bpark1000

Beware. Running a xenon flash tube like this for any time can quickly destroy it, unless the current is INSANELY low. (Some camera strobes would "hang on" this way, & were destroyed in minutes if allowed to continue.) You will know if damage is occurring if a black deposit forms inside the tube end near the cathode. Shut off the power, wait for capacitors to discharge, & feel the cathode end of the tube for heat. Do you know the current & voltage the tube is getting? Also, make sure you have the proper polarity on the tube. The cathode (minus) is the side with the wad of gauze inside (NOT the pin). Starting will be easier if the metal clip is tied to the positive supply. Keep your fingers away from the metal clip!

When lighting "holey" structures from the inside, you will get brighter results by painting all the inside surfaces of the wood white. The light that misses the holes will bounce around inside until it finds another hole. (You could also laminate a thin layer of white veneer on one side of your stock for the sides before you cut it up, & under the top & on the inside of the bottom.) For alignment templates, cut them from corrugated cardboard. Put a (temporary) post coming up from the bottom in the center. This post can be glued to the lowest cardboard circle. Each subsequent circle of cardboard can have a hole to fit over the post. That way, every layer is aligned. On the last layer, apply glue & forget the staples! After all is positioned, put board…

When lighting "holey" structures from the inside, you will get brighter results by painting all the inside surfaces of the wood white. The light that misses the holes will bounce around inside until it finds another hole. (You could also laminate a thin layer of white veneer on one side of your stock for the sides before you cut it up, & under the top & on the inside of the bottom.) For alignment templates, cut them from corrugated cardboard. Put a (temporary) post coming up from the bottom in the center. This post can be glued to the lowest cardboard circle. Each subsequent circle of cardboard can have a hole to fit over the post. That way, every layer is aligned. On the last layer, apply glue & forget the staples! After all is positioned, put board & weight on top. If the glue swells up & pieces move out of place before you can place them all, place & glue only some (put dummies on the other side), then apply the weight & dry. Then glue the rest in another pass. For strength, I would glue a thin ring of wood on top of the perforated part (that is a permanent part of the sides). If you do that, you can staple everything except that last ring.

Beware that your cutting board may have some "cold shuts" (places where the plastic comes together but does not completely fuse). These can form "trapped volumes" where food & bacteria can accumulate. This can also happen when the plastics are not totally clean, having coatings that can interfere with proper fusion, or by trapped air between the pieces. You want to make sure you have removed ALL the stuff the containers had in them (such as liquid detergent, & labels, not for toxicity, but for interfering with the fusion). When you plane the part, you could expose internal cavities that you can't see from the outside before the planing operation. To test: use a "dye penetrant test". Mix up concentrated solution of food dye &…

Beware that your cutting board may have some "cold shuts" (places where the plastic comes together but does not completely fuse). These can form "trapped volumes" where food & bacteria can accumulate. This can also happen when the plastics are not totally clean, having coatings that can interfere with proper fusion, or by trapped air between the pieces. You want to make sure you have removed ALL the stuff the containers had in them (such as liquid detergent, & labels, not for toxicity, but for interfering with the fusion). When you plane the part, you could expose internal cavities that you can't see from the outside before the planing operation. To test: use a "dye penetrant test". Mix up concentrated solution of food dye & soak board in it. Pull it out & wipe it dry. Sift white powder (can be flour, corn starch, talc) in thin layer, just covering underlying color. Then lightly mist with water. Where crack is in board, the color of the dye will come out & stain the powder coating.

It is not good practice to connect LEDs in parallel without ballast resistors. One LED can get hot & "hog" all the current, leading to its destruction. The resistor you used on each board (7.5 ohms?) should be multiplied by the number of LEDs (10) on each triangle, & resistors of that value (75 ohms?) should be connected in series with EACH LED, & the main resistor eliminated. (Resistors seem a little low, but you need to do current calculation to be sure (Vbat - VLED)/I =R

Your schematic is somewhat confusing. Is it for one triangle only? (I see 30 LEDs & that's number on one triangle). Are there 3 FETs, or only 1 in your entire circuit?I see a total of 5 resistors in your schematic: R3 & R4 for the gate-drive, & R6, R7,& R8 (I assume one on each triangle?) in series with the LEDs. In that same schematic I see 3 groups of 10 LEDs, all connected DIRECTLY IN PARALLEL.Quoting from your statement: "The attached Schematic is a simple one to understand, there's a total of 30 LEDs connected in parallel, they are all driven by this Mosfet as a switch setup which is then connected with the attiny MCU."There should be A total of 30 resistors on ONE triangle, ONE IN SERIES WITH EACH LED. Am I missing something? Do these LEDs have resis…

Your schematic is somewhat confusing. Is it for one triangle only? (I see 30 LEDs & that's number on one triangle). Are there 3 FETs, or only 1 in your entire circuit?I see a total of 5 resistors in your schematic: R3 & R4 for the gate-drive, & R6, R7,& R8 (I assume one on each triangle?) in series with the LEDs. In that same schematic I see 3 groups of 10 LEDs, all connected DIRECTLY IN PARALLEL.Quoting from your statement: "The attached Schematic is a simple one to understand, there's a total of 30 LEDs connected in parallel, they are all driven by this Mosfet as a switch setup which is then connected with the attiny MCU."There should be A total of 30 resistors on ONE triangle, ONE IN SERIES WITH EACH LED. Am I missing something? Do these LEDs have resistors built-in? (That is very unusual for discrete LEDs). The setup You have controls the TOTAL current to ALL the LEDs, but does not guarantee EVEN DISTRIBUTION of that current amongst the 30 LEDs. Forward voltage of an LED is not an absolute spec; they can vary, especially for different batches. The ones with lower voltage will take a larger share of the current & will heat up more. This can lead to "thermal runaway", because a warmer LED's forward voltage drops, making it take an even larger share of the current, making it even hotter...

I have seen this done somewhat differently. You have 2 containers, one heavily insulated below with pump, the other insulated on top & sides The water is pumped up to the upper container through a hose, where the freezing takes place by heat transfer through the bottom. Water returns to the lower one (spilling over a weir). The advantage of this scheme is that no insulation needs adjusting, & the lower reservoir provides a place for the minerals to concentrate. I don't know whether a small heater will be needed in the lower container. Regarding cutting the ice, why not use a coarse-tooth hacksaw or an old hand wood saw?

You mentioned wanting to counterbore holes but had no counterbore & the hazard trying to do this in a drill-press with a milling cutter. You can do this safely! First, drill the thru-hole. Then drill the counterbore with a DRILL bit ALMOST deep enough. This can be safely done on the drill press, as the drill bit is self-centering on the existing hole (if you are doing this in brass, be sure to stone your drill bit so it doesn't grab). Then, mount your milling cutter in the drill-press chuck. With the drill press OFF, lower the milling cutter well down into the counterbore hole you drilled. Hold the handle down to KEEP the cutter in the hole (other option is to raise the table so the cutter can't leave the hole, even with the handle up). THEN turn on the drill press. T…

You mentioned wanting to counterbore holes but had no counterbore & the hazard trying to do this in a drill-press with a milling cutter. You can do this safely! First, drill the thru-hole. Then drill the counterbore with a DRILL bit ALMOST deep enough. This can be safely done on the drill press, as the drill bit is self-centering on the existing hole (if you are doing this in brass, be sure to stone your drill bit so it doesn't grab). Then, mount your milling cutter in the drill-press chuck. With the drill press OFF, lower the milling cutter well down into the counterbore hole you drilled. Hold the handle down to KEEP the cutter in the hole (other option is to raise the table so the cutter can't leave the hole, even with the handle up). THEN turn on the drill press. The milling cutter is safely captured in the hole & you can safely drill to finish depth & get the flat-bottom counterbore. Turn off the drill-press BEFORE you lift the cutter out of the hole. Note that counterbores are OVERSIZE from stated diameter. You may need to drill with slightly oversize drill if the screw-head binds in the counterbore.

Safety first! As soon as you power this beast, the hazard level goes way up! First, the chain, sprockets, & all other moving parts except the bead rolls should be covered with a guard. Second, the machine should be controlled with a "dead-man" switch that requires constant pressure to keep the machine operating.

