Enchanted Rose With Diamonds!





Introduction: Enchanted Rose With Diamonds!

Epilog Challenge 9

This is an entry in the
Epilog Challenge 9

First Time Author Contest 2018

This is an entry in the
First Time Author Contest 2018

This is an enchanted rose I made for a birthday gift. The rose was a previous gift, and the glass dome was purchased when this idea was born.

The rose is a gold-trimmed rose product that can be found on Amazon or eBay. A second rose was used for the petals at the base.

The glass dome is from IKEA. The base is routed pine with rosewood stain and a polyurethane finish.

I built a simple "running hole" circuit with just a few resistors, capacitors, transistors, and LEDs into the base. I then ran four 0.01" fiber optic cables for each of the three LEDs. There is also a battery compartment with 4 AAs, and a switch that faces down so that it can be turned on and off by rotating the base.

At the end of each of the fiber optics is a tiny brilliant-cut diamond.

This instructable is not intended to be a simple "do these specific steps and then you are done" type of guide. This project has a lot of artistic choice and could have some very significant differences go into it, but still end up producing the same basic end result of an enchanted rose you can be proud of.

I spent well over 100 hours on it in about two months. Most of that time was testing and learning, and I hope this instructable can share what I've learned and help others with similar or even widely different projects.

Step 1: Materials

Here is an overview list of materials I used. A lot more went into this during my testing phases, but this is roughly what I would use if I wanted to build essentially a duplicate.

Primary Starting Materials

  • Glass bell jar
  • Artificial or dried/preserved rose x 2

Base / Wood / Finishes

  • 1”x12”x~24” Pine
  • Wood stain
  • Spray Polyurethane (or regular brush-on)
  • Felt
  • Protection
    • Eyes - Safety goggles are good for anytime you are using a power tool, especially for cutting.
    • Hands - Work gloves for cutting and latex / vinyl / neoprene gloves for finishing.
    • Lungs - A good respirator can be nice if there is fine sawdust in the air, but very important for things like spraying polyurethane.
    • Ears - Some of these tools are quite loud, so ear plugs are useful.


  • Prototype PCB
  • LEDs (I used three 3mm 20mA 3.3V white LEDs)
  • Resistors
  • Capacitors
  • Transistors
  • Switch
  • Battery holder and batteries
  • Wire
  • Heat shrink tubing


  • 0.01” plastic fiber optics
  • 0.75mm brilliant-cut diamonds
  • Small piece of plastic rigid sheet (or another means to affix bundles of fiber optics in line with the LEDs)


  • Two-part epoxy
  • Fabri-Tac
  • Wood glue
  • Hot-melt glue
  • Spray adhesive
  • Tape - cellophane and duct
  • IPS weld-on #16 (Didn't end up in the final result, but it was useful for prototyping with plastic parts.)

Step 2: Tools

Below is a list of tools I used. Most are not critical for completing a project like this; it all depends on what you have available and what you are attempting to accomplish. I will give general overviews of each and how I utilized them.


  • Dremel with assorted attachments and bits
  • Routers – table and plunge/fixed
  • Circle cutting jig for router
  • Clamps
  • Drill and drill bits
  • Wood carving tools / chisels
  • Rags (I used an old cotton t-shirt cut into pieces)
  • Sandpaper


  • Soldering iron
  • Solder sucker
  • Helping Hands with magnifying glass
  • Wire strippers
  • Digital multimeter
  • 300-in-1 Electronic Project Lab / Breadboard

General Purpose

  • Illuminated head magnifier
  • Chain-, round-, and needle-nose pliers
  • Diagonal cutters
  • Razor blades
  • Plastic scoring tool
  • Screwdrivers
  • Scissors
  • Files

Step 3: Deciding on and Gathering Materials

For this project, the starting materials (glass bell jar and two roses) were set, but I wanted to come up with a very unique end result that was nothing short of amazing!

I thought through and tested out many ideas, but my final design was mostly influenced by the desire for a clean and simple but elegant appearance, which meant as little visible inner workings (such as wiring) as possible.

I had some wood left over from other projects, and I salvaged some parts from old items around my house, such as a digital alarm clock, shelf stereo system, and portable CD player. I also had a decent selection of electronic components on hand. Most of the electronics I used can be salvaged from items like I mentioned.

I started designing this with a microcontroller (such as an Arduino UNO) in mind, but as I worked more towards simplifying and controlling power consumption (I wanted it to run on batteries, for as long as possible), I switched to simple ICs. I was working with a 555 timer and a 4029 counter, but I was not happy with the effect; it was too robotic-looking. Also, as I attempted to investigate other ICs, I was not happy about the lack of availability. You can get some online, but with lack of demand, and shipping costs, it just didn’t seem worth it. I used to get stuff like this from an electronic surplus store, or even Radio Shack, but not anymore.

I knew this did not need a complicated solution, and eventually found what I was looking for in a three-transistor astable multivibrator circuit. It uses no ICs and can be built with just three LEDs, three transistors, six resistors, and four or five capacitors. (One capacitor is essentially in parallel with the batteries, which may help handle load fluctuations, but isn’t required.)

When I first started planning this, I joked “I won't be adding diamonds or anything crazy expensive,” so of course, the first thing I did was order some diamonds. Next I needed a way to illuminate the diamonds. I tested out three different types of optical fibers. I tested some 5mm LEDs that I had, but felt they were too large. I only wanted 12-16 fibers around the rose, which meant only illuminating four fibers per LED. 5mm LEDs would be emitting most of their light around rather than into the fibers, so I purchased some 3mm LEDs.

I'm not trying to provide "the" way to do this project, but rather the way that I did it, and explain some of what went into the choices along the way. For someone else, it could make much more sense to have this controlled by an Arduino, or be hooked up to a PC via USB. Depending on what you have to start with, and what you want it to ultimately look like, there are plenty of ways to do something like this.

Most of the items I needed that I didn't have around, I purchased from Amazon or Wal-Mart.

