Since the coronavirus hit the country, I haven't been as motivated to work on furthering my skills on the Pocket NC at Oneshop. Unfortunately, Oneshop's owner has decided to close up shop permanently and sell off the equipment; this means I no longer have access to the machines. The makerspace hasn't had that many members, and so it was overall a net loss for the owner; aside from myself, there were maybe one or two other active members.

I offered to buy the Pocket NC V1 that he was going to sell. However, his asking price ($2,200) was more than what I thought it would be worth taking into account it being their Kickstarter model and doesn't have anywhere near the number of features their newer models have. I still wanted to get a Pocket NC to work on and decided to take advantage of the unique situation to even upgrade to one of their newer models. I couldn't afford to buy one new ($5,999), but the going rate for used machines was closer to my budget at $3,600-4,000, and I would only need to wait for someone to put one on the market. While waiting, I could then put together enough cash for it so I wouldn't be left penniless after the purchase. I posted on the Facebook and Google forum groups that discuss Pocket NC related information that I was looking for a used V2-10 model.

The following day I received a message from Pocket NC directly asking if I wanted one of their machines at a reduced price to continue posting about their machine and things I ran into while learning how to use it. I was super excited at this because it meant I could get my hands on the machine in a week or two instead of waiting an indefinite amount of time for one to pop up on the market, not to mention it is at a price I could afford pretty comfortably. 

They said they would ship it out this coming Tuesday and should arrive in a week or two! I am extremely happy with how things are turning out, and as the saying goes, when one door closes, another one opens up! They even asked if I've considered starting a YouTube channel and post content there in addition to my blog and Instagram page. I have thought about it, and this may just be the time to start one up!

After cleaning the Pocket NC, I ran the same operation that failed last time. I was able to complete it without fail, however, when I went in for another side milling job the b-table (the bed the clamp is attached to) wouldn't rotate. I attempted to re-home the machine, but it would keep giving me this error:

"Home switch inactive before the start of backoff move." I completely unplugged the Pocket NC and restarted it, again the same issue. I Then took the vacuum I'd been using and clearing away all the chips once more, thinking it must be something to do with chips flying and touching exposed wires. I also took the air compressor and did a quick once overall around the machine. Again same error, finally, I decide to unplug and plug back in the motor that rotates the b-table; this fixed the issue. I'm not entirely sure why but I think it may not have been plugged in all the way, thus allowing it to work when in the right position occasionally. I also had the same issue with the z height motor and the same solution.

Unfortunately, since I was playing it a bit safe with the toolpaths the run time for the last operation was 21 minutes and had already run through about 18 minutes. I really didn't want to sit there and watch it cut air for that long so I decided to do something a bit risky and stupid. I edited the code and removed about 60% of the lines to get it to start toward the end. However, I neglected to realize that it would go the shortest possible path from its starting point to the next line in the code. I ended up snapping my 1/8" endmill because it attempted to chew through 1/4" of aluminum at 34 inches per minute. I was very tired when I did this and if I thought about it I would have realized that would have happened and so decided to call it a day and head home.

Like I mentioned in my last post, I was not able to complete the aluminum impeller due to weird machine outputs. I would run the operation, and it would seem all fine for about a minute then when moving in on its z height, it would incorrectly think it was in position when about 1/4" away and start it's cut. I thought maybe it was just a programming error and went to check the operation simulator and found it worked as it should. In addition to this, it would also periodically pause every couple seconds, then continue. I took a video of the issue and sent it to the support team @pocket_nc for their advice. 

Here is the video I sent them: 

They got back to me within an hour with a few possible solutions to my problem though they said that since this is their Kickstarter model, they didn't have as much information and problem-solving solutions as their newer ones.

They said the issue could be chips interfering with the wiring on the stepper motors causing it to think it's moved when it hasn't. The solution is to thoroughly clear out all the debris on and around the machine. As for the pausing issue, they think the issue is related to the yellow pause button on the front of the Pocket NC and that something inside might be either pressing it or sending a signal to halt the operation then continue. 

I just finished cleaning the machine and even lightly lubricated some of the moving parts with Super Lube (synthetic grease). I just started the same side operation that I had errors yesterday, and I am crossing my fingers this works!