Make the robot's center of gravity as high as possible. Mount the accelerometer & the gyro at the robot's center-of-gravity. This will reduce noise on especially the accelerometer.

You didn't indicate how heavy the lumber is, but I suspect heavy! Those small, hard wheels have a tendency to settle into any depression & get stuck there. It is better to use larger diameter wheels with pneumatic tires. A pair of 20" garden-cart wheels should work well. Another option would be a pair of fat 10" under inflated tires/wheels.

This is a mover I made for moving large trees in pots over relatively level ground. Some of these weigh 500 lb, but most are under 200. Cart consists of a flat plate "hovering" inches off the ground, held by pair of 26" bike wheels. There is a long handle for leverage. (Photo shows 2 carts flipped up). While this is different problem from what you are doing, I found that any smaller wheels for this cart became "glued" to the tiniest depression, even in a "finished" lawn. The ground conditions I have are considerably smoother then yours, but the weight higher. My ground clearance is set low on purpose to minimize chance of the pot tipping over. (You don't have that problem!)

Christmas bubble lights contain a relatively volatile fluid in a totally-sealed tube. Heat at the bottom actually boils the liquid, the vapor rising & condensing back to liquid. I have made larger bubble tubes. The engineering details are a nightmare to get right. The liquid must be put in the tube & a vacuum drawn to get air out & boil the liquid. Then the neck of the tube is sealed off with torch. There must be a slug at the bottom to "catalyze" the boiling. (The liquid in the tube can superheat, then violently boil which breaks the tube!) After running awhile, the tube (if it is long & thin) will have bubbles only near the bottom (they condense as they rise). You need a long vapor-filled region at the top to reduce this effect. You need to find a non…

Christmas bubble lights contain a relatively volatile fluid in a totally-sealed tube. Heat at the bottom actually boils the liquid, the vapor rising & condensing back to liquid. I have made larger bubble tubes. The engineering details are a nightmare to get right. The liquid must be put in the tube & a vacuum drawn to get air out & boil the liquid. Then the neck of the tube is sealed off with torch. There must be a slug at the bottom to "catalyze" the boiling. (The liquid in the tube can superheat, then violently boil which breaks the tube!) After running awhile, the tube (if it is long & thin) will have bubbles only near the bottom (they condense as they rise). You need a long vapor-filled region at the top to reduce this effect. You need to find a non-toxic, non-flammable, readily volatile liquid to use. The light at the bottom needs to provide a lot of heat.

Another idea for this bubble lamp: replace the "ordinary" lamp with an RGB LED lamp that you can control with a computer. Write program to strobe the LEDs (about 2 milliseconds ON followed by 123 milliseconds OFF). You will see rapid succession of "freeze frames" of the bubbles. If you play with the timing of the pulses (for example, firing the red first, then green, you will see bubbles with red bottom, yellow center (if pulses overlap), then green on top. The software can change the colors periodically. The "magic" is to have the rep-rate about 8 Hz. This is slow enough to give the bubble time to move between the flashes, but fast enough for the eye to "connect" the images.Background lighting needs to be dark. This is difficult to photograp…

Another idea for this bubble lamp: replace the "ordinary" lamp with an RGB LED lamp that you can control with a computer. Write program to strobe the LEDs (about 2 milliseconds ON followed by 123 milliseconds OFF). You will see rapid succession of "freeze frames" of the bubbles. If you play with the timing of the pulses (for example, firing the red first, then green, you will see bubbles with red bottom, yellow center (if pulses overlap), then green on top. The software can change the colors periodically. The "magic" is to have the rep-rate about 8 Hz. This is slow enough to give the bubble time to move between the flashes, but fast enough for the eye to "connect" the images.Background lighting needs to be dark. This is difficult to photograph. Attached is photo of 3" diameter tube 2' tall outside at night on Halloween.

If you use water, use distilled, add some alcohol (about 20%). This will keep critters from growing in the lamp.If you don't use distilled water, the iron in the water will precipitate with time & air bubbles, staining the glass dark & blocking the light.

The problem with wind is that it is not as consistent as water. You would need a "little man" constantly tweaking that rheostat. That's why commercial turbines have "maximum power point tracking" circuits. Modern electronics makes this simpler.

Most of the "groundhog" & farmer's water-pumping windmills are focused on high torque & low speed. They have many blades, set at steep angle to the wind. They are not good for extracting power from the wind.With 18" blade, set at low angle to wind, you should be able to get 300 to 400W from the wind at 30MPH with proper alternator & tracking circuit. The blade's angle is correct when the blade's tip speed is 7 times the wind speed.You could try some of the larger "outrunner" airplane motors (smaller turbine), or hoverboard motors (larger, slower-turning turbine).

They are energizing the correct part (the rotor).

Your statement "Using a car alternator for a wind turbine might not be a good idea" except replace "might" with "IS". Car alternators are designed to be CHEAP. In their intended use, they are lucky to be 50% efficient (who cares? There is a 100HP engine behind it. What do you think that fan is for?) The steel laminations in the stator are way too thick causing high eddy-current losses. What some have done is to replace the rotor electromagnet with rare-earth magnets. This gets rid of the rotor power loss, but because of the thick laminations in the stator, the losses there are high. You also have to position the magnets carefully to minimize "cogging torque" which can prevent the turbine from starting up.There is something else that is not r…

Your statement "Using a car alternator for a wind turbine might not be a good idea" except replace "might" with "IS". Car alternators are designed to be CHEAP. In their intended use, they are lucky to be 50% efficient (who cares? There is a 100HP engine behind it. What do you think that fan is for?) The steel laminations in the stator are way too thick causing high eddy-current losses. What some have done is to replace the rotor electromagnet with rare-earth magnets. This gets rid of the rotor power loss, but because of the thick laminations in the stator, the losses there are high. You also have to position the magnets carefully to minimize "cogging torque" which can prevent the turbine from starting up.There is something else that is not right (that may be the more serious problem why you are not getting more power). You reduced the length of the blades, & the speed went up! That shouldn't happen. There is something wrong with the pitch/shape of the blades. THEY MAY BE BENDING. A commercial turbine, blades a little shorter than yours, (made with alternator similar to car alternator except better laminations & permanent magnet rotor) gets 500W maximum in 30MPH steady wind.One thing you should try is changing the pitch of the blades (rotate them about the rod you have fastening them to the center). Ideally, the pitch should be set so the blade's speed through the air is 7 times the wind speed. The blades should twist along their length (you can heat the pipe & twist it, or cut it into sections & have each set at a little different angle. Other option is to have NO BLADE near the hub.)Another thing you need is a "maximum power point tracking" system. The load on the alternator needs constant adjusting as wind speed changes. (For example, if wind speed doubles, turbine speed needs to approximately double (which will double output voltage). If the load is resistive, the current would double, quadrupling the power to the resistor. But the power available from the wind goes up by 8 under those conditions, so the current needs to quadruple. If you connect (rectified) alternator directly to a battery, the turbine will be held back at high wind & run too fast at low wind. You need DC/DC converter between the alternator & the DC bus you are feeding. (This can be simple in the hardware, because the inductance of the alternator windings can form the "hard part" of the converter, requiring only a FET and a diode). For high-wind protection, the same FET that does the DC/DC conversion can short-circuit the output of the alternator. This causes the turbine to "stall" & turn slowly. You could also build a voltage regulator to adjust the rotor current instead of the DC/DC converter (if you use the original rotor electromagnet). You then need circuit to sense when the wind is not enough to provide rotor current in this case & prevent discharge of the battery.You might have better luck with a hoverboard wheel motor for the alternator.

