Smashed Glass Coaster #2: Mirror Base FTW

Glass fragments bedded on clear epoxy atop a white base looked OK, albeit minus most of their glitter due to epoxy filling their cracks:

Glass Coaster - fragment edge detail
Glass Coaster – fragment edge detail

Filling the cracks with black epoxy makes them stand out:

Smashed Glass vs epoxy - magnified comparison
Smashed Glass vs epoxy – magnified comparison

So I assembled a coaster from shattered glass in a clear surround with black epoxy atop a mirror base:

Smashed Glass Coaster 2 - mid-layer glass pour
Smashed Glass Coaster 2 – mid-layer glass pour

Each fragment sits on a blob of black epoxy that eventually oozed out to fill the gap between the mirror and the transparent layer. You can see the oozing start around the two fragments in the upper left.

A top layer of black acrylic sits flush with the upper surface of the glass, seen here with the protective paper in place before pouring black epoxy into the gap around the perimeter of each fragment:

Smashed Glass Coaster 2 - masked top
Smashed Glass Coaster 2 – masked top

Peeling the paper away exposes an almost perfect surface, with the epoxy forming a slight curve between the black acrylic and the glass:

Smashed Glass Coaster 2 - overview
Smashed Glass Coaster 2 – overview

The mirror doubles the number of glass cuboids and their glittery gaps:

Smashed Glass Coaster 2 - fragment detail
Smashed Glass Coaster 2 – fragment detail

All in all, it turned out well, but the epoxy pouring and leveling is tedious.

It might be possible to assemble a coaster upside-down, with the black layer stuck to something like Kapton tape and the fragments carefully aligned in their openings to make the entire top surface a plane. The tape should keep the epoxy from oozing out of the gaps, although a perfect seal may be impossible.

Then fill the gaps with black epoxy, lay the clear middle layer in place, run a dollop of epoxy on each fragment, lay the mirror in place, and hope there’s enough epoxy to fill all the gaps and not enough to make a mess around the perimeter.

With a bit of luck, that wouldn’t require so much hand finishing.

The next coaster must have a perimeter shrinkwrapped around the fragments, if only to break the low-vertex-count polygon tradition.

Rounded Petal Acrylic Coaster

Having gotten the rounded-petal pattern generator working, applying it to acrylic sheets seemed reasonable:

Cut Acrylic Coaster - top cleaned
Cut Acrylic Coaster – top cleaned

The petals stand slightly proud of the black top frame, as the colored sheets were marginally thicker than the black sheet, but it looks OK in person. They’re all epoxied to a transparent base plate, so the bottom view is pretty much the same:

Cut Acrylic Coaster - bottom
Cut Acrylic Coaster – bottom

Because the bottom is perfectly smooth, I think it looks better than the top, which shows irregularities around the petals where the epoxy didn’t quite fill the gaps. There is one small bubble you won’t notice if I don’t tell you about it.

I laid a small bead of epoxy around the perimeter of the base, laid the black frame in place, ran a bead along the midline of each petal shape plus a drop in the round part, laid the petals in place, and hoped I didn’t use too much epoxy. It turned out all right, with only a few dribbles down the edge that wiped off easily enough.

I peeled the protective plastic off the top while the epoxy was still tacky, which pulled far too many fine filaments across the surface:

Cut Acrylic Coaster - frayed top
Cut Acrylic Coaster – frayed top

After the final cure, I managed to scrape most of them off with a thumbnail; I hope to never make that mistake again.

As you might expect, acrylic plastic’s pure saturated colors wipe the floor with Sharpie-scribbled white chipboard:

Chipboard coaster - rounded petals - front vs back cut
Chipboard coaster – rounded petals – front vs back cut

The black frame makes the whole thing overly dark, so the next attempt should use white or perhaps a transparent layer atop a mirror base.

