Re: ATM Re: [VacuumX] 24" vacuum chamber is getting closer :)

[Dominic-Luc Webb <dlwebb@canit.se>]

> Very true!  Unfortunately I can barely manage the fabrication of a simple O-Ring
> implementation, the machining for copper seals is well beyond my grasp at this
> time ;)


I have not yet tried copper but am considering it.


> > I would be interested in your solution for firing up the tungsten
> > filaments, presumably you used (green) TIG welding electrodes.
>
> For the 12" chamber I have been using 0.032" tungsten wire from smallparts.com.
> I plan a similar arrangement for the 24" chamber, but with 3 times as many
> filaments for a total of 18 in two concetric rings.
>
> Have you calculated the wattage used on your light dimmer, or is the dimmer just
> used as a signal source?  I'm using 6 filaments, each 40 mm length, in a series
> / parallel arangment, and consuming about 500 watts of power at a few volts and
> several hundred Amps.  The dimmer switch is rated at 600 watts so I am pushing
> it a little...


For your system, the wire appears to be 0.0813 cm diameter (area is then
0.0052 cm^2). The resistivity of Tungsten is 5.51e-6 Ohm*cm. This should
works out to about 1.06e-3 Ohm*cm. For your entire series circuit:

1.06e-3 * 6 * 4.0 = 0.025 Ohm  (intuitively, this sounds about right)

I presume connections between these filaments have essentially negligible
resistance. Hence, there is not much resistance to work with, as per my
original comment. The smallest TIG electrodes I can find are 1.0 mm, so
resistance is similar.

I am not sure what to say about your dimmer vs my dimmer and dimmers in
general. You discuss a few volts and hundreds of amps, which makes me
believe you have done something very different from me (voltage-current
conversion). Here is what I have done and how dimmers work as I understand
them.

A typical light dimmer (120 or 230 AC mains input) functions by
interupting the current so that the filament does not experience the
entire AC phase cycle. Consequently, the filament only experiences a
portion of the cycle and thus a portion of the energy. Because a
triac/diac pair is used, the current to the filament is literally
switched off, completely. This is a major departure from other techniques
that act by damping a continuous current or voltage.

There is no deliberate effort made in these circuits to convert voltage
to current although there are inherent current surges and timing
constrains in these circuits which cause small current rises, often
generating RFI. Filter coils (chokes) are added in an attempt to slow
down these rapid surges, and the presence of such coils also makes
possible further intereferences. The interuption and variance in mean
total energy permitted to the filament is achieved by adjusting the resistance
coupled to a capacitor which drives a triac. Triacs are chosen with at
least a couple times higher current rating than intended for the
prewarmed running circuit due to things like surges, etc, amongst these
is that caused by the cold resistance of tungsten which is 17 times
lower than once the filament is heated. The BT139 triac I have chosen is
rated at 230 VAC @ 16 Amp. I do not expect to operate at 230*16=3680
Watts. Indeed, I am not planning (I am an optimist, BTW) to even use half
this many Watts, thus justifying the use of this triac. Because this
dimmer circuit (designed for actual continuous operation to dim about
1000 Watts) acts by limiting the total energy reaching the filament
and becasue we do not need or even want the maximum possible heating
of the electrode, I figure this might work for aluminizing. I think
the real stress on the dimmer does not come from permitting the
full phase to pass through, but rather, by increased dimming. In testing
my original circuit, a couple components burned and this happened with
strong dimming. To this end, one needs quality capacitors (X-class) and
filter choke (should carry the full operating load). The major headache
of this circuit is dealing with the very low resistance calculated above
as 0.025 Ohms. A typical 230 VAC 100 Watt bulb has a cold resistance
of 35 Ohms. I need to think a bit on how many Watts are actually
required to melt and then vaporize the aluminum, but likely considerably
less than we are presently using. We both seem to be using the "bombs-away
Pentagon approach" at present. We both have the challenge of dealing with
hardly a 1/4 Ohm for the complete series, which is not hugely different
from the cables leading to the filament. One can sarcastically question
whether one is heating the filament, the wires to the filament, or the
dimmer. My design uses generous PCB space to carry this current, BTW. I
was not impressed with the commercial dimmers and would not want to use
them at their full ratings.

This brings us back to your fragile filaments. Perhaps much of that
breaking is due to extreme temperatures you are using. As per our earlier
discussions about sublimation, a lower temperature might be desirable.
Another issue is that you must recall that dimmers (the type I refer
to) are causing oscillatory phenomena which can be transduced into
filament vibration. This could be minimized by changing components,
changing the "dimming level" or changing the filament structure or the
nature of the support for the filament so as to dampen vibrations. I believe
you have an oscilloscope and you may be able to detect some of this
at the 2nd or 3rd harmonic for your line frequency (60 Hz*2 and 60Hz*3).
Not sure, just a thought.

I will do more work on my dimmer and electrodes before posting a web page
with specific circuit board and part numbers. I will now also consider a
120 VAC version. More ideas are welcome.


Dominic-Luc Webb