF transformer which is used to drive the heater
element. There are 2 large stats on the end nose that controls the beam
current, which is actually done by setting the heater voltage.

Because our systems do not use a motorized generator down the center
to operate the heaters. The main stat is what we called the preset, this
is the minimum heater voltage derived from the selected HV voltage at the
end of the rectifier stack. The other stat is a variable trim that ramps
up in proportion to the increase or more voltage as the process material
gets ramp up in speed to maintain dosimetry.


I could go on but I think this is enough to start with.. :)

Jamie