FIVE FROGS WOODWORKING

Second Challenge – reducing or preventing tear-out roughness on the backside

Once the alignment issue was resolved, next was to reduce the amount of tear out on the backside of the puzzle. At first, I attempted to do this by experimenting with different brands and styles of blades. Blades are sized by numbers. 1 is smaller than 3 which is smaller than a 5. If you want smaller than 1, you see numbers that look like 2/0, which is spoken “two-ought”. Now as the numbers get bigger, the blades get smaller. 3/0 is smaller than 2/0 and so on. All of my 1/4″ thick puzzles are normally cut with 2/0 or smaller blades.

As the blade moves up and down on the wood, the teeth are typically pointing down at the table. The sharp tip of the blade cuts through the wood, leaving a smooth top side. On the bottom, that sharp tooth pushing out of the wood causes the bottom edge to tear out, leaving a felt roughness. With a wooden backside, that is no problem. When the puzzle is done being cut, you do some light sanding to smooth it out.

The first puzzle I cut (shown in my last post) was cut with Pegas MGT 2/0R blades. The “R” in 2/0R means the bottom section of teeth is reversed so they point up rather than down. This can give a very smooth bottom cut, depending on where the table sits relative to the blade. In this case, they did not give a smooth cut.

In the highly magnified picture below, the left-hand image shows the top of one of the pieces. The top edge of the cut is nice and smooth. The cut is very crisp. That same piece is then rotated a little bit to view the bottom of the piece. You can see how rough it is in the right-side image. The photographic paper is torn and curling away from the cut. When the puzzle is assembled, that backside has a rough feel to it.

I tested about 12 different brands and styles of blades. I ended up cutting a total of five puzzles while experimenting with blade types. I found one blade that was noticeably better than the others (for this purpose!). Notice I said better, but not good.

The blade I found to give the smoothest cut on the bottom side was the Flying Dutchman 2-Way Cut blade. (I am abbreviating these as FD TC blades.) It has two teeth down and one tooth up, repeated the entire length of the blade. The negative part is that you have teeth pointed up coming through the top of the cut, giving a little rougher cut on the top. Here is a drawing of the blade. Flying Dutchman has another very similar blade called the Flying Dutchman Ultra Reverse (FD UR). It has the same tooth arrangement but costs a little more. I did not find a significant difference between the FD TC and the FD UR blades.

Here are some images of the cuts on one of the test puzzles using three different blades with no other tear-out prevention measures.

After these five puzzles, I was getting a little frustrated. I was pacing around the shop and happened to look over at my table saw. Bam! Right there was the answer for me: a zero clearance insert!

When cutting wood on a table saw and you want to minimize the splintering, you place a new blank piece of wood in the throat on the table, and then slowly raise the spinning blade through that new piece of wood. This causes the wood to be very tight around the blade and provides support to the underside of the wood being cut. This reduces splinters. (They do need to be replaced periodically as the blade slot gets wider and wider with use, like the one above.) Now how to do that with a scroll saw? As you can see below, the scroll saw has a rather large hole in the table with no allowance for making zero clearance inserts.

I happen to have some 12″ x 12″ sheets of brass stock. I am not sure what they were originally intended for, but I use them for getting signs/plaques engraved and for shim stock. I took one and cut right down the center of it until I was part way in. At that point I taped it down to the table.

Now we are talking! I started cutting another puzzle. As I cut, one problem rapidly became apparent. The sharp edges of the brass were catching on the puzzle as I spun it around, trying to cut/tear the bottom image. That is not good, so I took the next step. I have a large flexible white board that is a refrigerator magnet. It is too soft and flexible to prevent tear out but will work great on top of the brass. I drilled a little hole and placed it over the brass plate to smooth the transition out.

This was a vast improvement! I ended up cutting four of the puzzles I used for the exchange using this zero-clearance plate set up. The image below shows the FD TC cut with the zero-clearance plate on the left, and without the plate on the right. It is a very noticeable difference.

After the fourth puzzle, I noticed the edges were getting a little more ragged again. This is because the more I cut, the more the side to side and back-to-back movements of the blades made the hole in the brass plate get bigger and bigger. This provided less support to the bottom of the puzzle, resulting in increased tear out.