Be very careful with that compressor (or vacuum cleaner, or any other air-moving device). If you manage to build up even the tiniest amount of pressure or vacuum in the window, you will produce TONS of force on it. Better would be to point a small blower at one of the openings to slowly circulate the air. I would also use distilled water in the vinegar mix, & use a second pass with distilled water only to get all the acid out, then the alcohol. By the way, those "sealed" panes is PURE MOONSHINE. If the panes were truly sealed, when atmospheric pressure changed, huge forces would break the window! Just one inch mercury change on the barometer (common change when storm approaches) is 1/2 pound per square inch. On your window (I guess it is 18" x 4')…

Be very careful with that compressor (or vacuum cleaner, or any other air-moving device). If you manage to build up even the tiniest amount of pressure or vacuum in the window, you will produce TONS of force on it. Better would be to point a small blower at one of the openings to slowly circulate the air. I would also use distilled water in the vinegar mix, & use a second pass with distilled water only to get all the acid out, then the alcohol. By the way, those "sealed" panes is PURE MOONSHINE. If the panes were truly sealed, when atmospheric pressure changed, huge forces would break the window! Just one inch mercury change on the barometer (common change when storm approaches) is 1/2 pound per square inch. On your window (I guess it is 18" x 4') that would be 432 pounds of pressure against the glass! Any argon in there is probably gone before the window is hung! That's why I would NEVER buy such a window, & stick with something that can be disassembled!

All you need is a shade frame with hoops at the top & bottom. Get steel coat hangers, cut & straighten. Loop the wire over the top & bottom hoop. You can put a dab of glue to keep them from sliding sideways. Only tools needed are 2 pliers, one with a wire cutter.If you play the game right with magnet polarity & position them right at the corners, you can arrange them to attract the slides to one another, in addition to the frame.

You could make a steel wire frame for the lampshade, with square cells to match the spacing of the slides. To the back of each slide, in the corners, glue small magnets. Now you can change the slides to suit the occasion! Regarding all those family slides, invest in a scanner & scan them before you mount them onto the lamps. That way, you will have the images!

Is the motor submerged in the water? If so, it won't last! It would be better to put a magnet "stir-bar" (used in chemistry) inside the bulb & have another magnet on the end of the motor shaft to couple to the magnet inside the bulb.

What kind of coin is 100% nickel? In the USA, 5 cent piece is 75% copper & 25% nickel (& is not magnetic even at room temperature).Some people have made Curie engines from cigarette lighter flints. These have much lower Curie point, & can be "activated" by hair dryer.

When doing casting with plaster, you may have problems because the plaster expands on setting. It is also weak if you have thin sections. You can use instead "hydrocal-30" gypsum cement. This behaves like plaster in mixing & handling, but it is much stronger when set, & has less than half the expansion.

The better for term for this is "quizzing glass" (which is hand-held), not "monocle". If the lens is from a standard pair of full-vision reading glasses, it makes no difference whether it is upside down or at any other angle. But the concave side should always face the eye for best optical performance.

All of these enhancements are in software (except for series resistors to avoid destroying servos). I sent details to your email. Hope this helps!

Paint the inside of the dish with black paint. Set up, & set bright point light source shining down onto the water. Magnified light wave patterns will be reflected upward onto the ceiling. You can set up 45 degree mirror if you want to project the wave pattern onto the wall.

You should NEVER need mold release on silicone. NOTHING sticks to silicone except more silicone (that is of the "self healing" type). If you tried to apply mold release to silicone, most likely it would bead up & run off.Regarding not having vacuum setup, you already have most of what you need! The pressure pot (unless it is very large) can be used as your vacuum chamber (you need not bolt on the lid). Vacuum is so much more effective in removing bubbles. (Pressure only makes existing bubbles smaller. If you have a large bubble, the finished epoxy can actually explode when you release the pressure!) Even better is mixing while under vacuum, totally avoiding bubbles in the first place. If you degas your mix & you pour it into mold through tube inserted to the bott…

You should NEVER need mold release on silicone. NOTHING sticks to silicone except more silicone (that is of the "self healing" type). If you tried to apply mold release to silicone, most likely it would bead up & run off.Regarding not having vacuum setup, you already have most of what you need! The pressure pot (unless it is very large) can be used as your vacuum chamber (you need not bolt on the lid). Vacuum is so much more effective in removing bubbles. (Pressure only makes existing bubbles smaller. If you have a large bubble, the finished epoxy can actually explode when you release the pressure!) Even better is mixing while under vacuum, totally avoiding bubbles in the first place. If you degas your mix & you pour it into mold through tube inserted to the bottom, mold will fill from the bottom up & avoid making more bubbles.

Thanks for the update!14ms time constant is a lot longer than your period (1ms). Current ripple is only 10%. So if the controller "decides" to adjust the current, the dominating time is the 14ms, not 1ms. Your control loop's frequency response is limited to about 50 Hz. You could speed this up by using a higher supply voltage to the bridge (or substituting a coil with fewer turns of heavier wire which would increase coil current, but have same power), & using full bridge instead of single FET & diode. If you did this, you would also have the benefit of being able to configure the loop to adjust the globe's position to levitate at zero coil current (you can go both plus & minus about that point to deal with disturbances).

There are some things you can do to lower power requirements & simplify the circuits. You can eliminate 3 half-bridges in the rotation circuit, by connecting one end of each of the rotation coil sets together ("star" configuration) & the 3 remaining wires go to 3 half-bridges. Regarding the power required for the lifting magnet, you can reduce that by putting a thin magnet on the central pole face of the lifting magnet. This can be adjusted to provide 90% of the lifting force at your set levitation distance. (This can make it practical to increase your levitation distance. This makes the system more stable.)Comment about the 1ms pulse & the "distance the globe falls in 1 ms": The magnet has a large inductance. You (correctly) have a flyback diode in …

There are some things you can do to lower power requirements & simplify the circuits. You can eliminate 3 half-bridges in the rotation circuit, by connecting one end of each of the rotation coil sets together ("star" configuration) & the 3 remaining wires go to 3 half-bridges. Regarding the power required for the lifting magnet, you can reduce that by putting a thin magnet on the central pole face of the lifting magnet. This can be adjusted to provide 90% of the lifting force at your set levitation distance. (This can make it practical to increase your levitation distance. This makes the system more stable.)Comment about the 1ms pulse & the "distance the globe falls in 1 ms": The magnet has a large inductance. You (correctly) have a flyback diode in the driver circuit. The current in the lifting magnet DOES NOT change significantly in 1 ms. The magnet's inductance acts as a low-pass filter. The magnet's current has a large DC term, & a small ripple. One disadvantage of your driver circuit is that it can increase the lifting current quickly, but not lower it quickly. This MAY lead to an instability that is unsymmetrical. To comment further, I would need to know the lifting magnet;s inductance.

If you put 1 tablespoon salt in both the crust & the sauce, the pizza will contain 14 GRAMS of sodium, plus the additional amount in the cheese & other ingredients. If you cut this 8 ways (as in the photo), sodium content is at least 1750mg per slice, definitely not healthy for RMD of 2000mg maximum. Otherwise, the pizza is simple to make & tastes much better with no salt added (the cheese has enough). That is the advantage of making it yourself; you can regulate the content. You do not give quantities for cheese so I can't comment on the saturated fat content or total calories, which you also want to limit for health food.

Can't you use the CNC machine to make the probe? The only difference between a vertical milling machine and a lathe is that the lathe is a milling machine "on its side", and the milling machine has no tailstock. You can machine all the circularly-symmetrical parts in the CNC's collet, fastening the cutting tools (or abrasive) to the table. Make the parts from the collet end to the tip, and machine the next part. Any errors in prior steps will be corrected by the next machining step. The final step is machining the ball, which should correct for the all the other errors. Another dirty trick is to use the probe while the spindle is rotating slowly. No matter how crooked things are, the probe will just "look bigger than it is", but perfectly symmetrical, wh…

Can't you use the CNC machine to make the probe? The only difference between a vertical milling machine and a lathe is that the lathe is a milling machine "on its side", and the milling machine has no tailstock. You can machine all the circularly-symmetrical parts in the CNC's collet, fastening the cutting tools (or abrasive) to the table. Make the parts from the collet end to the tip, and machine the next part. Any errors in prior steps will be corrected by the next machining step. The final step is machining the ball, which should correct for the all the other errors. Another dirty trick is to use the probe while the spindle is rotating slowly. No matter how crooked things are, the probe will just "look bigger than it is", but perfectly symmetrical, when rotating.