OMTech 60 W Laser: Replacement HV Power Supply Waveforms

While I had the hatch open, I thought it would be interesting to look at the HV supply’s current waveforms:

HV laser power supply - current probe setup
HV laser power supply – current probe setup

The Tek current probe over on the right measures return current through the cathode wire, the point in the circuit where you might be tempted to install an ordinary analog (moving-coil) panel milliammeter, oriented so (conventional) current returning from the tube will produce a positive voltage.

Unfortunately, an analog meter isn’t up to displaying anything meaningful for this nonsense:

HV laser power supply - 5 mA-div - 50 ms 10 pct pulse
HV laser power supply – 5 mA-div – 50 ms 10 pct pulse

Admittedly, that’s a 50 ms pulse, during which an analog meter would barely twitch. The vertical scale is 5 mA/div, so the highest peaks exceed 35 mA, more than twice the tube’s recommended “14-15 mA”.

A closer look at the pulse startup waveform:

HV laser power supply - 5 mA-div - 50 ms 10 pct pulse - detail
HV laser power supply – 5 mA-div – 50 ms 10 pct pulse – detail

It sure looks like the chaotic current through a forced neon-bulb relaxation oscillator. Remember neon bulbs?

An even closer look:

HV laser power supply - 5 mA-div - 50 ms 10 pct pulse - tight detail
HV laser power supply – 5 mA-div – 50 ms 10 pct pulse – tight detail

That’s at 10% PWM, close to the threshold below which the laser just won’t fire at all. The power supply must ramp up to produce enough voltage to fire the tube while simultaneously limiting the current to prevent the discharge from sliding down the negative resistance part of its curve.

Apparently this supply isn’t quite up to the task.

A 10 ms pulse at 50% PWM gives the supply enough time to stabilize the current:

HV laser power supply - 5 mA-div - 10 ms 50 pct pulse
HV laser power supply – 5 mA-div – 10 ms 50 pct pulse

The 14-ish mA at the tail end of the pulse (note the baseline offset) matches my previous 13 to 14 mA measurements as closely as seems reasonable. That 2 ms of hash on the leading edge suggests the start of each cut or engraving line will be a bit darker than you might expect.

Another 10 ms pulse, this time at 99% PWM:

HV laser power supply - 5 mA-div - 10 ms 99 pct pulse
HV laser power supply – 5 mA-div – 10 ms 99 pct pulse

The peak 24-ish mA matches the previous measurements. Note that the peaks in all the previous pictures exceed the 99% PWM current level.

AFAICT, all PWM values below about 25% produce equivalent results: random current spikes with unpredictable timing and amplitude. Changing the PWM value does not affect the (average) tube current or laser output power in any predictable way.

Some samples to illustrate the point, starting with a different 50 ms pulse at 10% PWM than the first one up above:

HV laser power supply – 5 mA-div – 50 ms 10 pct

A 50 ms pulse at 15% PWM:

HV laser power supply - 5 mA-div - 50 ms 15 pct
HV laser power supply – 5 mA-div – 50 ms 15 pct

A 50 ms pulse at 20% PWM:

HV laser power supply - 5 mA-div - 50 ms 20 pct
HV laser power supply – 5 mA-div – 50 ms 20 pct

A 50 ms pulse at 25% PWM:

HV laser power supply - 5 mA-div - 50 ms 25 pct
HV laser power supply – 5 mA-div – 50 ms 25 pct

Now, that last one is different. After the hash during the first 8 ms or so, the power supply actually produces a stable 5 mA beam current, which is roughly what I measured using the power supply’s meter.

However, the other three are pretty much identical: the 10% PWM pulse does not delivers half as energy as the 20% PWM pulse. The waveforms may be different, but not in a meaningful or consistent way: the two 50 ms 10% pulses are different, but you’d (well, I’d) have trouble separating them from the 20% pulse.

