We just got a contract for a few different sets of parts today, and I'm really excited to build on my template library!

These pieces will be made in Delrin, and while building the toolpath template library, I also want to build a set of tools specifically for cutting this plastic, along with specific stepdowns and feed engagements. Fusion360 has actually got quite a bit built around making tool templates, and I think building an almost automatic system isn't all that complex. It will just take a lot of time and energy to do. 

Since this project will be in Delrin, I will be making the parts differently compared to working with metal. These parts have the majority of the features on one side, and minimal work needs to be done on the other, so I will be able to use double-sided tape and pins for holding and locating my work. 

In the picture above, you can see how I set up my workpiece, taking advantage of the manufacturers modeling workspace and setup my stock there; this allows me to have a mostly clutter-free design space where I keep the 3D model separate from my CAM work, which sometimes involves external geometry references. I set up one set of the parts in a line, one in front of the other with as little space between them as I could get a 1/8" endmill through while still leaving room to clean them up. I've got a considerable margin on either end of the row of parts I will use to drill location holes. This piece plans to cut the first side's surface (to break the tension) and then do the bare minimum work. When I flip it over (using two pins at either end to keep it in the same spot), I will finish each piece and cut them out from each other. Doing it this way gives me a way to locate each piece while getting the most rigidity.

I just wanted to put out an update for the Duif Workholding Project.

Unfortunately, I've been getting swamped in my other work, and I'm finding it increasingly difficult to allocate the time needed to make substantial progress. 

I finished up the remainder of the 10 baseplates and decided to do a tumbled surface finish. I found that the 'beaten' matt look to it is really appealing, and I think it makes it look more solid, which adds to the overall feel of the product. The next step is to anodize these and get the last two parts finished up before I can start marketing and selling them!

I am considering trying to do the anodizing myself and looked into the process a bit. From what I understand, it's a fairly complex process that doesn't require too much effort once setup. 

Reviewing a few videos on the matter, I've found that there are only a few steps to do it. You start with washing all oils off your part using a degreaser and lye. You then bathe it in a tub of battery acid and distilled water with a power source that actively sends electricity to the piece, the positive connection to the part hanging down off a rod, and the negative to a plate of lead (or cleaned aluminum) and let it soak for a few hours to get it 'anodized,' basically the process right before coloring. 

Once it's been anodized and thoroughly cleaned of battery acid, you then put the parts in a boiling bath of water and dye for a few minutes before it's finished. This is just the general outline of how it's done, and there is a lot more to the process than what I outlined above.

One of the great resources I found on this topic is a video made by ShopBuitl on YouTube entitled 'How to Anodize Aluminum', where he outlines the method he used and some of the tricks he picked up.

One major investment that will bring the most value, in the long run, is building systems rather than trading time for your outcome. Before I dive into this, let me preface it with how I define systems and time. A system is something that is pre-built or a series of steps to take in a given order. For example, a system of getting a machine warmed up would be a checklist to follow, a few tasks outlined in a specific order. It's not necessarily the work itself, but what comes before the work. 

The major benefit of setting a system in place is that it removes the number of mistakes made. If something is proven, and you follow the same steps, you should get the same outcome. It reduces the amount of human input, which is where the root of all problems comes from, thus lowering your overall workload. The goal is to spend as little effort on the tasks that don't move the needle.

The downside to having a system is the time and effort it takes to build and even follow. It's not a simple transfer of work where you can move your time spent setting up the machine over to a checklist, but rather a thought-out process while you're going through it doing your best to optimize and remove the possibility of errors. 

Building a system gives you the time and freedom to work on other projects while your system works for you. It may not be the fastest way to do something, but it's proven, and you don't have to spend time trying to start it from scratch. 

In the case of my boss, who's been in the industry for 40+ years, he already has the knowledge and system built-in, so he doesn't spend very much time trying to figure out how to go about programming a specific part. In situations like this having a system isn't as critical but definitely still beneficial. He's built up an amazing speed for his work over time through repetition of working on so many different prototypes. There still is a certain amount that is just grunt work that doesn't require a lot of mental focus to get done, so having a system, or in this case, pre-set toolpaths already programmed greatly reduces the amount of time spent on the overall program. 

I drew a quick line graph showing the relationship between time spent in effort with systems versus scratch each time. The blue line represents starting from scratch with each project method, which has a moderate amount of effort and steadily gets lower. Whereas the orange line denoting the systems-based method shows a large upfront amount of effort required to put it in place, then drops below the blue line and steadily gets lower from there. You can see that the short term, starting from scratch with each progress brings you the most value for time spent, and you steadily get better, but the systems-based method has you pull a lot of effort to build it, but when it's set in place, it requires little effort to do the same amount of work. 

I'm primarily referring to manual, mindless work that doesn't require much thought to do. With the more complicated work, you want to have that direct mental input to do things manually. 

One massive, dare I say life-changing thing I learned from this rocket project is the power of the pocket clearing toolpath, and except in rare cases, I think it's greatly superior to the adaptive clearing. Saying this is going against the grain a bit and maybe a bit controversial, but hear me out. 