Step 4: Electronics

There are lots of ways you could power and control LEDs for a project like this. My main objective was having it be battery-powered and last as long as possible, while also hiding all of the electronics, including the switch to turn it on.

I first thought of using a 9V battery, but after a small amount of research, I found that AAs hold significantly more power than 9V batteries by volume. I wasn't interested in trying to fit larger batteries, and space wasn't so tight as to call for AAA, so AA was an easy pick.

I had a two-AA holder in the portable CD player that I was salvaging, but my initial testing used ICs such as a simple 555 timer, so more voltage was necessary. I cut the battery compartment out of the CD player case, cut it in half, then used IPS weld-on #16 to glue it together (with an additional battery's length) with some strips of plastic that I also cut from the case. Now I had a four-AA case, and ~6 volts should last a decent amount of time running the circuits I was working with.

Once I got further along with the project, I decided that the battery case took up more space than desired. I ended up just removing the metal battery terminals and contact plate from the plastic and hot glued them directly to the wood at each end of the battery space.

In the first video, you can see: several petals I removed from the "spare" rose I used; the preliminary, four-AA battery holder; a bundle of fiber optics with some different end structures I was evaluating; some LEDs (one of which has the bundle of fibers taped to it) blinking in a binary counter pattern; and an early test circuit doing the counting that blinks the LEDs.

I was making progress, but I was not satisfied. The lighting effect I was getting was not pleasing, and the whole thing would likely stop working once the batteries dropped below 75% of their ideal, factory-new voltage.

Three-transistor astable multivibrator

The solution I went with is a variation of an astable multivibrator, but with three transistors rather than the usual two. You can see this in action in the second video.

With this circuit, everything continues to work as the batteries die over time, down to less than 1.5v total, with the only issue being that the LEDs get dimmer. That's much better than not working at all once it reaches 4.5v.

A simple description of how this circuit works is that capacitors are charged through resistors (which add a bit of time to the process), then transistors are turned on, which drains other capacitors, which then charge again over time.

The three larger (thousands of ohms) resistors, along with the capacitors, control the timing of the circuit. One of the resistors is smaller than the others to ensure an unbalanced state at power on; otherwise it could reach a stable start, which would mean that all of the lights would just turn on and stay on. The unbalanced state also produces a difference in timing between the LEDs, which I felt was a nice effect.

The three smaller resistors (100 ohms in my case) are primarily there as current limiters for the LEDs, but they also affect the timing of the circuit.

If you copy my circuit exactly, you will probably get similar results, but since this is not a stable circuit design, every slight variation in component values could have noticeable effects. The best option would be to try it with various values for the resistors and capacitors and see what you like. You can change the larger resistors and capacitors pretty significantly, and it will likely still work, just with different timing.

Step 5: Salvaging Electronics

Salvaging from old devices is a good way to have a variety of components for projects like this. Resistors, capacitors, transistors, switches, LEDs, and occasionally some standard through-hole ICs (many devices use very custom and surface-mount ICs, which are less useful) can all be collected.

If you have some solder wick or a solder sucker, you can use those for removing the bulk of the solder.

Then, if there is enough lead showing on the component side, use some small needle-nose pliers to grip the lead (which will help shield some of the heat from the component), heat the solder with a soldering iron, and pull from the other side with the pliers to remove the component.

If there isn't enough lead to grip on the component side, with the soldering iron, press against the end of the lead on the solder side. If you have removed most of the solder beforehand, the component may be completely free to remove at this point. If not, pull on it gently from the component side while heating the solder end. This may take several passes to get all of the leads free, especially when there are more than two.

Step 6: LEDs

The forward voltage rating of your LEDs, along with the voltage you are supplying, can be used to calculate the proper amount of resistance to prevent damage to your LEDs, but I found this less than useful. The circuitry used also causes resistance, and due to its astable nature, it was not easy to determine.

The LEDs only need the resistor to make sure they do not experience too much current. If you purchase LEDs, they should specify what their current limit is. Mine are 20ma. Using a digital multimeter, you can measure the amount of current going through the LEDs. Start off with higher-than-likely-to-be-needed resistor values (such as 1K or 470ohms) and test the current by using the multimeter set to milliamps to complete the connection of part of the circuit as if it were a wire. Between the resistor and the LED, or between them and power, are good places to use for testing.

You can do this with just the LED and resistor, but you want to find the best resistor for the LEDs in the circuit. Use the "Max" setting of the meter to see what the current gets up to when the circuit is running. If the max current measured going through the LED is significantly lower than the current it can handle, try out a smaller resistor. If it is higher than the rated value, use a larger resistor.

Use of a potentiometer (variable resistor) will help speed up this process significantly. Set the potentiometer to the max for the two terminals you will be using (middle and either side) and hook it up in place of the regular resistor. Set the multimeter up to track the max current and slowly lower the resistance, making sure to give the circuit time to go through all stages as you go.

Once you get close to the current limit of your LED (or to the max brightness you desire), stop and measure the resistance of the potentiometer between the pins that you are using. That is the resistor value you want to get close to.

Step 7: Assembling the Electronics

First I dry fit several of the components on the prototype PCB. If you are using salvaged parts (like in the demonstration video), you may have to clean off any solder and straighten the leads before you can insert them into the PCB.

I tested out and rearranged several times before I got to a layout that I liked. I attempted to optimize for placing leads of components right next to the leads of components they were to be connected to as much as possible. If that wasn't possible, having a clear path between them was the next best option. It also helps if the leads are long enough to be bent to bridge the gaps between them.

Once you have a layout you like, place the components in the PCB and bend the leads slightly to hold them in place.

Using the schematic image in an image editor (such as Paint or GIMP, or you could print it and use a pencil), I drew lines through each lead of each component as I completed soldering those joints in place. If there was a wire between two components, the end that connected to a particular component needed to be soldered in place before that lead was marked.

Once a component had all leads connected to whatever they were getting connected to, I circled the component.