So, I had to either replace the brass plate or find another solution. I suppose I could have just moved the brass plate over an inch and made another slot, but that is a spendy piece of material and I did not want to completely turn it into brass confetti by the time I was done. I threw in the towel and did what I did not want to do. I ended up using a piece of sacrificial plywood on the bottom of the next nine puzzles for the exchange.

By attaching another piece of plywood as a backing board, it does the same thing as the zero-clearance concept. The saw blade exiting the puzzle passes into the second piece of plywood, which acts as a backing board and prevents the tear out on the puzzle. The second piece of plywood does get tear out, but who cares?

The wood I used for the sacrificial backing board is a thinner MDF product with thin alder veneer layers on the outside. These ended up being puzzles in their own right. Once sanded to remove the tear out whiskers, they can be painted however a person would like. The image below shows how I glued this material with a hot melt glue gun to the puzzles to be cut.

As long as the backing board stayed tight against the puzzle board, this gave great results. While gluing the two together, I selected backing boards that matched any cupping or warping in the puzzle board to try and ensure this. The image below shows this tight relationship which resulted in almost no tear out on the puzzle image.

However, as the days went on, those assemblies I had glued up started to expand/contract with the temperature changes in the shop. We went from 70-degree days to 98-degree days with nights getting down into the 50’s. I only cut in the mornings while it was cool, but the wide temperature swings over several days caused some of the assemblies to develop permanent gaps between the puzzle blank and the backing board. This effect would have been reduced if I did not glue up all 9 puzzles at one time and let them set in the heat. Lesson learned.

For these glued up assemblies with gaps, I had to press down firmly to close the gap while attempting to spin the puzzle blank around to cut the pieces out. Pressing down while spinning is not a great thing. It makes it hard to control the cut. On some of these puzzles, my figurals really kind of look a little mangled.

So, the 13 puzzles I provided for the exchange have minimal tear out on the backside and feel relatively smooth. About five of them are exceptional in that regard. All of them are good puzzles from a puzzling standpoint, but a few of them are not as smooth as I would have liked. The next two images show the front and back of one of the 13 puzzles exchanged.

Summary

These puzzles were a great project for me. I learned quite a bit in the process. When everything was just right, a nearly perfect double-sided puzzle resulted.

Each puzzle had five figurals. In the center was a reduced size frog signature piece. At the three o’clock position was a wristwatch. 12 o’clock was a mantel clock. 9 o’clock was an hourglass. 6 o’clock was meant to be Old Man Time but ended up being the grim reaper.

I cut a total of 18 of these puzzles. The last 13 cut were used in the puzzle exchange for the 2024 Puzzle Parley. Of the first five cut, two were sold as “seconds” at the table sales during the Parley. Three remain with me as of this writing. I will not be listing them on my Etsy store, as I consider them to be seconds. If anyone is interested in obtaining one, contact me and we can work something out.

I also had 8 of the puzzles that resulted from the backing boards. Two of these were painted as examples of what could be done with them. Six were left bare as “Bare Naked Wood Puzzles”. All of them were taken to the Parley as well, and five were sold. The remaining ones are available for sale by contacting me direct.

I did some other experiments while cutting these as well. Not part of the challenge per se, nor part of this write up, but I experimented with using 5 different types of wood, several types of different papers including two metallic papers, and different finishes to protect the paper.

Happy Puzzling!

Bob

I have wanted to create a double-sided puzzle for a number of years. The first one I ever encountered personally was from the 2012 Puzzle Parley. It was created by Fools Gold Puzzles and has a reverse image on it to make the puzzle more difficult. This puzzle is superbly done, including figurals of butterflies that are perfectly aligned front and back.

In 2022, I made my first double-sided puzzle using an image supplied by a customer. This was of a bell pepper that had been cut in half and formed a face. The puzzle would actually be two puzzles, one for each half of the pepper. I ended up making three attempts before I was satisfied enough with the outcome to send the puzzle to the customer. In this case, the image on the backside of the puzzle was the mirror image of the front side. If done well, it makes it harder to solve the puzzle, as you do not know which side of the piece is up or down.