Clearing large shallow areas with an adaptive clearing toolpath doesn't make sense when it's completely enclosed; the calculated feed to get the same bite for each cut is definitely ideal, but in most cases, adapting from a pocket will often have equal run time going to and from the cuts, thus spending 50% of the tool movement cutting the air. If time weren't a concern, then adaptive clearing would almost be the best case in every situation.

The major downside to the pocket clearing method is that you are typically putting the tool at odd loads, and in a lot of cases, you even slot the material you're cutting. Depending on how fast you have the machine running, it could chatter quite a bit, thus leaving a bad finish. But if you calculate your stepdowns to allow for slot clearing while running full speed, you don't have to worry about taking corners too fast. This may mean that to clear a deeper pocket, it takes multiple roughing stepdowns, but it would save quite a bit of time in the long run.

Even better than the time-saving aspect of pocket clearing is how versatile it is in getting into tight spaces. Adaptive clearing won't typically let you cut a groove that isn't at least double the tool diameter, or it will take a million small steps in a circular motion down the path. I wasn't really aware that the Pocket clearing method actually allows you to take slots with minimal clearance side to side. This, coupled with the smaller stepdowns, allows for some fantastic speeds while producing a great looking part!

For this project I was able to contour the part, avoiding the pins that keep it in place, and simultaneously finish both walls all with the 3D pocket clearing toolpath. It is truly amazing and will probably be my most used toolpath 

Continuing work on the Rocket project and ran into some interesting problems related to the work origin setup.

The whole project was designed and made by my boss in Mastercam only using unstitched 2D faces. When I imported it into Fusion360, I had to do a ton of cleanup work to turn it into solid bodies to work off of. The problem was that the geometry could not move what so ever as the fixture plate was already made, and both sides of the plate were located off the same work offset.

When I ran the second operation, it came out slightly shifted and didn't align properly with the previous side. 

The problem was it was off by a weird amount, something like 13 thou, which meant that something wasn't aligned properly and wasn't an easy manual fix. My initial thought was I somehow moved the geometry while working on it, thus creating the misalignment. 

It turns out that my work zero was set incorrectly, and I selected the top right corner of the stock and not the model itself, thus making my stock the reference which, when flipped to the second operation, was offset by half the distance between the stock plate and the model. The fix was to move my work home to the proper point, which I did by sweeping a reference hole from the second op fixture, then manually moving the distance to the G54 home. I ran the first operation once again, getting the correct alignment, and will find out tomorrow if I set everything up properly when I run the second op.

So moral of the story, make sure you set your work zero at the same point as the previous software used to make the part.

For the Duif Workholding baseplates, I ended up deciding to tumble them instead of trying to polish before sending them out to get anodized. I'm actually delighted I went this direction as the surface finish has a very large satin look to it (large in the sense of the visual pattern). It blends everything really well and makes it look more solid.

I found an absolute game-changer of a tool in Fusion360 today that I had no idea existed before this point! I watched one of Lars Christensen's what's new in fusion360 videos, and he shared about the manufacturing modeling tool. This is a sub-design area completely separate from the design workspace but has almost all the same tools. The difference is that whatever changes or features you make in this design environment do not apply to the actual model. 

For example, let's say you have a cube with a slot straight down the center of it, and you want to program a toolpath to completely ignore the slot. You would normally create a sketch or body and select that as your boundary, thus adding an extra step swapping back and forth between the design and manufacturing workspace. This is somewhat sketchy as the reference geometry isn't tied to the model itself. With the Manufacturing model tool, you can select the slot and use a quick edit tool to completely remove it, making a ghost-design that your programming references off of. 

It's also got a feature editor that allows you to easily remove fillets and chamfers, as well as whole blocks of space like holes, pockets, and other features. I've only just scratched the surface of the capabilities this tool has, and I cannot tell you how excited I am to find this out!

I'm also working on a production 5-axis project, which required me to cut many dovetails in blocks of stock. I took advantage of the manufacturer model and created my stock reference model and made it so much easier to edit and keep the design workspace free of random components and bodies! I'm also using this as an opportunity to build up my CAM template library and added the dovetail operation to it.

Working on the rocket project again and ran the first operation today!

I'm honestly a little shocked by how well the new programming techniques I picked up over the past few days worked! Everything was swift and efficient, with amazing surface finish results! 

I'm really pleased with the surface finish on the rocket's fins and took full advantage of the radial toolpath to get an exploding star-like pattern. The downside to using this toolpath is that as the lines get farther from the center, the stepovers get increasingly larger, so you can see the thrusters' bottom got clear cusps. 

I picked up an interesting technique today while engraving some lettering on the part I'm working on. I couldn't use an engraving bit due to the sharp nature of the centerline that wouldn't look right on this piece. I was limited to using a 0.04" (the thickness of your fingernail) ball endmill to cut the letters out. I initially tried to generate a normal 2D pocket toolpath, but it gave me a few errors and wouldn't work right with some of the font's tight corners. So my boss suggested tricking the program into thinking it was a smaller tool than it actually was, thus defining all the sharp corners and features. The downside was that if you used the lettering's straight contours, you would have an oversized and bloated look to the text. This is where I took advantage of the engraving; by only cutting five thou deep, I could use the tip of the ball endmill to touch the surface, imitating a smaller endmill with lighter cuts. 