Steps for soldering:

  1. Let the soldering iron heat fully.
  2. If you aren't using flux-core solder, scrape the end of the solder in flux paste to gather some onto it.
  3. Clean the tip of the iron so that it is shiny – I prefer the brass wire sponge type versus a wet sponge.
  4. Touch the end of the solder to the tip of the iron to get a drop of liquid solder onto it.
  5. Touch the iron (leading with the liquid solder drop) to the lead of the component and the metal pad of the PCB.Touch the end of the solder wire to the joint, but not directly to the iron.
  6. Once the lead and pad are hot enough, the solder will liquefy and wick up and around all the contacts in the area. If there isn't enough to fill in the hole and cover the pad, just push the solder wire into the liquid solder pool to add more.
  7. Remove the solder wire from the area, pull the iron away, and keep everything very still while it cools. The liquid solder will suddenly change from a shiny silver color to a more hazy grey color. If it looks whitish or rough-textured, then it's not a good joint. Just remelt it and try again to let it cool while keeping it very still.

For the most part, I went through all of the components and did a quick solder on each to tack them in place. Then I went back through each component and either pulled the lead straight out tight with pliers while remelting the solder, or pulled the lead sideways to a nearby component while remelting the solder. As I did this, I marked in the schematic whenever the connection between two components was complete.

I added solder to any of the joints that didn't have the pad and holes completely filled in and solid-looking. Also, when the lead of one component went over an unused hole to get to another component, I pressed the lead down and filled in that hole / pad / lead combo to secure everything more completely.

For any connections that needed to be made that were impeded or too far away to use the leads, I used a ~22 gauge wire, inserting one end into each lead from the front so that it went through the hole with the lead of the component, and soldered it in place.

As each lead / solder joint was completed, I used small diagonal cutters to trim the leads to just beyond the solder mound.

Step 8: Fiber Optics

I had several 50/125 fiber optic data cables that I had been hoping to find a use for, so I experimented with those. The insulation / PVC coating was a pain to remove. Stripping it like a wire could get me about 8” of fiber before it broke. Loosening the PVC by soaking it in acetone worked pretty well to get a couple feet of fibers, but the 125 microns (~0.005") was just too weak, in terms of physical strength and light transmission.

I ordered some 0.03" plastic end glow optical fiber, which produced some nice bright lights, but the fiber was too stiff and bulky for this project.

Finally I got the 0.01” version of that plastic fiber, which worked out very well. It feels similar to fishing line. One of the reviews I saw for this complained that it tangled very easily, so I knew what I was getting into. A piece of yarn tied around the loops keeps it pretty stable. You then need to tediously unwind individual strands.

It transmits a good amount of light, though with the end having such little surface area, it doesn't look like much unless you look at it straight on. The end of it can be enlarged by placing it near heat, such as from a soldering iron. This makes the area of the light a little larger, which makes the light more noticeable, but more importantly, provides a good surface for gluing the diamonds. And the diamonds take that very directional light and spread it out so it can be seen from all directions.

I used some rigid plastic from one of the cases of my old devices to shape a small rectangle and bored three small holes in a line with some very small scissors.

Then I formed a mounting structure with the following steps:

  1. Gather a bundle of four fiber optics.
  2. Insert the bundle a little through one of the holes.
  3. Wipe epoxy on the fibers close to the hole.
  4. Pull the bundle through a little more to get the epoxy into the hole.
  5. Repeat the above steps two more times, for a total of three bundles of four fibers.

Leave this to dry for at least an hour, and then use a razor blade to trim the little ends of fibers and excess epoxy flush with the plastic. Then use some 1000+ (the higher the better) sandpaper to polish the ends.

Step 9: Diamonds

My quest for diamonds started with looking into diamond dust / powder.

I saw several listings on eBay where light was shining down onto the diamond dust, and it looked nicely reflective / sparkly, but I was concerned that it would be hard to utilize that for a good effect with the rose. I asked a seller about how well light would refract through the diamond dust if you put some on a light, and the response was not promising.

The facets created when cutting the diamonds are a big part of their effect, so I looked for small, cut diamonds. It’s quite surprising how small they go. You can buy brilliant-cut diamonds smaller than 1mm in diameter, by the hundreds. Imagine taking a grain of sand and meticulously grinding 56 facets into it to form a recognizable shape.

When looking into diamonds, there are a few things to understand in order to know what you are getting. There are three main terms used to grade diamonds:

Cut – This is just a rating of poor/fair, good, very good, or ideal/excellent for how properly the proportions of the cut were done. In the ideal cut, all light going into the top of the diamond is reflected back out the top, along with some prism and sparkle effects. A poor cut means the light is scattered around more and “leaks” out the sides.

Color – This is an alphabetical scale of the amount of color present (or rather, not) present in the diamond. Unless you are looking at fancy (colorful) diamonds, there is a scale from perfectly colorless: “D,” to light yellow: “Z.” There is still almost no color until you get past “J.”

Clarity – This is a description of how perfect the structure of the diamond is. What you will see as the grade are the abbreviations for Flawless (FL), Internally Flawless (IF), Very Very Slightly Included (VVS1 or 2), Very Slightly Included (VS1 or 2), Slightly Included (SI1 or 2), and Included (I1, 2, or 3). The subscript numbers are subdivisions within the grade where 1 is better than 2 and 2 is better than 3. Even the worst grade here just means there are some specks or tiny cracks in the diamond. What some people might call "character" (when they aren’t spending hundreds or thousands of dollars for perfection).

Carat and shape are also part of the equation, but those are pretty straightforward descriptions, rather than measures of quality. There are also different scales used by different groups, but the above are what I saw most when looking for diamonds.

I purchased a 50 pack of “good” graded brilliant-cut round diamonds with G-H color (essentially colorless) and I1-I3 clarity. They aren’t perfection, but they are perfect for this. The cost was $16, including shipping from India. They are about 0.75mm in diameter, and add up to a total of 0.10 carats. They are tiny…very tiny, but since they were going on the ends of 0.01” diameter fiber optics, they needed to be.