The problem I had with this puzzle was the amount of tear out on the reverse side of the puzzle. The side that was up when the puzzle was cut was nice and smooth, but the side that was down was very rough feeling. The three sets of puzzles were fine as puzzles, but due to the roughness on the back, it was easy to figure out which side is which. I made this one project and had not tried another one.

A few years ago, I found the scanned images of a pocket watch built by the firm of Moynier & Fils in about 1825 (shown at the top of the blog entry). This public domain pair of images is from the New York MET digital library. This is a perfect pair of images for a double-sided puzzle. I have been sitting on this for a number of years, waiting for the perfect time or incentive to make the puzzle.

The 2024 Puzzle Parley took place last weekend, which includes a puzzle exchange between some of the participants. The guidance for the exchange puzzles included that the puzzle should be around 6″ x 6″, have 60 or more pieces, and should challenge you in some way. This was the perfect opportunity to use this image to once again attempt to make double-sided puzzles. In this case, it is rather obvious which side is which. The advantage to this is that you really would get two different puzzles in one. The watch face is relatively easy to solve, the backside is relatively harder to solve. Even if I could not solve the roughness of the cut issue, it did not really matter. The resulting puzzles would still be good puzzles.

The two challenges for me in making this puzzle would be to 1) ensure that the front and back are perfectly aligned and 2) minimize how rough the back side of the puzzle is.

First challenge – Aligning the two images

If you are making a square or rectangular puzzle, it is relatively easy to reverse the image and add reference alignment marks. For the bell pepper puzzles, I added reference marks on the image using software, printed the image, flipped the image 180 degrees and reprinted it.

Next was to mount one image on the wood, and then extend the alignment marks to the edge of the wood and transfer the marks to the backside. Once the marks are on the backside, position the second image while matching all the marks. Easy peasy, right?!?!

For the pocket watch, it was not so easy. The images were different, so I could not just flip one image and reprint it. Additionally, the two images were not scanned exactly on the same vertical or horizontal axis. My first attempt at alignment was to make reference marks on each individual image based on obvious common points of the stopwatch, such as the scalloped edges.

The next steps were as before with the bell peppers: mount the first image, extend the marks to the edge of the wood, transfer those marks to the backside, and mount the second image.

As you can see from the next photo, the positioning of the reference marks was not really very consistent between the two images. I cut this puzzle with the watch face up. As we look at the back, all of the grey on the right edge is material that should have been cut away. If you look at the left side, some of the watch edge is missing.

I tried this method for the first two puzzles I cut. Deciding that it was not working out, I changed the method. From this point on, I mounted the front side of the puzzle. That was taken to the saw, and four portions of the board were cut away. Matching portions of the second image were cut away with a razor blade and then the second image was mounted on the back.

This turned out to be a satisfactory method of mounting the fronts and backs, even if it did take more effort and time than the first method tried.

I have decided this is going to be a multi part posting. The next post will discuss how I overcame the roughness on the backside of the puzzles.

Make sure you look for part 2 to read more of my journey on this puzzle.

Happy Puzzling!

Bob Walzer

I am going to walk through the steps I use to mount an image to the wood. This is my next project to cut. The customer emailed me with a digital file she wants cut into a puzzle. It is an AI generated image that came from some sort of free image sharing website. In our discussion, it was determined the puzzle is to be at least 200 pieces. It is square, and with my average piece size, this turns out to be about a 9.5″ x 9.5″ puzzle. That is a little unique and required me to print it on a 13″ x 19″ sheet of paper. Since my in-stock plywood is already cut to size, it will also require using a piece of plywood that size.

Here is the printed image.

Next was to trim it to a more manageable size. First, I used a straight edge and box cutter. I was able to salvage the bottom blank section as an 8.5″ x 11″ piece. Maybe I can use it for a Puzzle Parley exchange puzzle.

Next, I cut the image closer to size with my rotary cutter.

Here is the trimmed image.

This is going to be mounted on a piece of Cherry Appleply. My wood has been pre-sanded down at the wood shop before bringing it up to the temperature-controlled house. Here is the backside of the puzzle.

The side that the image will be mounted on is swept clean once again, and then rubbed with a tack cloth to remove any remaining dust.