Be careful when you substitute. Depending upon the circuit tube is in, there are other characteristics that must match (such as whether the triode is "high mu" or not) for proper circuit function. Fortunately, for "receiving" tube circuits, usually nothing will "blow up" if everything doesn't match the original, but circuit function may not be proper. I had a fit replacing a 6X9 tube in Philco Predicta TV. There was just nothing that even partially matched! I had no choice but to buy the exact replacement. We run into these games because TV manufacturers would swap pins on old tube type & declare a new tube type for their product (to force consumers to purchase their tubes).

What I would do is to make 3 sides of the container out of aluminum, and the front only from clear plastic, to totally eliminate the need for any operations on the plastic other than cutting to size & drilling holes.

You don't need that ratchet strap. Use a loop of bungee cord permanently fastened below, and just pull it up & over the separator module. The suction helps clamp the top on. Regarding fire-polishing acrylic plastic: DON'T! Looks great now, cracks & crazes later, unless you put the whole box in an annealing oven immediately after assembly. Acrylic plastic is hard to work with. I would use polycarbonate instead. Regarding the hose coming from the top & tipping: loop the hose down & secure the hose to the body low.

No web pages, just read the data sheet on the A/D converter. A/D's are clocked from a reference clock. After the prescribed number of cycles, the A/D is ready for readout. If you know that time is (usually published as that sets how frequently samples can be taken), you can ignore "ready" signal. By giving Command Convert right after the read, you leave the maximum time for conversion. Make sure there is no jitter in the Convert signals. If you code "straight line" and invoke it from an interrupt timer, and do the read/Convert 1st thing, you will minimize the jitter.

To label your seeds, get an old narrow-slat (light color, matte finish) vinyl window blind. Expand to full size, and lay on ground. Cut the strings that run vertically through the slats & pull the strings out. Now you can remove the slats by sliding out sideways. Cut the slats into convenient length to form labels. Use pencil to write on the label, and poke into the soil. You can even write on them when wet! Water & sun will not affect pencil marks. You can also use pairs of the slats as seedling scoops. Poke into soil on each side of the seedling, press together, and lift out the seedling with roots & soil intact, and insert into transplant pot.

To get ADC samples with less hassle: read the ADC, then immediately AFTER the read do a command convert. This way (working backwards), the ADC has almost the whole time between samples to convert, and you need not wait or handshake. As long as the time between samples is more than the time to convert, you need not check if the data are "ready" for the read. Another way to increase processing speed is to code the filter routines in assembly. There are many dirty tricks that can be played (such as "rotating the filter coefficients" rather than shifting the data (data shifting requires read-modify-writes which are slow), and using look-up tables for the multiplies. You can also cascade filters after a front-end LPF & reduce ("decimate") the data…

To get ADC samples with less hassle: read the ADC, then immediately AFTER the read do a command convert. This way (working backwards), the ADC has almost the whole time between samples to convert, and you need not wait or handshake. As long as the time between samples is more than the time to convert, you need not check if the data are "ready" for the read. Another way to increase processing speed is to code the filter routines in assembly. There are many dirty tricks that can be played (such as "rotating the filter coefficients" rather than shifting the data (data shifting requires read-modify-writes which are slow), and using look-up tables for the multiplies. You can also cascade filters after a front-end LPF & reduce ("decimate") the data rate in the following filters. An IIR filter can (with minimal additional math) split the energy into LPF & HPF.

If you send me email address, I can dredge up the setup schematics and videos of the trooper working and send them to you.

I built a similar machine, but used a long-throw subwoofer with electronic drive in place of the diaphragm. What I found that the pre and post-motion affected the efficiency of the machine. A short, abrupt pull back, then the forward thrust, followed by another pull-back after made much better rings. I use this as part of a Halloween storm-trooper with gun display. The trooper shoots as people as they pass by.

If you replace the long lever arm with a cam that the string wraps around, you can set up the mechanical advantage higher at the start to overcome inertia, then reduce it for the long cruise. Adding a stationary idler that the string wraps over at the end would eliminate the "bang!" and its attendant loss of the last bit of energy.

There is another player in this game. WD-40 is now available in hand-pump-spray cans (like common window cleaner). The can looks like an aerosol can, but it is not pressurized. It has a screw-on lid, but built-in ratchets forbid removing the cap to refill after it is used up. I chewed up the bottom edge of the screw-on lid, where the ratchet teeth are, with diagonal cutters, allowing removing the cap, and re-filling from bulk container. I prefer the pump-spray to pressurized spray because you can control the amount delivered to the drop, versus the "fire hose" delivery from pressurized containers. The only thing I lose is the tube.

What is the wax for?

Other people use baking soda solution for the etchant. It makes the water "less gross". The etch will go better if the work and the scrap electrode are set parallel to one another.

Steel is being used. The "eating power" comes from the power supply rather than the chemistry. This process is called "electrolysis", and is used to both etch and plate metal. You can't etch circuit boards this way because as soon as you etch through, the copper islands become isolated from the power supply, and etching stops.

If you are making the backing pattern by 3D printing, why not print a negative of it, and use this, on the front, (with some thin padding) to press the foil into the depressions?

There are many details about making the engraver machine, but none as to the details how the tape is handled and applied for the masking step, other than showing photo of finished board. Is the tape applied to the board first and then engraved, or is it applied to some sort of backing and engraved, peeled off the backing and applied to the board? If the former, how are all the little holes removed, and the burnt edges of "goo" dealt with? If the latter, what is the backing, and how is the fragile tape removed from the backing, and aligned and applied to the board?

I sent the charts for my hexapod to your email. You can adapt this scheme for your robot. Send me any questions.

If you get the mains reversed, so the "hot" side is on the negative side of the capacitor, you will be connected to the line, in series with a 12V drop. Be safe and isolate, and use a separate earth ground! NEVER RELY ON LINE POLARITY! With the ground, even if the isolation fails, the breaker will be tripped rather than getting shocked. A commercial product ("vacuum leak tester" made by Electro-Technic Products), did the same, common-ing the secondary to the primary side in their handheld line-operated unit https://www.electrotechnicproducts.com/even when they have grounded cord! They think a 1/4" gap in the probe will prevent shock. It won't! Once the high voltage establishes an arc across, the 120V can pass at will. I had the privilege to meet …

If you get the mains reversed, so the "hot" side is on the negative side of the capacitor, you will be connected to the line, in series with a 12V drop. Be safe and isolate, and use a separate earth ground! NEVER RELY ON LINE POLARITY! With the ground, even if the isolation fails, the breaker will be tripped rather than getting shocked. A commercial product ("vacuum leak tester" made by Electro-Technic Products), did the same, common-ing the secondary to the primary side in their handheld line-operated unit https://www.electrotechnicproducts.com/even when they have grounded cord! They think a 1/4" gap in the probe will prevent shock. It won't! Once the high voltage establishes an arc across, the 120V can pass at will. I had the privilege to meet (and berate) the engineer who designed it. His reply: "never thought of it". I modified mine (removed primary, fished out the secondary, connected that to ground).

SAFETY FIRST: KNOW THE HAZARDS! You should not touch ANY part of this coil when plugged in (or input capacitor charged), and that includes the high voltage output! No, the hazard is not the high voltage, IT IS THE LINE VOLTAGE. There is a DIRECT CONTINUITY current path from the line "hot", through the secondary, to the top. you can be electrocuted from the line even if no high voltage generation is occurring! The only way to make this top electrode safe is to have the Tesla transformer secondary ohmically isolated from the primary (it then becomes safe because the air-core transformer cannot pass 50 or 60 Hz line frequency power to any significant extent, and there is no continuity for the 120V to come through). The "bottom" end of the secondary would nee…

SAFETY FIRST: KNOW THE HAZARDS! You should not touch ANY part of this coil when plugged in (or input capacitor charged), and that includes the high voltage output! No, the hazard is not the high voltage, IT IS THE LINE VOLTAGE. There is a DIRECT CONTINUITY current path from the line "hot", through the secondary, to the top. you can be electrocuted from the line even if no high voltage generation is occurring! The only way to make this top electrode safe is to have the Tesla transformer secondary ohmically isolated from the primary (it then becomes safe because the air-core transformer cannot pass 50 or 60 Hz line frequency power to any significant extent, and there is no continuity for the 120V to come through). The "bottom" end of the secondary would need connecting to EARTH ground. This will defeat the oscillator feedback and render the coil inoperative (but the top will be safe!) Feedback could be provided from the secondary to the FET gate with a small current transformer (wound on a ferrite core). This transformer should be wound 1:1 with 2 separate windings (insulated from each other to withstand line voltage), one across the TVC diode, the other in series with the secondary's bottom end and EARTH ground. The phasing between the primary and secondary of the Tesla transformer is critical. if the coil does not oscillate, try reversing the primary coil's connections.