I didn't get any ideal cut diamonds to compare, but since I wanted light to be emitted from all sides of the diamonds rather than just straight out in one direction, I expect the lower quality is actually a benefit for this project.

Step 10: Adding Diamond to Fiber Optics

The video gives an overview of the process I used for attaching the diamonds to the fiber optics.

I tested out a few different options for the termination of the fiber optics, and what's in the video is what I was most pleased with.

I tried:

  • Just a simple clean cut.
    • This would work nicely for projects where the fiber optics are embedded in something and only viewed from one direction, but not this project.
  • Melt the end with a flame, possibly catching the end itself on fire.
    • This actually creates a nice ball of plastic on the end that emits light in all directions.
    • It's a little hard to control fully, especially when in flames, but overall a nice result.
    • I would have been happy with this, if I wasn't already planning to use diamonds.
  • With a diamond lying on the flat "table" side with the point or "culet" pointing up, touch a fiber that had been dipped in epoxy to the culet of the diamond.
    • This worked okay, but just didn't look as great as desired.
    • There was a larger-than-desired blob of epoxy on the fiber, and the diamond didn't feel like it had a solid connection to the fiber, both in terms of physical strength and appearance / flow of light.
  • Place a diamond on a piece of tape with the culet facing down, and then bring an epoxy-dipped fiber down on to the table of the diamond.
    • There was a bit of a hassle getting the diamond to stay upright, and getting the fiber lined up with it, and held stationary (even using the Helping Hands) long enough for the epoxy to cure.
    • Overall the process was too tedious for a not-so-consistent result.
  • Secure a simple cut fiber with the end facing down, brush a small amount of epoxy on the end of the fiber, pick up a diamond, and press the table of the diamond up to the fiber.
    • This worked pretty well, except with the diamond needed to be adjusted (mostly just pushing it around with a toothpick) repeatedly until the epoxy cured enough to hold it in place. With the small diameter of the fiber and low weight of the diamond, it kept getting sucked up the side of the fiber, since in that position, the table of the diamond was against more surface than on the end.
  • Using a soldering iron, heat the end of the fiber to cause expansion before performing the above process.
    • This was best process I found. Since the end of the fiber now had a similar diameter to the table of the diamond, they clung to each other instantly and stayed in place.
    • This also provided a very nice profile of light shining into the diamond, creating the desired effect.

Step 11: Building the Wooden Base

The base is really the heart of this project.

The rose and glass dome were existing products that had been purchased, and weren't modified much. The diamond-tipped fiber optics were a lot of work and add a lot to the result, but if someone glances at this from any distance, the base is what stands out. The base is what pulls everything together, and if it looked bad, the whole thing would look bad.

I went through several iterations of testing out ideas for the base, and I have ideas for changes if I were to do it again.

I tried working with oak, but I did not like how it came out. I think there is a good reason a lot of oak products are larger square-cut pieces. The grain lines were so much denser than the wood between them that the router would get pushed out a little on them, causing a ridged effect when going across the grains. It also is so hard of a wood that the routing needed to be done pretty slowly, which easily caused burn marks. I'm sure there are solutions to these problems, but I wasn't motivated to find them.

Pine is cheap, looks pretty good, and is much easier to work. I also was concerned about expansion and contraction. Since I would be attaching a glass dome to the wood, I didn't want it to change size significantly due to humidity changes and put stress on the glass. Pine is actually one of the best choices for this issue. Cedar would be better, but it is a lot less easily obtained in wide planks.

I tried carving petals along the edges of a bevel-routed circle of wood, to match the table the rose is on in the cartoon movie, but I was not happy with the results that I could achieve in a reasonable time frame. If I was going to spend a lot of time practicing and starting over, or perhaps if I had a CNC carving tool, this could look really good, but I expected it would come out rough at best.

I tried routing a single circle of pine with a classical router bit. This looked pretty good, but since I was going to be putting electronics into the base, I wanted more depth. So I ended up recreating the classical router bit profile with the pieces of wood and the component elements of the bit: cove and beading.

For all of my testing, I used sections of the wood that were less desirable due to knots or just ugly grain patterns. I also frequently put together mock-up type assemblies of the pieces to gauge how it would look.

One issue to watch out for with wood is warping. I started with a pretty flat board, but after cutting out and working with my two "final" base pieces for a while, they sat less and less flush with each other. Because of the way they were facing when I routed them, the top piece was upside down if I placed them on each other in a way that they were most flush.

The top piece was the most warped, so I started over and cut out a new top, and paid attention to route it with the correct side facing up in the end. Think of using a cookie cutter to make two circles in a strip of dough. Remove the excess dough, pick one of the cookies up, and without rotating or flipping it, slide it over to cover the other. Make sure to keep track of the tops, since when routing, it may make sense to put the wood upside down for the profile to be the way desired.

Step 12: Cutting in Circles

I had long, 12"-wide pine boards, but I wanted round pieces for my base. I had a nice 1/4"-thick piece of polycarbonate that was about the same width as my router's fixed base and long enough to be useful for any potential future circular needs.

I placed the router on the end of the polycarbonate, marked it with a pen through four screw holes in the base, drilled out the holes, and then drilled a wider "countersink" on the holes that was deep enough for some screws I had to sit a little below the surface. I also sanded the screw heads with some high-grit sandpaper to flatten them a little and smooth out any rough bits sticking out. Then with the polycarbonate screwed to the router base, I used a couple of my router bits, including my widest one, to bore a hole through it.

Then I just marked points at the radii that I needed and drilled holes that snugly fit a wood screw I had with a smooth upper shank. The routing profile of the bottom piece of the base means that the surface on which the top piece sits is a smaller diameter than the bottom piece.