To summarize:

  • The first several millisconds of any pulse will consist of randomly distributed spikes with very large tube currents.
  • For PWM values greater than 25%, the tube current will settle down to the corresponding current after 5 to 10 ms. Before the current settles down, the tube will be firing those random spikes.
  • For PWM values less than 25%, the tube current never settles down: the entire pulse, no matter how long, will be short, high-intensity spikes, without a consistent DC-ish level.

No matter what an analog meter might show.

I have no way to know if this power supply is defective, but I’ll certainly ask …

OMTech 60 W Laser: Replacement HV Power Supply

The original HV power supply in the OMTech 60 W laser went casters-up just barely inside OMTech’s six month tube-and-supply warranty period. For the record, the laser controller reports this status info since mid-March:

Laser Stats - replacement supply
Laser Stats – replacement supply

I think the Total job laser on time line says the power supply failed after firing the laser for a little over eight hours. The OMTech manual says the laser tube should last 1000 to 2000 hours (low vs high power), which suggests I should stock up on power supplies.

Its replacement just arrived:

OMTech replacement HV supply
OMTech replacement HV supply

It (bottom) seems to be a knockoff of the original ZYE Laser supply (top), with a similar model number and a “serial number” resembling a date from last year. All the connectors matched up, which isn’t too surprising.

The three most interesting inputs:

  • L = controller’s active-low L-ON enable output
  • IN = controller’s PWM output
  • P = jumper to G (circuit ground) — not water flow sensor

Also note the two AC power-line terminals directly adjacent to the TEST button, then consider insulation and stand-off distances before poking the button with your index finger.

The power supply has a digital current meter, so I plotted output current against PWM input:

Laser Power Supply - mA vs PWM - overview
Laser Power Supply – mA vs PWM – overview

Taking more points at the low end, with vertical bars indicating single-digit flicker on the meter:

Laser Power Supply - mA vs PWM - 0 to 20 PWM
Laser Power Supply – mA vs PWM – 0 to 20 PWM

I have little reason to believe the meter reading indicates the true current with any accuracy and I know CO₂ laser output power does not scale linearly with the current.

But it’s cutting again, which is a step in the right direction.

Smashed Glass vs. Epoxy

Just to see what happens, I laid some smashed glass in puddles of epoxy:

Smashed Glass vs epoxy - samples
Smashed Glass vs epoxy – samples

Backlighting with the LED light pad reveals more detail:

Smashed Glass vs epoxy - backlit samples
Smashed Glass vs epoxy – backlit samples

The chunk on the left is the proof-of-concept shot glass coaster with a form-fit black acrylic mask atop a clear epoxy layer on a clear acrylic base. The chunk at the top is raw shattered glass fresh from the pile. The two chunks on teardrop acrylic scraps are bedded in transparent black and opaque black tinted epoxy.

A look through the microscope at all four, laid out in that order, with the contrast blown out to emphasize the grain boundaries:

Smashed Glass vs epoxy - magnified comparison
Smashed Glass vs epoxy – magnified comparison

You may want to open the image in a new tab for more detail.

The raw chunk has air between all its cuboids, so it’s nicely glittery. All the others have much of their air replaced by epoxy.

Clear epoxy produces an essentially transparent layer where it fills the gaps, because its refractive index comes close enough to the glass. The stretched contrast makes the gaps visible again, but the backlit image shows the unassisted eyeball view.

Transparent black dye sounds like an oxymoron, but it fills the gaps with enough contrast to remain visible. The overall chunk is not particularly glittery, but it’s OK.

Opaque black dye produces a much darker tint; the slightly tapered thin layer between the glass and acrylic (the small white circles are air bubbles) cuts down on the transmitted light. The gaps remain nearly as prominent as in the air-filled chunk, although with very little glitter.

Bedding the glass in epoxy against an acrylic sheet should reduce its tendency to fall apart at the slightest provocation, although the proof-of-concept poured coaster showed the epoxy must cover the entire edge of the glass sheet to bond all the slivers in place.