You should have "broken" the outside edges over (like you did on the inside), and contoured out the backside of the nose. Then you should cut the hinge mortises on front and back. Next, a curing finish (such as 2 part epoxy should go on, including in the lens grooves. Finish should be heavily applied (preferably by dipping, or even under vacuum to impregnate well). Then wipe off excess, dry, and polish before the lenses go in. If you don't have heavy finish, the frame will be impossible to clean. It needs to be able to withstand water immersion if you plan on using them regularly!

Protecting the IC should be easy. First, DO NOT ELIMINATE the "upper" resistor in series with the lamp. There are 2 reasons. First, that resistor will limit the current spike into the IC. 2nd and more important, the lamp's V-I characteristic has "negative resistance" & operating point of the lamp will be unstable.https://en.wikipedia.org/wiki/Pearson%E2%80%93Anson_effectThe the "positive" resistance in series must be greater than the negative resistance, to make it stable. This is common in most discharge lamps (fluorescent lamp, sodium vapor lamp, mercury vapor lamp, "CFL"). This is why such lamps need a "ballast" to operate. Yes, that "extra voltage" is there and contributes to circuit losses, but is a "nec…

Protecting the IC should be easy. First, DO NOT ELIMINATE the "upper" resistor in series with the lamp. There are 2 reasons. First, that resistor will limit the current spike into the IC. 2nd and more important, the lamp's V-I characteristic has "negative resistance" & operating point of the lamp will be unstable.https://en.wikipedia.org/wiki/Pearson%E2%80%93Anson_effectThe the "positive" resistance in series must be greater than the negative resistance, to make it stable. This is common in most discharge lamps (fluorescent lamp, sodium vapor lamp, mercury vapor lamp, "CFL"). This is why such lamps need a "ballast" to operate. Yes, that "extra voltage" is there and contributes to circuit losses, but is a "necessary evil". For expensive high-power things like gas laser tubes, the loss in the ballast resistor is 1/3rd that in the tube! That's why there are patented circuits to "present" high series resistance at high frequencies, but low resistance to DC.Look up this patent:https://www.freepatentsonline.com/result.html?sort=relevance&srch=top&query_txt=US4107580&submit=&patents_us=onYou could add-in this circuit in place of "top" resistor to see what efficiency you could get. You should be able to lower the ballast voltage drop 10x. It will also reduce the ripple. DO NOT REDUCE the inductor value! Increase the capacitors value in the doubler circuit to reduce voltage ripple. For IC protection, add 2 series 1K resistors between top of 330 ohm sense resistor and the IC sense input. At the center tap, put signal diode (1N4148) to the IC's + power rail (anode to resistors, cathode to the rail). The diode & resistors will "do nothing" during operation, but clamp the voltage to the rail during start. You should also have independent voltage control loop "diode'd" in, to prevent voltage runaway if the lamp fails or is removed.

It would be a good idea to paint the cardboard frame flat black on the front side (facing the mirror) so it doesn't show.

You can reduce the variability between lamps by regulating the current through the lamp instead of the voltage. Swap the neon lamp and the series resistor, so one end of the resistor is grounded. Now you have a voltage proportional to current. You can split this resistor into 2 and pick off the "tap" that regulates the current to what you want. You may be able to reduce the (total, 33K) value of the "ballast" resistor, increasing efficiency. Couple in the original voltage feedback with a diode so if the lamp fails or is absent, the circuit won't "run away" and damage itself.

Here are some improvement ideas. First, for each audio frequency band, have 2 light channels. One of the lights turns on when the audio energy gets higher, and the other turns off. This way, no matter what the audio is doing, there is always light. Second, use triacs (or SSR's that allow phase control) to drive the lights. This will permit smooth dimming, not just off-on. Third, universal problem with "color organs" is that the dynamic range of the audio is greater than the lights (in quiet passage, lights are always off, in loud passage, lights are always on). You solve this by having "automatic gain control" on the input audio, which adjusts the gain to keep the lights "active". The gain adjustment circuit should have a time constant…

Here are some improvement ideas. First, for each audio frequency band, have 2 light channels. One of the lights turns on when the audio energy gets higher, and the other turns off. This way, no matter what the audio is doing, there is always light. Second, use triacs (or SSR's that allow phase control) to drive the lights. This will permit smooth dimming, not just off-on. Third, universal problem with "color organs" is that the dynamic range of the audio is greater than the lights (in quiet passage, lights are always off, in loud passage, lights are always on). You solve this by having "automatic gain control" on the input audio, which adjusts the gain to keep the lights "active". The gain adjustment circuit should have a time constant of about 20 seconds. If the front-end analog circuitry and the A/D converter have enough dynamic range (12-14 bits), you can do the "gain control" in software. If you want to get real fancy, use the Arduino to generate digital multiplex code (DMX). This can be used to drive standard DMX stage lights, or stage "dimmer packs" (which can drive incandescent lights).

When cutting a radius at the end of a slot in the board, drill the hole FIRST, then bring the saw cuts to it. This will prevent cracks from starting at the root of the cuts, and makes aligning the cuts to the hole easier. You will not have to fight the drill bit as you will be drilling into solid material. Why didn't you do more of the contour outline work on the CNC machine (that you obviously have)?

It's easier to center a hole on a slot than vice versa. Why do I say this? When making the jig, set the table saw for a very shallow cut and cut the slot. Then drill the hole from the slot-side, centered on the slot. Then back to the table saw, and cut the slot to the final depth, without changing the saw's setting. Another strategy is to cut the slot to finished depth first, then drill hole from slotted side (diameter to match width of slot) all the way through (with 120 degree twist drill). Slot will guide drill centered to it. Then flip over and drill with finished hole diameter (with twist drill). Using Forstner or other flat-bottomed wood bit with center point for the dowel-hole? Drill hole from slot side with twist drill as before, but only enough to form "spot" i…

It's easier to center a hole on a slot than vice versa. Why do I say this? When making the jig, set the table saw for a very shallow cut and cut the slot. Then drill the hole from the slot-side, centered on the slot. Then back to the table saw, and cut the slot to the final depth, without changing the saw's setting. Another strategy is to cut the slot to finished depth first, then drill hole from slotted side (diameter to match width of slot) all the way through (with 120 degree twist drill). Slot will guide drill centered to it. Then flip over and drill with finished hole diameter (with twist drill). Using Forstner or other flat-bottomed wood bit with center point for the dowel-hole? Drill hole from slot side with twist drill as before, but only enough to form "spot" in bottom of slot. Then drill through with tiny twist drill, also from slotted side. It will center on the "spot". Flip over, and drill with Forstner bit. It will center on the tiny through-hole, and will be centered over the slot. There is no need for any measurements to achieve the alignment!

You will have trouble with bacteria and algae growing in the ornament. You can prevent this by substituting about 30% alcohol for some of the water. Regarding the motor corrosion problem, put the motor outside the ornament and have it turn a small magnet. Another magnet (ceramic magnet will not corrode) inside will spin (like magnetic stirrer).

My experience with the "leaky fibers" is that you have to put a megawatt of light in to get a milliwatt of light out the side (and most of the light blasting out the other end). How did you manage all that "spilled light"?