In your press, you should have two large pieces of release paper that are the size of the platens. One side of the paper is shinier than the other side. That shiny side is the side that goes towards the item you are mounting. This paper can be reused many times, but make sure it is clean and does not have any residual adhesive from prior projects.

One of the key elements of successful bonding is to minimize the amount of moisture present in the process. I store my cut and sanded wood in the basement workroom of my house where the temperature and moisture are relatively stable. Additionally, some sources (including the dry mount press owner’s manual) recommend pre-drying the materials to be mounted. To do this, wrap the mounting material with brown paper or kraft cloth and place it in the warmed-up press. This tends to drive the moisture out of the wood and into the paper. If your wood is particularly damp, you may want to do this step several times with different pieces of paper to get it as dry as possible. I just use a paper bag from the grocery store. I put this in the press while I am getting the image and adhesive ready. Since I am only using one end of the board, I am not covering the entire thing.

While that is heating up, I get the image and the adhesive ready. I use Fusion 4000 adhesive. I have a large roll I bought years ago. It is getting pretty well used up. I found out this stuff is no longer made, so I went on a search to buy some new old stock. I found a package of precut pieces on eBay, which is what I am using for this project. I have since bought two more large rolls from a source in Great Britain. Believe it or not, the rolls from Great Britain ended up costing less than my original roll I bought in 2012, even after factoring in the shipping cost.

Select a piece and trim it to the size of the image. You want it just a little bit bigger than the image, maybe about a 1/16 or 1/8 inch all around.

Once the adhesive is trimmed, you need to prepare the back of the image. I have learned that some of the modern photo paper is treated on the back so that it feeds through the printers more smoothly. Unfortunately, this treatment can cause adhesion difficulties. It has been recommended that the back of the image be lightly abraded with sandpaper. You are not sanding it, just roughing it up. I just take a single pass over each area.

Next, I wipe off any dust from the process and then final clean it with isopropyl alcohol.

Next, place a small piece of release paper on your work surface. Place the image face down on the release paper. Place the adhesive on the image back. Place another small piece of release paper on top. Use a tacking iron to melt one side of the adhesive to the image.

This anchors the adhesive to the image. If you try to put everything in the press without this step, you will have a harder time.

Next, pull the warmed-up wood from the press and remove the paper bag. (Remember to close the press to keep it from cooling down.) Let the wood cool enough that the adhesive does not melt. Position the image/adhesive assembly where you want it on the plywood. Use the tacking iron and small piece of release paper to tack the opposite end of the assembly from where it was tacked in the last step.

Now, load it all in the warmed-up press.

• bottom piece of release paper
• plywood
• adhesive
• image
• top piece of release paper

Make sure the image and adhesive are properly lined up and the top piece of release paper is covering them.

Close the press and clamp it. Cook it for however long your adhesive requires. I use 4 minutes.

Once the time is up, pull out the assembly and place it face down on a clean surface. Place some weight on it until it cools.

The mounting process is done! I take one final step before taking it down to the wood shop to cut. I spray a protective finish on the image to protect it from finger oils and spilled drinks. I am currently using an archival varnish that claims to be non-yellowing. I have a little spray tent with a box fan behind it that blows the fumes out of the window. I also set up some cardboard ducting behind the fan to ensure the fan discharge goes out the window instead of sneaking back around into the room. I will not spray lacquer in the house with this set up, but I find it is more than adequate for varnish and paint. Lacquer is nasty stuff.

I know some people use a plastic laminate on their puzzles, but to me that is just one more layer of stuff that might peel.

Here is the sprayed image, ready to be cut! Need to let it set and dry for the specified amount of time before handling.

So, that is my mounting process. Between adjusting an image just right to print it, printing it, mounting it, and spraying it, there may be between 45 minutes to an hour invested before it even gets close to a saw. That does not include the time that goes into cutting and sanding the plywood when it first arrives.

When I started cutting puzzles, I really struggled with images lifting off of the plywood backing material. Between that and the actual paper separating itself, I spent a lot of time researching how to correct the problem. I do not claim that the way I do this is the right way or the only way, but it is the method I use that has brought me the best results.

I hope you found this interesting reading and maybe even helpful to you.

Happy Puzzling!

Bob