OMTech 60 W Laser: Failed HV Power Supply

Setting up a piece of MDF and hitting the Frame button produced a lightly scorched line around the part perimeter, plus a slightly diagonal track leading from / to the Home position in the far right corner:

Fire while framing tracks
Fire while framing tracks

Doing another pass with LightBurn’s rubber-band frame produced the faint dotted circle.

Huh. Didn’t useda do that.

The laser should not fire while framing and, having just installed LightBurn’s 1.2.01 update, suspicion instantly fell on the most recently changed thing.

Which turned out not to be the case, as LightBurn’s tech support pointed out:

This is generally an indication of a failed high-voltage power supply, not a software issue.

OMTech’s support requested a video of the equipment bay, which didn’t seem like a useful way to convey the situation. Instead, I sent pix.

This picture shows the status of the 60 W laser power supply while the laser is incorrectly firing:

OMTech 60W Laser - uncommanded framing fire
OMTech 60W Laser – uncommanded framing fire

The power supply has two LEDs on what looks like, but is not, an Ethernet jack near the bottom:

  • Orange P LED: good water flow
  • Green L LED: controller’s PWM signal

The LASER orange LED near the top turns on when the HV output is active and the laser should be firing.

In this case, L LED is off and the LCD shows “Laser signal OFF”, but the LASER LED is on and the LCD shows 2 mA beam current: the laser beam is ON, even though the controller has not activated the PWM signal.

Not only that, but I discovered the laser would fire while framing even with the lid up and the “safety interlock” sensor active.

Totally did not expect that.

For comparison, the power supply status during a manual pulse at 49% power:

OMTech 60W Laser - manual pulse 49%
OMTech 60W Laser – manual pulse 49%

In that case, the L LED shows the PWM signal is active, the LASER LED is on, and the LCD shows 14 mA of current to the tube. That’s how it should work.

Although the function of the TEST button seems very lightly documented, pressing it did not turn on the output (the LASER LED is off), despite lighting the L LED:

OMTech 60W Laser - Test button pressed
OMTech 60W Laser – Test button pressed

OMTech confirmed my suspicion:

We are afraid that the laser power supply is defective

A replacement should arrive in a few days.

Protip: always practice laser eye safety.

Coaster Generator: Rounded Petals

Making a coaster with petals from the NBC peacock turned out to be trickier than I expected:

Chipboard coaster - rounded petals
Chipboard coaster – rounded petals

Protracted doodling showed that I cannot math hard enough to get a closed-form solution gluing a circular section onto the end of those diverging lines:

Chipboard coaster - rounded petal geometry doodle
Chipboard coaster – rounded petal geometry doodle

However, I can write code to recognize a solution when it comes around on the guitar.

Point P3 at the center of the end cap circle will be one radius away from both P2 at the sash between the petals and P4 at the sash around the perimeter, because the circle will be tangent at those points. The solution starts by sticking an absurdly small circle around P3 out at P4, then expanding its radius and relocating its center until the circle just kisses the sash, thus revealing the location of P2:

t1 = tan(PetalHA);
sc = (Sash/2) / cos(PetalHA);

<< snippage >>

P3 = P4;        // initial guess
r = 1.0mm;      // ditto
delta = 0.0mm;
do {
  r += sin(PetalHA) * delta;
  P3.x = P4.x - r;
  dist = abs(P3.x * t1 - sc) / sqrt(pow(t1,2) + 1);
  delta = dist - r;
  message("r: ",r,"  delta: ",delta);
} while (abs(delta) > 0.001mm);

P2 = [P3.x - r*sin(PetalHA),r*cos(PetalHA)];

The dist variable is the perpendicular distance from the sash line to P3, which will be different than the test radius r between P3 and P4 until it’s equal at the kissing point. The radius update is (pretty close to) the X-axis difference between the two, which is (pretty close to) how wrong the radius is.