I just hand-drew the curves and scanned them. There are tons of stuff on the web, and a lot of it is only "halfway true". here is example, this image byhttps://www.google.com/search?q=led+voltage/current+characteristics&rlz=1CAIEIT_enUS807&sxsrf=ALeKk02tJTTXa4TKyArcVFlZRxvSkOFXfQ:1607712756872&tbm=isch&source=iu&ictx=1&fir=29NVqrc3S0hzBM%252CIJrOiXdXo3o7PM%252C_&vet=1&usg=AI4_-kSzKJMiLVGY6R3vBguYr1klIT0L6A&sa=X&ved=2ahUKEwjirZjezMbtAhWj2FkKHVXcAHoQ9QF6BAgGEAE#imgrc=ENv6Oea7roKj2MBy last minute engineers.com, shows different curve for white and blue. A white LED is a blue one with a phosphor in front, so the curve should be the same! Those curves were most likely drawn by artist, not measured curves. After the turn-on threshold is reac…

I just hand-drew the curves and scanned them. There are tons of stuff on the web, and a lot of it is only "halfway true". here is example, this image byhttps://www.google.com/search?q=led+voltage/current+characteristics&rlz=1CAIEIT_enUS807&sxsrf=ALeKk02tJTTXa4TKyArcVFlZRxvSkOFXfQ:1607712756872&tbm=isch&source=iu&ictx=1&fir=29NVqrc3S0hzBM%252CIJrOiXdXo3o7PM%252C_&vet=1&usg=AI4_-kSzKJMiLVGY6R3vBguYr1klIT0L6A&sa=X&ved=2ahUKEwjirZjezMbtAhWj2FkKHVXcAHoQ9QF6BAgGEAE#imgrc=ENv6Oea7roKj2MBy last minute engineers.com, shows different curve for white and blue. A white LED is a blue one with a phosphor in front, so the curve should be the same! Those curves were most likely drawn by artist, not measured curves. After the turn-on threshold is reached, the curve becomes straight and steep (it is exponential flavored by the wire-bond resistance). Usually, green has slightly more drop than blue. As for temperature sensitivity, here is list of colors from most sensitive to the least: yellow, orange, red, green, blue, UV. My point is this: there are many fine details about LEDs (as with other components). Practical engineers design circuits to avoid sensitivity to those details that the manufacturers can't even control (and don't try, because smart engineers design to not be sensitive to them!) Yes, you could spend a year studying your batch of LEDs V/I/temperature curves, and maybe "get away" with other circuits, but why do so? Transistors' current gain can wary over decades! What do engineers do? They design circuits to work "if the gain is greater than x", and use a transistor that guarantees that over temperature on the data sheet. LED V-I curves are very steep in the forward conduction zone, and can vary all over the place (depending on wirebond details, and are not detailed in any data sheet). (What data sheet? For many LEDs sold from China!) Giving each LED circuit its own "current regulating" resistor completely eliminates the possibility of that problem! Some of the cheap Chinese lights wirebond LED chips in parallel in their "COB lights" 2 months later, there are a bunch of "dark emitters" in the array. (Never buy a light that claims "COB" which means "chip on board". Translated, means "too cheap to give each chip its own real package".)

I have looked at your code, and in particular, the difficulties with reading the button and switching to a new display mode (and that you use an interrupt to do so because many of the display routines are "forever" loops). Since you do not know where you were interrupted, you resort to hard reset. Then you lose variable to select the next display mode. This brings up questions: What happens if the interrupt lands in the middle of the LED update routine? If you can read EEPROM location 0, can't you do the same with RAM? Deposit the selectedEffect into unused RAM using the ASM statement, then read it back after reset. Why not also provide for the looping of selectedEffect back to zero in the changeEffect() routine by adding:if(selectedEffect>17{selectedEf…

I have looked at your code, and in particular, the difficulties with reading the button and switching to a new display mode (and that you use an interrupt to do so because many of the display routines are "forever" loops). Since you do not know where you were interrupted, you resort to hard reset. Then you lose variable to select the next display mode. This brings up questions: What happens if the interrupt lands in the middle of the LED update routine? If you can read EEPROM location 0, can't you do the same with RAM? Deposit the selectedEffect into unused RAM using the ASM statement, then read it back after reset. Why not also provide for the looping of selectedEffect back to zero in the changeEffect() routine by adding:if(selectedEffect>17{selectedEffect=0}Better would be to eliminate the need for hard reset. Write a routine to read the button. Put that routine inside all looping display modes (could be called every time a new "image" is sent to the LEDs). Button Test routine provides for abandonment of the display routine, back to the CASE statement where new display routine is branched to. Another method (better) is to banish all while(1)'s from the code (or have only 1, which encapsulates EVERYTHING include read button). The code structure is straight-line, initiated by periodic interrupt from a timer, and ending with Return From Interrupt. This code is effectively a loop. At the start of this code would be button test (polled) and adjust selectedEffect, then dispatched to the appropriate display pass. I did calculation, and it requires 3 milliseconds to write your 100 pixel display (24*100/800000). You could set interrupt rate to 100/second for smooth display motion. One display iteration could have 7ms, and 3ms would be for the NeoPixel update. No routine would ever be "caught in the middle" of its execution; no "stack overflows" could result. A lot of your display routines could be simplified by employing lookup tables, instead of (in essence) embedding the table in the code. If you are interested in more detail, I built similar display, but with only 6 LEDs but bubbling water, and critical timing. Code is in assembly, and is MUCH smaller. Instructables "attach file" does not work for my computer. Send email for code listing and photos.

You need to account for the "lead" of the saw (where the blade does not cut parallel to the saw table). Make test cuts in scrap. If the cut tends to spiral inward or outward, adjust the pivot point forward or backward to eliminate this tendency. Use a blade with a wide tooth-set. Do not use narrow-kerf blades! Someday, bandsaw manufacturers will WAKE UP and make a bandsaw with a table that can rotate a few degrees to eliminate the hassle this causes!

You state that you are "uneasy" about the table saw setup. FOR GOOD REASON!When wood binds against the blade, it is pushed back. With usual table saw arrangement (wood getting pushed directly in, fence parallel to the blade), backward motion tends to relieve the bind, and there is no tendency for the work to rotate because the binding force is in line with the pushing force. But in your setup, the wood piece is pivoted about the center point of the circle, which is laterally offset from the blade, and connected to a massive sled. If the work binds the blade suddenly, the inertia of the sled will resist moving back. The offset force will cause the work to rotate about that pivot, toward the blade, causing the bind to get tighter, contacting the blade in the back. The only thi…

You state that you are "uneasy" about the table saw setup. FOR GOOD REASON!When wood binds against the blade, it is pushed back. With usual table saw arrangement (wood getting pushed directly in, fence parallel to the blade), backward motion tends to relieve the bind, and there is no tendency for the work to rotate because the binding force is in line with the pushing force. But in your setup, the wood piece is pivoted about the center point of the circle, which is laterally offset from the blade, and connected to a massive sled. If the work binds the blade suddenly, the inertia of the sled will resist moving back. The offset force will cause the work to rotate about that pivot, toward the blade, causing the bind to get tighter, contacting the blade in the back. The only thing preventing this is the friction between the work and sled and your holding blocks, which is no match for the saw. The work will lift, and it and the sled will be shot back! Once the work is roughed out, the danger of binding is little as the blade is not embedded in the work. For all of the cuts, setting the blade as high as possible reduces the amount of pushback, and keeping the cuts short so the back side of the (up-moving) blade is not contacted (geometry more like the bandsaw setup). Having a guard in place "protects" the high-set blade and prevents the work from lifting. The high setting also makes the sled position easier to determine, and the final spin-cut truer. I hope you have Saw Stop on that saw!