Through creating test pieces, I sized up how big I wanted to cut the top piece and made a hole closer to the router for that. I made the top piece only wide enough to cover the top of the bottom piece with about 1/4" gap around it where the top of the bottom piece would show. This way, it formed another profile detail rather than a seam on a face.

After routing the bottom piece, measure the distance from the center hole to where you want the top piece to reach. Or just set the router and jig on top of the bottom piece and line up the cutting bit with the outside of where you want the top piece and the long part of the jig towards the center. In either case, mark the jig and drill a hole for the screw. It does not matter if the hole is centered on the jig; only the distance from the hole to the router bit matters.

Here are the steps I used with this jig:

  1. Place the wood on a surface that can be cut into a little with the router, such as scrap wood, or support it securely in a way that nothing will impede the bit along the circle being cut.
  2. With a straight bit in the router, set the depth so that it doesn't protrude past the polycarbonate.
  3. Insert the wood screw in the pivot hole for the radius desired and screw into the center of the circle to be cut in the wood.
  4. Test rotating the jig 360 degrees to make sure there are no obstructions and that the router power cord can stay loose and out of the way the whole time.
  5. Start the router, and once it is up to speed, adjust the depth so that the bit penetrates the wood a little.
  6. Rotate the router and jig, providing support for the router so that it doesn't attempt to lean when in areas that don't have a stable surface under the outside half of the router. Be sure to keep track of the power cord and move it out of the way when necessary.
  7. Bring the router back around so as to not twist the power cord. Leaving it running for this pass will help clean out the cut.
  8. Check the cut and make sure you are happy with it.
  9. Lower the bit further and repeat steps 6 and 7. Depending on the type and size of wood, size and speed of your bit, and your comfort level, you can adjust the depth to cut more or less at a time.
  10. As long as your cutting bit has a shank at least as deep as your wood, you will eventually cut all the way through; if your bit is too short, you could make sure your center hole goes all the way through, flip the wood over, and repeat from the other side as well.

This might not be the safest or most efficient process, but it worked well for me. Using the plunge base is likely safer for adjusting the depth while running, but it felt too involved to me.

I used the Dremel circle cutter with an engraving cutter bit to carve the groove for the glass dome to fit into.

I set the circle tool distance to the radius of the glass dome and tried it on a test piece. After testing the fit, I adjusted the distance slightly for the best fit.

Step 13: Routing the Edges of the Circles

The cove and beading profiles were routed using a table router.

After the circle is cut out, you will want to smooth out / perfect the edge. Any little imperfections will be very noticeable when routed with the beading or cove bits.

I used the Dremel circle cutter with a 1/4" sanding drum bit (the half inch sanding drum bit does not fit through the circle cutter attachment) to sand the edge just enough to get a clean smooth surface about 1/4" down the edge. This is to have a good surface for the bearing of the cove and beading router bits to roll along.

Because of the profile of the cove and beading router bits, the bearing ends up on the bottom side of the wood you are routing. So be sure to smooth the edge closer to the "bottom" side of the wood.

Another option would be to use a round item of the proper diameter as a template to run the bearing of the router bit along.

Just be sure that the template or the smoothed section is at least slightly smaller than the rest of the profile of the wood.

The table should have a removable starting pin that is useful for seating the work piece against when using these bits with bearings. There is likely also a safety cover that goes over the bit, and a Shop-Vac hookup for removing the shavings. Use all of these features.

Here are steps for using the table router for the cove and beading:

  1. Start with the bit's cutting edge only a small amount above the table surface, and turn on the router.
  2. Press the wood up against the starting pin, then over toward the router bit.
  3. Once the wood comes into contact with the bearing, apply pressure on the right side on the wood, rotating it counterclockwise but maintaining pressure against the starting pin and the router bit.
  4. Once you have rotated the wood all the way around, you will notice a reduction in resistance.
  5. Pull the wood back from the router bit.
  6. Turn the router off.
  7. Raise the router so that more of the blade will come into contact with the wood.
  8. Turn on the router and repeat from step 2.
  9. Once the bearing reaches the smooth / template area, you can stop. You can keep going higher for the desired profile; just be certain to not raise the bearing all the way past this area.

How much you raise the bit each time depends on many factors; just start small and raise more or less based on how it feels. Pay attention to how much more blade surface is being added each time because depth is a factor in addition to height.

If there are any divots in the wood (which will be routed away once the bit reaches full height) deeper than the smooth template, avoid pressing the bearing into the divots.

Step 14: Creating a Place for the Electronics

Using a router attachment for the Dremel along with chisels, drill bits, and probably some other items, I carved a complex cavity into the upper portion of the lower base piece. I am actually happy with the results, but I would not recommend or repeat this.

Besides the complexity, this caused a situation where I needed to have the electronics in and the fiber optics protruding from the hole in the top piece before I could glue the two pieces together and finish the base with stain and polyurethane.

I would suggest the following:

  1. Cut out a simple full-depth square or rectangle from the bottom piece of the base.
  2. Glue the pieces together and finish them.
  3. Build the electronics into the space already cut out.
  4. Build a cover to enclose the electronics.

The one thing I would be concerned about with this approach would be preventing light from the LEDs escaping up through the hole that the fibers and stem go through. So you could do something to cover the LEDs or the hole and shield any light, or set up the LEDs facing as far away from the center hole as possible and just run the fiber optics around and over to the hole. Or just let the light shine through for a different effect.

I used several pieces of plastic in building this, since it can be handy when building custom structures. I mostly used the case of a digital alarm clock and portable CD player for the source material. When an area I wanted had perpendicular sections adjacent to each other, I used a saw blade to cut along at least one of the corners first.

When a single straight line cut is needed, scoring the plastic with a scoring tool and then bending it is a good solution. A pair of sheet-metal seamers provides an easy straight edge for scoring, and a clean grip to hold the plastic on one side while bending.

After cutting or snapping a plastic section, I cleaned up the edges with a razorblade, file, and / or sandpaper.