As far as I can tell, this will eventually converge on the right answer:

r: 1.0000mm  delta: 13.3381mm
r: 6.1043mm  delta: 6.2805mm
r: 8.5077mm  delta: 2.9573mm
r: 9.6394mm  delta: 1.3925mm
r: 10.1723mm  delta: 0.6557mm
r: 10.4232mm  delta: 0.3087mm
r: 10.5414mm  delta: 0.1454mm
r: 10.5970mm  delta: 0.0685mm
r: 10.6232mm  delta: 0.0322mm
r: 10.6355mm  delta: 0.0152mm
r: 10.6413mm  delta: 0.0071mm
r: 10.6441mm  delta: 0.0034mm
r: 10.6454mm  delta: 0.0016mm
r: 10.6460mm  delta: 0.0007mm

Obviously, efficiency isn’t a big concern here.

Having found the center point of the end cap, all the other points fall out easily enough and generating the paths follows the same process as with the simple petals. The program performs no error checking and fails in amusing ways.

As before, laser cutting the chipboard deposits some soot along both sides of the kerf. It’s noticeable on brown chipboard and painfully obvious on white-surface chipboard, particularly where all those cuts converge toward the middle. I applied low-tack blue masking tape as a (wait for it) mask:

Chipboard coaster - tape shield
Chipboard coaster – tape shield

Whereupon I discovered the white surface has the consistency of tissue paper and removing the tape pretty much peels it right off:

Chipboard coaster - white surface vs tape
Chipboard coaster – white surface vs tape

Putting the chipboard up on spikes and cutting it from the back side, with tabs holding the pieces in place (so they don’t fall out and get torched while cutting the next piece), should solve that problem.

In the meantime, a black frame conceals many issues:

Chipboard coaster - rounded petals - front vs back cut
Chipboard coaster – rounded petals – front vs back cut

I must up my coloring game; those fat-tip markers just ain’t getting it done.

The GCMC and Bash source code as a GitHub Gist:

// Round Petals Test Piece
// Ed Nisley KE4ZNU
// 2022-07-17 Coasters with round-end petals
layerstack("Frame","Petals","Rim","Base","Center","Tool1"); // SVG layers map to LightBurn colors
//-----
// Library routines
include("tracepath.inc.gcmc");
include("tracepath_comp.inc.gcmc");
include("varcs.inc.gcmc");
include("engrave.inc.gcmc");
FALSE = 0;
TRUE = !FALSE;
//-----
// Command line parameters
// -D various useful tidbits
// add unit to speeds and depths: 2000mm / -3.00mm / etc
if (!isdefined("OuterDia")) {
OuterDia = 100.0mm;
}
if (!isdefined("CenterDia")) {
CenterDia = 0.0mm;
}
if (!isdefined("NumPetals")) {
NumPetals = 6;
}
if (!isdefined("Sash")) {
Sash = 5.0mm;
}
// Petal values
PetalAngle = 360.0deg/NumPetals; // subtended by inner sides
PetalHA = PetalAngle/2;
PetalOD = OuterDia - 2*Sash;
PetalID = CenterDia + 2*Sash;
PetalOAL = OuterDia/2 - Sash - (Sash/2)/sin(PetalHA);
//message("petalOAL: ",PetalOAL);
P4 = [PetalOD/2,0.0mm];
// Find petal vertices
P0 = [(Sash/2) / sin(PetalHA),0.0mm];
t1 = tan(PetalHA);
sc = (Sash/2) / cos(PetalHA);
if (P0.x < PetalID/2) {
a = 1 + pow(t1,2);
b = -2 * t1 * sc;
c = pow(sc,2) - pow(PetalID/2,2);
xp = (-b + sqrt(pow(b,2) - 4*a*c))/(2*a);
xn = (-b - sqrt(pow(b,2) - 4*a*c))/(2*a);
y = xp*t1 - sc;
if (FALSE) {
message("a: ",a);
message("b: ",b);
message("c: ",c);
message("p: ",xp," n: ",xn," y: ",y);
}
P1 = [xp,y];
}
else {
P1 = P0;
}
P3 = P4; // initial guess
r = 1.0mm; // ditto
delta = 0.0mm;
do {
r += sin(PetalHA) * delta;
P3.x = P4.x - r;
dist = abs(P3.x * t1 - sc) / sqrt(pow(t1,2) + 1);
delta = dist - r;
message("r: ",r," delta: ",delta);
} while (abs(delta) > 0.001mm);
P2 = [P3.x - r*sin(PetalHA),r*cos(PetalHA)];
PetalWidth = 2*r;
if (FALSE) {
message("P0: ",P0);
message("P1: ",P1);
message("P2: ",P2);
message("P3: ",P3);
message("P4: ",P4);
}
// Construct paths
PetalPoints = {P1,P2};
OutArc = varc_cw([P2.x,-P2.y] - P2,-r);
OutArc += P2;
PetalPoints += OutArc;
if (P0 != P1) {
PetalPoints += {[P1.x,-P1.y]};
InArc = varc_ccw(P1 - [P1.x,-P1.y],PetalID/2);
InArc += [P1.x,-P1.y];
PetalPoints += InArc;
}
else {
PetalPoints += {P0};
}
//--- Lay out the frame
linecolor(0xff0000);
layer("Frame");
if (CenterDia) {
goto([CenterDia/2,0mm]);
circle_cw([0mm,0mm]);
}
repeat(NumPetals;i) {
a = (i-1)*PetalAngle;
tracepath(rotate_xy(PetalPoints,a));
}
goto([OuterDia/2,0]);
circle_cw([0mm,0mm]);
//--- Lay out internal pieces for oriented cutting
// baseplate
layer("Base");
relocate([OuterDia + 2*Sash,0]);
goto([OuterDia/2,0]);
circle_cw([0mm,0mm]);
// central circle
if (CenterDia) {
layer("Center");
relocate([OuterDia/2 + Sash,-(OuterDia - CenterDia)/2]);
goto([CenterDia/2,0mm]);
circle_cw([0mm,0mm]);
}
// petals
layer("Petals");
repeat(NumPetals;i) {
org = [PetalWidth/2 - OuterDia/2,-(OuterDia + Sash)];
relocate([(i-1)*(PetalWidth + Sash) + org.x,org.y]);
tracepath(rotate_xy(PetalPoints,90deg));
}
// Debugging by printf()
if (FALSE) {
layer("Tool1");
linecolor(0xff1f00);
goto([Sash/2,0mm]);
circle_cw([0mm,0mm]);
goto(P0);
circle_cw([0mm,0mm]);
goto([0,0]);
move([OuterDia/2,0]);
goto([0,0]);
move(OuterDia/2 * [cos(PetalHA),sin(PetalHA)]);
goto(P2);
move_r([0,-PetalWidth/2]);
}
#!/bin/bash
# Round petals test piece
# Ed Nisley KE4ZNU - 2022-07-17
Flags='-P 4 --pedantic' # quote to avoid leading hyphen gotcha
SVGFlags='-P 4 --pedantic --svg --svg-no-movelayer --svg-opacity=1.0 --svg-toolwidth=0.2'
# Set these to match your file layout
ProjPath='/mnt/bulkdata/Project Files/Laser Cutter/Coasters/Source Code'
LibPath='/opt/gcmc/library'
ScriptPath=$ProjPath
Script='Round Petals.gcmc'
[ -z "$1" ] && petals="6" || petals="$1"
fn=RoundPetals-$petals.svg
echo Output: $fn
gcmc $SVGFlags \
-D "NumPetals=$petals" \
--include "$LibPath" \
"$ScriptPath"/"$Script" > "$fn"
view raw roundpetals.sh hosted with ❤ by GitHub