Here is the thing that I want to emphasize (I am not sure if you understand the critical point regarding stability). An LED voltage/current curve is very "flat". Once you overcome the forward voltage, increasing only millivolts more causes the current to climb ENORMOUSLY. (That's why the resistor is needed in the first place). This curve is temperature sensitive. Is 2 LEDs are parallel connected, in series with common resistor, the currents will not match exactly. The one getting more will heat up more, which leads to that one's V/I curve dropping, leading to it getting even more current. The other LED cools, raising its curve. It can go as far as to take almost all the current from the other. Depending upon the resistor you select and the current rating of the LED (and…

Here is the thing that I want to emphasize (I am not sure if you understand the critical point regarding stability). An LED voltage/current curve is very "flat". Once you overcome the forward voltage, increasing only millivolts more causes the current to climb ENORMOUSLY. (That's why the resistor is needed in the first place). This curve is temperature sensitive. Is 2 LEDs are parallel connected, in series with common resistor, the currents will not match exactly. The one getting more will heat up more, which leads to that one's V/I curve dropping, leading to it getting even more current. The other LED cools, raising its curve. It can go as far as to take almost all the current from the other. Depending upon the resistor you select and the current rating of the LED (and whether the LEDs are on a common heatsink), the LED can be destroyed. At best, the brightness is not uniform between the 2. Such a circuit is very unstable! The putting in series is a totally different issue (although in this case also solves the stability problem). The system efficiency improves dramatically. If you put 3 in series, the same current is "used 3 times", and less power is lost in the resistor. Do not carry this too far. You need enough voltage "dropped" across the resistor to stabilize the circuit. When the "load lines" for both the LED(s) and the source cross at a right angle, the circuit is stable. When they are nearly parallel, watch out! (I have V/I diagrams, but @#$ instructables photo upload won't work!)

Beware edge banding: in time, it falls off! It tends to get caught by objects sliding over the MDF, and "picked" off. It is especially bad on sharp corners. Its bonding agent is basically "hot melt" glue, and that breaks down to powder with age. If you are that picky about finish, don't use MDF! Use quality plywood, and epoxy bond half inch thick boards (not veneer!) to the edges. TechTheaterLifer has it right for stage props that are not viewed close-up. Make sure the shellac is fresh.

When you wire LEDs, you should not connect them directly in parallel. One may heat and "hog" all the current from the others, leading to its destruction. Each one should have its own series resistor. Since you have 9V available and red LEDs drop only about 2 volts, you could save current (and resistors) by putting up to 3 in series with only 1 series resistor.

Instead of welding, bolt and braze. Brazing can be done with good propane torch, and the metal is not subject to the high heat shock so it doesn't crack later.

In my opinion, there is a hidden danger using the table saw method as you show. Of course, everyone knows that a table saw pushes back on the work, and that the work can jam against the blade, especially if the fence and blade are not set parallel, or the wood has locked-in stresses. When the work is a many-sided polygon, pinned at its center, any backward rotation will cause a jam against the blade, VIOLENTLY kicking back the work and sled, and possibly leading to fingers contacting the blade! I wouldn't even THINK about doing this without a guard in place, setting the blade as high as possible (to minimize the pushback tendency), and setting the guard as low as possible to just clear the work. Setting the blade high also gives truer cut when you get to the final "spin&q…

In my opinion, there is a hidden danger using the table saw method as you show. Of course, everyone knows that a table saw pushes back on the work, and that the work can jam against the blade, especially if the fence and blade are not set parallel, or the wood has locked-in stresses. When the work is a many-sided polygon, pinned at its center, any backward rotation will cause a jam against the blade, VIOLENTLY kicking back the work and sled, and possibly leading to fingers contacting the blade! I wouldn't even THINK about doing this without a guard in place, setting the blade as high as possible (to minimize the pushback tendency), and setting the guard as low as possible to just clear the work. Setting the blade high also gives truer cut when you get to the final "spin" pass. Your comment about bandsaw being limited in radius is true only to the extent you don't have room off to the side of the table. The clearance of the saw only limits the amount you can cut off IN ONE PASS. You can precut the corners off until you are less then that. You must set the pin carefully, accounting for the "lead" or cutting direction, or the blade will be pulled to the side, and produce a spiral and then jam. You can adjust for this by moving the sled forward/backward and making test cuts. Use a blade with a high tooth set. Avoid "thin kerf" blades. Set guides close and keep blade tension high. Another method to make circle is to first make template from scrap (using your's or other methods), fastening this to the work, and using router with follow-bearing against the template.

You should have made the barrel nuts from brass stock. This eliminates the need to "brass finish" them, and they are easier to tap then steel.

I wanted to make sure you got the materials I sent, and whether they were of use?

To make the anti-slip rings, cut rings from a pipe coupling.

Rather than make the haunted action directly in the program, drive it from a table that the program "reads". This way, when you want to change the haunted sequence, you need only change the table, not the code. You can arrange the table to one or more bytes to write to the ports, and a "delay byte" (how long to wait before executing the next table entry). For this book, a single byte will suffice for the output, and another for the delays (if you choose 1/4 second as your "time unit", you can have delay up to 64 seconds between actions). The code reduces to setup, and read the table, copy the byte out, and set the delay. A zero byte in the delay can signify "end-of-table" (loop back to beginning). You can put a test sequence at the beginning (…

Rather than make the haunted action directly in the program, drive it from a table that the program "reads". This way, when you want to change the haunted sequence, you need only change the table, not the code. You can arrange the table to one or more bytes to write to the ports, and a "delay byte" (how long to wait before executing the next table entry). For this book, a single byte will suffice for the output, and another for the delays (if you choose 1/4 second as your "time unit", you can have delay up to 64 seconds between actions). The code reduces to setup, and read the table, copy the byte out, and set the delay. A zero byte in the delay can signify "end-of-table" (loop back to beginning). You can put a test sequence at the beginning (to test all the actions in sequence) which is executed only once on power-up. The loopback goes beyond the test sequence.

Assuming your CNC has ballscrews, you need to either cut on the other side of the blade, or reverse the direction of rotation of the saw motor (& flip over the blade) so you are making "climb" cuts rather then "conventional" cuts. This will push the work down, rather then trying to lift it up (like a radial or miter saw). That was the major cause of the calamity you had happen. When using blade stack, you need to go slow. The motor overloaded and stalled. Cause #2 for calamity.

Simple solution to the ricochet. Fill the bucket quarter-way with wood chips. A barrel full of wood chips can capture all grenade fragments (detonated in the center of a 55 gallon barrel). Discard the chips after use as they will be full of nails.

Did. let me know if it is useful, or if you have questions.

When you insert code to debug (such as your status monitor) LEAVE THE CODE IN! It should not interfere, and more importantly, the code you run is the code you debugged!

You don't need a germanium transistor. There is plenty of voltage with many-turn coil and even weak magnet. It is done here commercially with silicon transistor (I think 2N3904. If not, that will work) https://www.teachersource.com/product/top-secret

The mechanism you built was patented by Anderson (3,783,550 Novelty Electric Motor) for their "Top Secret" toy, in 1974. Their coil was wound on a nail, about 3/4 inch long by 3/4 inch diameter. it keeps the top spinning for weeks on a 9 volt "transistor" battery. This is also used in novelty toys with spinning and swinging mobiles. Toy shown disassembled here, and you can buy for $12 https://www.teachersource.com/product/top-secret For efficiency, I would extract the energy from the pendulum with "magnetic escapement gears" (for at least the first few stages). You want to minimize the coupling to the pendulum to not disrupt its speed.

You don't need a germanium transistor. There is plenty of voltage with many-turn coil and even weak magnet. It is done here commercially with silicon transistor and 9V battery (I think 2N3904. If not, that, or other common small NPN transistor, will work.) To make this circuit run on single cell, you need to reduce the number of turns on the coil in series with the collector. For changing from 9V to 1.5V, you need to reduce the turns by 6, and make the wire's CROSS-SECTION 6 times larger (wire DIAMETER sqrt(6) larger). Leave the number of turns/wire gauge in the base circuit the same. The coils should fit in the same space. Current consumption will be 6 times higher (but power will be the same).https://www.teachersource.com/product/top-secret

I don't understand your statement. The magnets I am talking about attract and grab the numbers to the carriage, not produce movement. The movement is independently provided by stepper motor driving a screw (to engage/disengage the number from the carriage by moving the magnet toward/away) and the other 2 steppers on the carriage (for X/Y movement).

Your choice to use permanent magnets versus electromagnets on the carriage is a wise one! To use electromagnets, you would need huge power to get enough strength, and have problems with heating. With the permanent magnets, all you need is a motor to move them to/from the surface, which requires little power, and only during the "turning on & off" of the magnetic force. People underestimate the strength of rare-earth magnets. Just 1/4 inch thickness is equivalent to electromagnet with 5000 ampere-turns of current!