For attaching the plastic to plastic, a solvent-based cement (like IPS) works well for many common materials. For plastic to wood, the Fabri-Tac or epoxy would work well, but I mostly used a hot-melt glue gun.

Step 15: Putting the Base Together

After many iterations of test fitting / adjusting / perfecting everything, it got to the point where I no longer felt like any reasonable changes would help. So then it was time to move on and do the nearly irreversible step of gluing the top and bottom pieces together.

This would not be as concerning of a step if the electronics are put in afterwards, as I recommend. This was a learning experience, and I wasn't displeased with things enough to start over.

Since I had carved a small section of the top piece to fit a small amount of the PCB, the orientation was locked in before I got to this step, but if you just make a large enough area in the bottom piece to put the electronics in after, then it will not be so decided.

Depending on your concerns about warping and appearance, make sure you line up the grains of the top and bottom pieces properly. I wanted the grain going the same direction with the pieces in the same top / bottom orientation as they were in the board they were cut from. I felt like this makes the result look more like one piece of wood. I didn't test this, but aligning the grains 90 degrees to each other could help give it strength and control warping more, like layers of plywood.

I spread wood glue on the bottom piece since it has the cut outs, while trying not to put any past the edge of the top piece, which is a little smaller in diameter. Then I carefully placed the top piece in position, using very little pressure until I was certain I was happy with the positioning, and then pressed more firmly. I then sandwiched this between two of my test pieces, separated with a small cushion of paper towels, and held the whole thing together with two large clamps. Leave it alone for at least a day to thoroughly dry.

I tried to do this relatively soon after I had cut out and started working with my replacement top piece. My original top piece developed a good amount of warping and didn't sit well on the bottom piece. I had also not aligned the top / bottom sides correctly, and that just made the warping more noticeable. The longer the pieces are left separate after shaping them, the better chances are that they will warp.

You could possibly avoid this whole step by using a thicker, solid piece of wood, and just route a recess for your electronics. I like this idea, but am not confident in finding large enough pieces or wood in good shape at a reasonable cost, but I have not put too much effort into looking.

Step 16: Finishing the Base (Though Not Actually Done With It)

Before applying any finish, give the wood a thorough sanding. If there are noticeable scratches or dings in the wood, starting with 150 or lower grit should allow you to get it looking clean and relatively smooth. I then used 220 grit for a final pre-finish sanding. At this point, it actually has a pretty good natural appearance, but it just doesn't look, well, finished.

I chose rosewood-colored stain, mostly because of the name. The slightly red hue was also desirable. I chose spray polyurethane for the purportedly quick, easy, and quality finish.

This is the part that chemically protective gloves are a very smart choice.

Step 17: Add Color With Stain

I cut an old t-shirt into strips to make rags for applying the stain. I had a small bin filled with water for disposal of the used rags, and a small water bottle cut in half as an overflow stain container.

Steps for staining:

  1. Wear clothes you don't care about and rubber gloves.
  2. Get everything you need arranged and accessible. (You don't want to be fumbling around with stain-covered gloves on.)
  3. Shake the can of stain thoroughly. (I apparently did not do this enough for my first test, and there was hardly any pigment on the wood.)
  4. Using a flat-head screwdriver, or similar, pry open the can of stain. (You can stir the strain now if you prefer, or if it has settled since you shook it.)
  5. You could pour some stain into another container, but what I did was deeply submerge a clean rag into the stain and then wring it out into my overflow container.
  6. Put the lid back onto the can of stain and tap it closed with a mallet or the back of a screwdriver.
  7. With a good amount of stain on the rag, though not dripping, start rubbing it onto the base.
  8. Go with the grain where you can, but with a small round piece, it's more important to just get thorough and even coverage.
  9. Be sure to get in all of the nooks and crannies of the routed profile, as well as a little onto the bottom near the edge so that bare wood isn't visible from the side at all.
  10. Try not to press on any of the already stained areas with your (gloved) hands.
  11. If you need more stain, and you have used all that was in your overflow, either pour some from the can, or use a clean rag to soak some up. It is a good idea not to put a used rag into your source material.
  12. After you have covered the whole thing, wait between 5 and 15 minutes (less for a lighter result, longer for darker) for it to penetrate the wood and then wipe the excess off with a clean rag.
  13. Dispose of your stain-soaked rags in the water filled bin.
  14. If you want a darker finish, you can add another coat after 5-6 hours.

Step 18: Add Protection and Shine

After the stain has dried for a day, you can apply the polyurethane. With just the stain, the wood has a nice soft, natural feel, but adding a seal coat will help protect the finish and make it pop. There are several other options for this step besides spray polyurethane, but it seemed like the most promising in terms of quality of results and ease of use.

I think next time I would start with a can of brush-on polyurethane, and after brushing on a few thick coats, use the spray as a final fine coat.

While using the polyurethane, you should definitely wear a good respirator.

It is also strongly recommended to work in a well-ventilated area with a good cross draft. The problem is you are essentially spraying a fine coat of glue all over your workpiece, so any dust nearby (or carried from far away in the wind) is sure to become part of it. If you have access to a spray booth it would probably be less of an issue.

The can of Minwax Fast-Drying Polyurethane I used says to apply thin coats and recoat within 2 hours. It does not mention how long to wait between coats. Looking online, I found some people asking that question and mostly getting that same statement parroted back to them as an answer. I saw some saying to wait two hours. I called them and eventually got an answer of at least 15 minutes between coats.

After you have put as many coats as you want, put it in a clean, dust-free area, and forget about it for a few days. After just a few hours, it will be "dry," but it is still soft and has almost a sticky feel to it. Also it will smell very strongly. After a few days, it will harden, causing it to feel a lot better to the touch, and it will likely be done off-gassing.