After my last post, I generated some new speeds and feeds with the calculator provided by NYC CNC. I decided to play it safe and take the lower ends of the output to use on the Pocket NC as it has significantly less power than the machines the calculator was initially meant for. When I was reading up on the information around speeds and feeds provided by @saunders they mentioned that there is a balance that you need on the side load of your endmill where if you cut into your material too much your tool of course can't eat it all, and, this is what I found interesting, if you don't cut into it far enough it can cause the edges of your bit to rub against the material not allowing it to take a large enough bite to properly work. 

Side note: I set the stock on the bed diagonally rather than perpendicular like I was before. I did this because, in the past, it seemed like no matter how straight I thought I would put the stock in the clamp, it would always seem to be slightly off, which has put me in some pretty close calls with my endmills. Since the stock was perfectly square, I had no issue in inserting it into the vice. 

Right away, after starting the first operation, I was getting some excellent results. In the past, I would often set the sideload to be larger than it could handle; this wouldn't normally be an issue in other materials, but aluminum is not as forgiving and would require a lot more playing around with the settings. Cutting feed rate: 34 in/minute, side load: 0.006" and maximum stepdown: 0.13". I was getting a very consistent flow of chips all being large enough to prevent the endmill from rubbing the material.

I attempted to increase the sideload to 0.0025" as the recommended starting input, but the machine couldn't quite handle that and chattered quite a bit. I also noticed that when the machine was pulling the tool back to itself, it would screech pretty bad. I have a feeling this is due to it being pulled and not pushed like when going in other directions.

When the initial operation was about a third of the way through, the sound started to change from a gentle hum to a low rumble and got increasingly worse as time went on. I eventually stopped the machine and found that chips started to weld onto the endmill. I cleared out the aluminum and set the compressed air setup I used for the Shapeoko to get the chips out of there as soon as possible as well as cooling down the machine, which fixed the issue right away. 

I was able to complete the top operation without any further issues but when attempting to do the sides I started having some weird errors which I will touch on in the next post.

It's been a minute since I last posted. I got busy with other work and didn't have that much time to continue my education on the Pocket NC/Fusion360, then the virus struck and threw things off even more. 

Anyhow, after my initial test cuts with aluminum (see my last post), I knew I needed to do things differently. 

Someone on Instagram commented in the post I made about the aluminum cuts on the Pocket NC, mentioning it may be deflection and that the clamp system can't keep it stable enough for the endmill to take a more significant bite. He said that more clamping might fix the issue or at least help quite a bit. Unfortunately, due to the way the Pocket NC clamp system on the v1 is designed, it's difficult to add any other workholding fixtures. He suggested I should put solid block covers over the grub-screws to absorb some of the vibrations. 

I also wanted to read up more on speeds and feeds in general as up until this point; I've just been using some recommended rates then adjusting them slightly to my needs without knowing exactly how it all worked. I went to NYC CNC by Saunders Machine Works and found a getting started with speeds and feeds post with a video.

He went pretty in-depth, explaining some of the startup recipes, calculating speeds and feeds, and even chip thinning and horsepower. At the bottom of the page, there is a download to speeds and feeds calculator with a ton of beneficial information when programming new operations. 

I imputed the data according to the Pocket NC and tooling specs and have new speeds and feeds to test next time I get to Oneshop. I found that though the new speed rate was very similar to what I was already using, the depth of cut and sideload were pretty significantly different. I'm a bit wary of if the machine can handle these rates and may bump them down a bit and work my way up so I don't ruin any endmills.

Before milling the polyurethane, I tried to do the same operation in aluminum. 

I started by using one of the blocks I cut and surfaced one side to ensure I had a perfectly flat face to work with. Unfortunately, I wasn't able to use the same speeds and feeds I was using on the Shapeoko and ended up almost snapping one of my endmills when attempting too aggressive of a cut. I cut down the bite it was taking and tried it again, it worked, but since it was taking away such little material, it would take forever to do the entire operation. One way to cut down a lot of time was just to cut the rectangular stock and make it square; this took quite a bit longer than I was expecting due to the way I was cutting it. After that, I kind of gave up on cutting aluminum for that day and decided to try the polyurethane instead.