No matter what you do, tool heads loosen because the wood swells and shrinks with the seasons. When wood swells, wood cells on the outside of the handle get crushed by the tool head, which doesn't swell. Then wood shrinks, becoming loose. That's why bolts in wood get loose. This is referred to in the industry as "compression set". I have used Gorilla Glue to repair shovels, rakes, hammers, and yes, axes for 20 years. I have never had a head come off because the glue failed. I have had the handle break or rot, and if that happens, I cut it off flush with the head, drill a hole through, and shoot a torch through to burn out the remains of the handle. You really don't need strength. You need all voids filed, and shear strength which all glue's are good at. The glu…

No matter what you do, tool heads loosen because the wood swells and shrinks with the seasons. When wood swells, wood cells on the outside of the handle get crushed by the tool head, which doesn't swell. Then wood shrinks, becoming loose. That's why bolts in wood get loose. This is referred to in the industry as "compression set". I have used Gorilla Glue to repair shovels, rakes, hammers, and yes, axes for 20 years. I have never had a head come off because the glue failed. I have had the handle break or rot, and if that happens, I cut it off flush with the head, drill a hole through, and shoot a torch through to burn out the remains of the handle. You really don't need strength. You need all voids filed, and shear strength which all glue's are good at. The glue is trapped between 2 strong substances. An example of this is Locktite for securing threaded fasteners. Regarding forge welding, doesn't the steel have to have high carbon content for it to work? (The rasp is certainly, and the rollers in the chain do, but what about the side plates?)

This scheme is common for those trying to make circuit boards (there are Instructables and Youtube videos on the subject). But unless everything is perfectly set up, the transfer tends to be spotty. But for your purpose, this is fine! This works because the laser ink is plastic dust (that gets fused to the paper). This plastic is dissolved by acetone, so some transfers. Do tests to work out the technique. Have rigid surface under the canvas, with the print on top, taped along one (long) edge. Apply the acetone with a slightly dampened rag from behind the paper with light rubbing. You can roll the paper up and peek underneath for transfer skips, roll back down, and re-apply the rag. Have the lines on your print made good and dark, and not too thin. Keep the acetone away from the …

This scheme is common for those trying to make circuit boards (there are Instructables and Youtube videos on the subject). But unless everything is perfectly set up, the transfer tends to be spotty. But for your purpose, this is fine! This works because the laser ink is plastic dust (that gets fused to the paper). This plastic is dissolved by acetone, so some transfers. Do tests to work out the technique. Have rigid surface under the canvas, with the print on top, taped along one (long) edge. Apply the acetone with a slightly dampened rag from behind the paper with light rubbing. You can roll the paper up and peek underneath for transfer skips, roll back down, and re-apply the rag. Have the lines on your print made good and dark, and not too thin. Keep the acetone away from the tape or you will have a gooey mess. Other tricks is to tape the canvas (leading edge only) to stiff paper and feed through the laser printer (in the straight-through path) for small enough pictures and well-sized canvas. For huge canvas, project the picture on the canvas (mounted on the opposite wall) and paint the lines as you see them.

Why not draw contour lines over the image in the PC, along with very course grid lines. Print this to full-scale in sections, if your printer is not large enough. Use coarse grid lines to align the various prints. Cut and tape these together. If you have carbon paper, cover the canvas with this, then your taped print. Trace over the contour lines to transfer them to the canvas. If you make these prints in mirror and use laser printer, you can lay face-down on canvas and transfer the lines with acetone lightly applied from the back of the print with light rubbing. The transfer may be spotty, but it won't matter.

For holding tool heads on wooden handles, polyurethane "Gorilla glue" is the thing to use. The glue bonds tight, doesn't require mixing, is waterproof, and if the fit isn't perfect, the glue foams up to fill the gap. Let set 24 hours, then trim off the excess with rasp. Tool heads WILL NOT COME OFF!

You would be better to use a photodiode rather then an LDR. LDR's are notoriously nonlinear, and prone to age. You will need an amplifier between the photodiode and the Arduino. To get more dynamic range, have several gain stages (gains factors of 2) and feed each one into different Arduino A/D input. The "highest" one that is not saturated is your output.

What is the brand/type of the casting epoxy resin you are using? I see UN3082 but can not find.

Spacing is 8 to the inch. Rotor made by turning epoxy fiberglass tube until it just slips into copper pipe (diameter of fiberglass about 1 3/16 inch). outside of fiberglass and inside of pipe coated with epoxy resin, then slipped together and allowed to set. Then mount on lathe, and cut just through the copper 8 to the inch. Holes drilled in copper through to center for soldering wires. Outside brush guide made from 4 pieces of 1/16" thick, 2 unclad, 2 single-side 1 Oz copper clad. Interior square hole just large enough to fit rotor. Cuts through copper cladding isolated contacts (alternate on 2 sides giving room for terminals, 4 to the inch each side). Contacts made with bronze wire (too thick, but I could not get thinner material from McmasterCarr). Each wi…

Spacing is 8 to the inch. Rotor made by turning epoxy fiberglass tube until it just slips into copper pipe (diameter of fiberglass about 1 3/16 inch). outside of fiberglass and inside of pipe coated with epoxy resin, then slipped together and allowed to set. Then mount on lathe, and cut just through the copper 8 to the inch. Holes drilled in copper through to center for soldering wires. Outside brush guide made from 4 pieces of 1/16" thick, 2 unclad, 2 single-side 1 Oz copper clad. Interior square hole just large enough to fit rotor. Cuts through copper cladding isolated contacts (alternate on 2 sides giving room for terminals, 4 to the inch each side). Contacts made with bronze wire (too thick, but I could not get thinner material from McmasterCarr). Each wire contacts on 2 opposite sides of the rotor. There is a lot of friction, but that didn't matter as it was on large, heavy, missile test platform. I have photo, but Instructibles "Add images" won't work,I can send image by email. I also made tiny one (8 channel) for a "record player" sized centrifuge.

Is there some reason why you didn't use a single, large, piece of plywood, and cut all the rings out of that? Was it importent to have the sections of grain for each label?

Put clear plastic cover over the clock faces, and paint 45 degree black stripes on the cover, a little wider then the hands. That way, when you want the hands to "disappear", you can have the hands hide behind the black stripe. no more annoying 45 degree hands in unused faces! If you put 3 hands on the central 2 clocks in each number, you can properly form "3", "4", "6", "8", and "9".

I would use clear silicone rubber rather then construction adhesive. It adheres well to both glass and ceramic. If you get new tube and trim tip to small hole, you can dispense right out of the tube. Get the type of silicone ("silicone 1", or "acetoxy silicone). This type smells like vinegar (look for warning about acid fumes, not ammonia). I used small hose connected to vacuum cleaner for the pickup probe. I did a 10" diameter ball. If mirror gets damaged or broken, it can be pulled off and a new one glued in with silicone.

Use a "1 pass" sensor (collimated light in the back, straight light sensor in the front). You will need automatic tracking device (check out the one in the piano!)

Why not make the reader suitable for "real" piano rolls? Roll standard is 11 1/4" wide, 88 notes, 9 to the inch, centered. Speeds are in the range of 6 to 10 feet per minute. If you used "one pass" sensors (light source behind, simple detectors on top), you could achieve the tighter spacing. You will also need an "automatic tracking device" to keep the roll centered properly.

Tip for simplifying the "reverse kinematics" calculations: Rather than calculating every incremental move, calculate points "further apart", and use linear interpolation to calculate the intermediate points for smooth movement. I also recommend 2 other things in the software: one is a "startup routine" that starts the PWM to the servos one-at-a-time spaced out 1/2 second each to reduce the large current surge on startup when all 6 servos suddenly all attempt to jump to position. The other thing is a slew-rate routine that forbids a servo's movement more than "X" counts per iteration. This protects servos from abuse and limits current spikes which can overwhelm the power supply. I have charts for all 3 of these routines if you are interested.

I will need email to send material. Instructables file attach doesn't work.