While testing, I was not 100% happy with the texture of the result. It looks great when just casually looking, but if you look closely, it is a very uneven surface. I attempted several methods of smoothing and polishing it on my test pieces, including higher and higher grit sandpaper and polishing compound, with limited success. I could get a very nice smooth, mirror finish over about 80% of the area I was working on, but there were always small pits and valleys from the grain of the wood that ruined it. When I kept going, I broke through to the wood. I tried several coats of the spray, but it just wasn't thick enough. This is why I'm leaning towards brush-on polyurethane, to build up a thick coat that could be polished to perfection. Water-based polyurethane might also be better, due to far less of a vapor hazard.

In the end, I took my 220-grit sanded and stained base, sprayed about five coats of polyurethane spaced out between 15 and 30 minutes each, sanded lightly with 600 grit (to knock down any dust bumps) and wiped off with a tack cloth before a final coat, then let it dry fully until it didn't smell anymore. I didn't try to sand or polish it, and it looks really good if you aren't expecting a sheet of glass.

Step 19: Final Touches to Base

The bottom of the base was not pretty at this point. I used my router to thin down a piece of the pine, and then used my small miter and a chisel to shape it into a battery cover. I drilled a screw hole in one end and glued some plastic to the bottom of the other end to form a hinge. I recessed the area where the screw head was so that the screw did not protrude much past the surface of the base. I put batteries in place and screwed the lid closed.

I used wood glue mixed about 50/50 with sawdust to fill in the center hole about 1/4" deep on the bottom, and to tighten up the hole the switch was in.

Next I covered the bottom with felt:

  1. Wrap painter's tape around the edge of the whole thing, with one edge of the tape lined up with the bottom edge of the base to protect the finished side.
  2. Set it upside down on a newspaper-covered surface.
  3. Spray the whole bottom with medium strength spray adhesive, using smooth, even passes, starting and ending past the base.
  4. Spray one side of a sheet of felt (that is at least slightly larger than the base) in a similar way.
  5. Wait a minute or two for it to tack up.
  6. Spread the felt over the base, quickly and smoothly. Quickly pull it up for re-positioning if it's not perfect.
  7. Let dry for a while, then using some small, thin, sharp scissors, trim the excess felt around the base.
  8. Feel for the screw in the battery cover.
  9. Cut an X in the felt across the top of the screw.
  10. Peel back the felt from the screw and then unscrew it, birthing the screw through the felt.
  11. Press the felt back down onto the wooden cover and reinsert the screw enough to be secure in the cover, but not the base.
  12. Gently lift up the screw to cause the cover to move to locate the edges.
  13. Cut the felt along the edge with the screw and the two edges next to that one.
  14. Feel for the switch and cut out the felt to uncover it and allow space for it to move. Just a slit in the felt and then pressing / tucking it down on the sides of the switch provided a clean look.
  15. Add hot-melt glue to the switch to make it about even with the outside surface of the felt. The glue provides decent grip when pressed against a smooth wood or glass surface.

Step 20: Arranging the Rose and Fiber Optics

The rose I used was taller than the glass dome and base combined, so it needed to be cut.

I started by drilling out the center screw holes in the base to be able to fit the rose stem all the way through. I set the base on some small boxes to raise it a few inches off the table while giving access to the center.

I inserted the rose and set it all the way down to the table so that it was significantly shorter than the glass dome. Then I placed the dome on the base and set it into the groove that I carved for it.

I moved the rose up and down by raising the stem from below. Once I found a height I liked (which was with the edges of the large petals lined up with the start of the curve of the glass dome) I marked the location. I removed the rose from the base and cut it (with a metal hacksaw blade) so that it would roughly sit on the sawdust and wood-glue plug that I filled the bottom of the hole with later.

After the base was finished, I inspected it from all directions to determine which looked the best and was easily identifiable as the front just from a glance at the grains. I then looked the rose over and decided which of its sides looked the best, then secured it with its best face forward using a small amount of epoxy.

At this point, the fiber optics were coming up through the hole alongside the rose stem, and I still needed to get them arranged. Before I secured the rose with the epoxy, I bunched them all into a 12-fiber clump and, with as gentle of a curve as possible, snaked them around so that they were coming up out of the hole on the "back" of the rose. I then placed the small amount of epoxy so that it would not impede the fibers.

Most of the fibers were longer than necessary. I gently pulled them all out as far as they would go and separated the longest ones for the top of the rose. I taped them in place where I desired them, and pushed them back into the opening to determine if they could be short enough. If not, I cut them down to the desired length. I then did this for another set that went close to the top, then some that I had curl out to the glass and back.

Once I had them all set to lengths that I wanted, I proceeded to actually attach the diamonds to the ends. I did a few at a time since the smallest amount of epoxy that could be mixed was significantly more than necessary for just one, but I didn't want to take more than a few fibers off the rose (since attaching the diamonds involves securing them pointing down as described in Step 10) at a time and lose track of where they should go.

I did a test with Fabri-Tac to secure the fiber, since it would be much easier and faster and would be somewhat re-positional, but it did not have the desired results. The glue is made with acetone as a solvent, which damages the fiber optics in a way that causes significant light leakage where the glue was applied. This could be a rather useful trait for a different type of project, but not this one.

After I had attached the diamonds to the first batch of fibers, and let the epoxy cure, the steps were:

  1. Tape the diamond tipped fibers in place.
  2. Free the next batch of fibers to be worked on and secure with ends pointing down.
  3. Prepare a small mixture of epoxy.
  4. Attach diamonds to new set of fibers using the process from Step 10.
  5. Using a toothpick as a mini-brush, secure the previous batch of fibers in place with the epoxy.
  6. Wait at least an hour for the epoxy to cure.

For some of the complex bends, the same fibers would be secured in additional locations in subsequent passes.

I arranged a couple of the fibers so that they were sticking out to the sides far enough to hit the glass and curve back to point towards the rose. I really liked this effect, but it took a lot of placing and removing the glass dome. Just be careful not to get sloppy on the 23rd time and have it slam into one of the petals or pinch some of the fibers.

Once you have the fibers all secured in place, go down the stem securing the growing bundle of fibers on the "back" of the stem.