I also was able to finish up the rest of the baseplates for the Pocket NC Duifworkholding project. Still figuring out exactly how to polish the plates' surface and get them ready to ship it for anodizing. 

Working on the rocket project again this morning, trying a slightly different approach to programming today. 

One of the things my boss would sometimes get on my case was how much I relied on Fusion to program "for me," referring to the 3D toolpaths that require very little to generate a workable program. He was brought up with all the manual machines, so I figured this was more of a mindset thing than anything else, and since he doesn't regularly use a lot of the auto-generated tools, he didn't really know the power they had.

Starting at the makerspace, I was mostly self-taught when it came to programming, so I mostly did what worked and didn't try to improve it as I wouldn't work on the same project again. Because of that, I primarily used 3D adaptive as my roughing and finishing toolpaths, which worked with a lot of time-consuming tweaking. The problem is that that mindset carried over to my current job, and I am constantly struggling to get over it and use a systematic method to program the parts. This rocket project has been especially complex due to having to finish both the parts and the remaining stock around them. Basically, cutting very clean grooves between piece and stock, in contrast with most parts only requiring a good finish on the product and the wasted stock, would get removed regardless. 

Today I decided to start building a series of steps that I can use to produce any given part. I also want to create many CAM templates that I can use for future projects; this will require a lot of extra work upfront and remembering to make any edits in the program itself even after the parts are finished up if I changed anything on the machine to ensure I have proven toolpaths.

I've gone back and forth quite a bit on finishing some of the faces in this project and the best way to go about it. I found that manually drawing faces with specific boundaries then using hard geometry in 2D pocket clearing toolpaths works surprisingly well. Using this method, I simultaneously finished off the parts' flats and cleaned the side walls in a single toolpath. 

The main drawback to manually drawing out your geometry is that your reference drawings are no longer up to date with the actual part if you shift the hard model. Basically, if you have a square box (your part) and you create a sketch on one of the faces of the box, then shift the box to the side, the sketch does not move with it; thus, any program that references that sketch will stay the same, while others will shift with the part. Generally, once you have a part setup, you don't really make changes to its position, so this isn't that big of an issue. 

In summary, creating manual sketch geometry to use as a reference in your CAM programming is an absolute time-saver. It seems like it would take longer to create a custom sketch or body for each major feature of your project but the time saved trying to manipulate the toolpath settings more than makes up for the perceived time spent. The downside is you're creating an external feature that is not directly linked to your 3D model, but in most cases, this is not a problem. I can definitely see how this one trick makes up a good portion of how fast my boss programs. 

Last Saturday, I got to put in some extra time on my Duif Workholding project, and my goal was to produce ten sets of the dovetail vice with baseplates. Unfortunately, I severely underestimated how long it would take and found myself five hours in with only a few of the baseplates complete. 

I haven't really done this small scale of "production" work before, and so went about it in a slightly different method than one-off parts. Run time and consistent quality were something I had to take into account, so I had to program the piece so that it would only require me to come and swap the parts out when it finished. 

I started by cutting many plates from a larger piece of bar-stock to use for the baseplates. I did this on the large bandsaw, which has a self-feeding mechanism so I could set it running and walk away. Because I wanted consistent stock sizes, I set up a quick and dirty hard stop using a chunk of aluminum to slide the large bar against. However, I didn't consider that since the diameter of the metal was so large, that as the bandsaw blade was cutting through it, the back slot where it just passed through would crimp together. Thus applying pressure on the blade prevented its movement and caused it to catch in the material. I didn't think this would be an issue as the stock was only resting gently against the hard stop, but even that gentle touch was enough to push the hot metal inward slightly. My boss gave me the solution: to have a finger spacer that you would put up against the hard stop and set the stock in place, then remove the spacer before clamping it down. 

For the milling side of the project, I was able to take advantage of using soft jaws to hold the round stock in place, then milling the backside of the plate, flipping it over, and doing the final side. Starting with the backside gave me a stable workpiece to sit completely flat on when flipping it over versus doing the front first, which had angled sides making it difficult to get a good clamping grip on. I made the mistake of tightening my part in the jaws too much, which didn't manifest itself in a problem until I cut the dovetail in the center groove of the part. Because the plate was quite large, I could only grip it by a small portion on either side, so when you clamp it together, the center gets pushed up ever so slightly. When I cut the dovetails, it acted as a spot to release pressure and actually pushed the aluminum into the cutter itself, leaving a very odd surface finish. My solution was to take lighter cuts with the dovetail instead of re-cutting the soft jaws. Definitely, not a great solution as the material still pushed upward, but it worked for what I had. 

I sadly only had the time to finish up seven of the baseplates. Still, the rest is ready to go for their final operation, and the programming for the final two parts is mostly finished, so I should, fingers crossed, get these completely done within a week or so.

Working on this and a self-centering vice project (new post to come about this) brings my hours to a total of 36:00:00 for this project so far. I'm getting close to finished with the products and should hopefully start selling them shortly.