I'm going to do a bit more research and find out exactly how far I can push the machine on the aluminum without breaking the endmill. However, since the Pocket NC can only run at a max RPM of 8,500, it's very limited in how fast it can eat the material.

After making the impeller so many times (several not posted about), I kind of realize it's going to be nearly impossible to get the type of finish I want with wood. It's just way too fragile and breaks in chunks, not ideal for small delicate parts. Apart from the stock I was given by the local shop owner was a cylinder of clear polyurethane. I've only worked with wood and aluminum before, so I was unsure how hard it was to cut. I assumed since it was a plastic-based material, it would be pretty soft. I was wrong, and after nearly breaking one of my bits from setting the speeds and feeds to be too aggressive, I had to slow it down quite a bit.

I ran the same operations I did for the wooden versions and adjusted a couple of the settings to keep run time down. I used 3D adaptive clearing with a 1/8" ball endmill with 0.003" stock to leave so I could come in with the same tool later and clean it up. It took about 4 hours to do the roughing passes, at which point I didn't have enough time to finish it that night. 

The next day I came back since I didn't remove the part from the machine I was able to pick up where I left off without worrying about having to restart. I used morphed spiral toolpath to finish it off since there was only 0.003" of stock to remove. I set the speeds to be relatively aggressive for this tough of material at 30 inches per minute. For the stepover, I set it to 0.01," which is a bit higher than I usually would go for a finish. I got a pretty decent finish on the part with a very smooth underside and no broken blades this time. I then used a 5-axis swarf toolpath with the same 1/8" ball endmill to simultaneously finish the bottom of the part and cut it off the base.

Third attempt at making an impeller on the Pocket NC

I fixed the CAM toolpath from last time, so I thankfully didn't have the same oversight. I was able to get all the roughing passes with only one error. I accidentally set one of the last toolpaths to read as a 1/8" square mill when I was using a ball mill. I was able to catch the issue pretty quickly and restart it with the right tool. I have noticed the software that controls the Pocket NC is a bit finicky and crashed for no reason on two occasions. 

After the roughing finished, I used the same 1/8" ball mill to do the finishing passes. I set the toolpath to do parallel passes with 0.007" stepover at 30 IMP. I noticed pretty quickly it would dig into areas where it couldn't reach before and cause everything to vibrate, which broke a few of the fins. I decided to take a look at what else Fusion had to offer for finishing passes and see what would work better. I found that the morphed spiral toolpath adapted nicely to its sharp contours and set it to take smaller bites; this worked amazingly well. I was able to even get into the crevises under the blades, leaving a smooth transition between them. I am thrilled with how it turned out, and even with the broken blades, I would have to say it's the best of the three I've made as far as learning goes. 

You can really see the difference, on the first image, on the underside of the left blade is the parallel and the underside of the blade on the right is with the morphed spiral toolpath.

My second attempt at making the Impeller wasn't as successful as the first. 

I wanted to cut back on machining time and reduce the file size of the g-code, so I changed some of the settings. I raised the fine stepdown to 0.025" from the previous 0.008". I also increased the overall size of the Impeller to prevent parts from getting too small. I did keep the finishing passes the same as they were in the previous version.

I ran into an issue pretty quick, the stock I was using was a bit larger than the previous time I machined the Impeller, and my tool reach wasn't quite long enough to machine down to the center of the part. I was able to fix this by using a few facing toolpaths on the corners of the stock, which gave me another 1/8" reach. 

I was having some issues with the post-process trying to obtain the g-code, it would fail as soon as I tried to export it. I can't for the life of me figure out why. I played around with which toolpaths I exported and found that certain ones don't work together in the same file, I'm very lost as to why it's doing this. The workaround I found was to select as many toolpaths at once and do them one at a time, time-consuming, but it works.

I ran through five of the seven roughing operations and must have accidentally set the sixth toolpath to have no stock to leave, which caused one of the fins to snap and left too small of a peg to stand on, and so I couldn't do any of the finishing passes. Since the part was pretty much useless at this point, I decided to try out one of the five-axis toolpaths. Namely, the swarf tool, I programmed it to finish the underside of the part and cut it from the base. The underside of the Impeller had a very slight dome-like incline, and the swarf worked out very well (see the last image).