When cutting the fibers to length, I tried to insert them into the base so that they were roughly in the middle of the how far they could be inserted or pulled out. This way none would get pulled too tight and damage the fiber or put stress on the end structure, but they still had room be inserted back into the base as I went down the stem making them flush to the back.

This wouldn't be a concern if you do my recommended method in Step 14 of having a large open area in the bottom of the base where you can add the electronics and later the cover, because you could just have all the fibers loose on that end while attaching them to the flower, and then trim them to length and terminate them up against the LEDs afterward.

Step 21: Working With Rose Petals

This step will depend on the specific rose you are working with.

This project came about because of a rose that had already been given as a gift. There was a desire to use that rose to create a new gift. Since the rose already had sentimental value and leaving it as intact as possible was a factor, a second of the same type of rose was purchased.

I dissected that second rose to learn what I was dealing with and what it could handle. I contemplated lots of variations, such as an S-curved stem and lighting going through the rose. With a different rose, those may be possible, but not really with this one.

I carefully removed individual petals mostly using a small cut-off wheel. I painted exposed edges with gold-colored paint and sorted through them looking for an arrangement that had as little unattractive parts showing as possible.

After the primary rose was attached to the base, I used two-part epoxy to glue a pinwheel pattern of petals to the base. This also helped hide the junction of the rose to the base.

Once the petals were all in place where desired, I went through and added liberal amounts of epoxy into the open spaces around the stem into the base and under all of the petals. This was to ensure everything was very well secured so that nothing would get knocked loose from handling.

Step 22: Attaching the Glass Dome

I wasn't sure if I should attach the glass dome with epoxy or the Fabri-Tac.



  • Strong bond with little risk of coming loose.
  • Doesn't contain solvents that will damage the wood finish.


  • So permanent that if the glass dome needed to be removed for anything there was a high chance of damage.
  • Spreading the mixed epoxy into the groove for the glass dome may have been a little messy.
  • If there is any significant wood shrinkage or expansion, the bond might be so strong as to put too much stress on the glass.



  • Very good hold, especially for resisting quick jolts.
  • Able to be slowly and methodically separated if necessary.
  • Bottle comes with a nice tapered tip that makes applying in groove straightforward.


  • Unclear if there is a chance it could lose bond strength from natural conditions or time and unexpectedly break loose.
  • Uses acetone, which is a solvent likely to damage the wood finish.


I was leaning towards the Fabri-Tac, but I needed to test it. I used it to glue a small canning jar to one of the test pieces I had stained and coated with polyurethane.

The bond felt sufficiently strong. It didn't seem like accidental separation would be an issue, even if it degraded some. Especially with the glass going into a groove, the bond profile is sufficient that it would likely still stay gripping in place even if the glue became a soft gum.

I was able to purposefully separate the glass from the wood with slow and constant pressure. This left a gummy residue on both the glass and the wood.

The only less-good feature was that even after rubbing away all of the residue, there was an obvious ring of damage to the finish. However, I was more concerned about the acetone spreading away from the joint and affecting the finish around it. Especially within the glass, where likely some of the acetone would aerosolize and be trapped. I did not see any signs of damage, outside of the area that the glue was in direct contact with the finish, so I was satisfied with the results. Since this would all be done within a groove, it would likely be very easily contained.

So I used the Fabri-Tac, and I am very happy I did.

I was so excited to be finishing this that I glued the glass dome on before I did the extra securing of the petals and rose with plenty of epoxy I described in the prior step.

So I pried the dome off, secured everything very thoroughly, and cleaned inside the glass with some window cleaner and a paper towel.

Then I glued it back on for the second time.

Then I cleaned the outside of the glass and noticed all of the dust and fibers that were clinging to the inside of the glass.

So I pried it off again, got some "lint free" microfiber cloths, closed myself into my bathroom, ran the shower and vent fan for a while to get as much dust out of the air as possible, cleaned the inside of the glass with window cleaner using microfiber cloths, and glued it again.

The results were still not what I would call perfection, but probably as close as can be expected without a cleanroom.

Step 23: Review / Final Thoughts

Overall I am very happy with the way this came out; it was a big hit with the recipient.

Below is a quick overview of the broad stroke steps to create this, in case someone actually reads this whole thing straight through and has lost track of what day it is:

  1. Acquire one or two roses and a glass dome, and a whole bunch of other things.
  2. Build a small circuit to light some LEDs in an interesting pattern.
  3. Build an attractive base to hold the rose inside the glass.
  4. Make a place in the bottom of the base to put the circuit board, batteries, and the ends of fiber optics where they will get the light from the LEDs.
  5. Attach the rose to the base, with the fiber optics running up the back of the rose through a hole in the center.
  6. Attach diamonds to the ends of the fiber optics, and attach the fiber optics to the rose in an interesting arrangement.
  7. Put some fallen petals on the base.
  8. Cover the rose with the glass dome.

In some ways, it's quite simple. The one thing I want to stress most about this whole process is: test. Test every single thing you are going to do.

Once you are confident you understand what you are going to do with one aspect of the project, stop working on it and test something else. I built my initial circuit for this in just a few minutes. I didn't put it on a PCB until a month later, and by that time it was a completely different circuit. I didn't work on (nor did I neglect) the circuit for that whole month, I just worked on pieces of every aspect of this project, then went back and revisited every aspect, tested how everything was going to go together, then revised almost everything. And then did it again.

Testing this much means you get to see if something works the way you expect or hope; you get practice in everything before you really do it; and best of all, you have plenty of opportunities to stumble upon things that you wouldn't have thought to try if you were just going through the motions of implementing a complete step-by-step plan.

Please share any projects that this instructable inspires or helps you to build!



  • Sew Warm Contest 2018

    Sew Warm Contest 2018
  • Paper Contest 2018

    Paper Contest 2018
  • First Time Author Contest 2018

    First Time Author Contest 2018

We have a be nice policy.
Please be positive and constructive.




Really nice ! Great job

I love the lights! Those are a nice touch :)