Looking forward to the planes you'll make using this lovely machine!

We missed your airplane building and flying videos! CNC process is very interesting and I was hoping for some flying footage too, bingo! I think you can CNC an FT style larger wing like the Kraken maybe? That'd be a sweet sweet build.

Me too we missed your airplane building techniques it was fantastic really

Everything is looking pretty good, I'm excited to see what you make with the improved manufacturing capability!

@markgreco1962

2 жыл бұрын

Right on. !

If anyone was going to make a CNC Hot Wire Cutter, it'd be you!

The inductive switch: the problem with the screw is that it is round. The sensor is only specified to reproduce the stop position, if half of the sensor is covered with metal, a half moon trigger so to say. The thing you sense have to be flat and in parallel with the sensor, to measure correctly. Interesting video as always 👍 Thanks for sharing your experience with all of us 👍😀

@avejst

2 жыл бұрын

Another thing: please use shielding cable for the spindle. And connect the shielding part to ground at the converter end. DON'T connect the shield at the spindle end!

@iforce2d

2 жыл бұрын

Yes, the cable is shielded and connected exactly like you say.

Nice :) For a drag chain ptfe or silicone wires are great based on our 3D printing experience with Voron. NC switches are great for safety but don’t know if your mb rated for 24v signal back that inductive sensors might deliver so bat85 diode won’t be a bad idea to install. Check the Voron build manual if needed. Also just a mouse button endstops are working incredibly well and costs effective for 60c per piece :) Two X or Y endstops could ensure squaring on the axis so nice feature with endstops anyway. Foam machining wing is awesome ;)))

absolutely sick man

The cnc did a decent job on the foam. To save a lot of weight use water based polyurethane instead of resin.

Yeah typical 3:50am yt video😅😅

As a quality check, you could calculate the volume of your raw material and measure the weight to get an idea of the density and if that's different to the one foam with good cutting quality. That way you dont have to buy fully blind. What also could affect the cut is the amount of humidity within the material and you try to pre-dry it in a vacuum bag for a day or so. Thanks for sharing your work! Looking forward to the next video! Cheers from Germany

@iforce2d

2 жыл бұрын

After much comparison I found the best way to check is to squeeze the foam a bit. The one with poor result is a tiny bit softer and bounces back more when you let go. It's only a small difference but after squeezing a whole bunch of samples I have here, I think I could get it right next time I'm at the shop buying foam :)

@Tu4r3g21

2 жыл бұрын

Ok, now my last answer containing some fancy material advice that are probably way over the top (Rohacell IG foam) is just gone. 😂 But one of the major problems you were facing during machining seems to be the brittle trailing edge, or the big pores that prevented the small details to show, got me thinking. Do you think it would be possible to use thinner material with smaller pore/higher density and pre-bond two sheets before machining and place the bond seam right in the middle of the wing profile that you'll end up with a high strength/ free of pores trailing edge? Just an idea.

@PiefacePete46

2 жыл бұрын

@@iforce2d In your previous video, when mistercohaagen saw you fondling the finish, he commented "sanity is a fragile thing ; )" On the strength of that, I don't think it would be a good idea to include video of you "squeezing a whole bunch of samples"! :o) The man at the foam supplier might think you have lost the plot when you want to squeeze his produce too.

Just a rule of thumb is no less than 32 strands per wire and no less than 6 diameters for bends. Extension cords are pretty good for the money. Silicon covered wire is very good.

you can cnc a negitive and put it under the wing edge when you do the reverse side to stop the flex

@coreviz

2 жыл бұрын

Or make a wooden mold and filling it with insulating foam and starting mass production, hehe.

In my experience with machining that foam, it has a grain and a top and bottom. That greatly affects the density, rigidity and quality of cut.

@PiefacePete46

2 жыл бұрын

Bruce Baxter: Let's face it... it is intended for building insulation, the manufacturers never expected anyone to give a toss about consistent density etc.

3d printing a clamp on fan and duct to suit the spindle and motor works well as cooling and evacuation of swarf.

@PiefacePete46

2 жыл бұрын

Bruce Baxter: 3d what? (I don't think Chris has a 3d printer) Substitute CNC for 3d print. :o) Having said that, I did see a simple 3d printed fan, about 55mm diameter with a circular rim that was just a push-fit onto the shaft of the cutter. It did a great job of blowing swarf away from the cutting area, at minimal expense. It looked a little like the kids toy flying disc that you launched with a wound-up helical spring.

@brucebaxter6923

2 жыл бұрын

@John Colvin Yep. It’s very similar. The fan sucks instead of blows and there is a shroud that blows it past the motor for increased cooling.

I’m building a Long-Ez in my garage (full size). We have similar problems with foam trailing edges but the designer came up with a genius workaround. I can send you an infographic explaining the process if you like. After glassing, the TE comes out razor sharp.

@jotham123

2 жыл бұрын

Am also interested, any way we can find out too?

@LSUtiger607

2 жыл бұрын

@@jotham123 See my additional comment above

@hlcdriver

2 жыл бұрын

You can just see the workaround in this picture: eurekacnc.com/images/IMG_6159001.jpg The first lay up would be the bottom surface & around the leading edge, peel plying over the leading edge. Once that's cured, cut off the excess foam at the trailing edge, sand it smooth, fill in any gaps with flox & lay up the top surface. The slight joggle in the underside trailing edge would be filled when the time comes to finish the surface.

For moving applications i recommend the chainflex cable from Lapp

Judging by the variations in the surface finish, there is a rougher finish when conventional milling, which is what would be expected. Climb milling usually results in better finish, and is the normal mode for CNC operations. So reducing the step-over and keeping cuts near the finished depth to climb cuts should help. Try some experiments to determine, I am just some doofus from the internet. :D

You say about the hot wire cutter leaving a more solid finish... How about a hot end for the CNC? Kinda of like a 3d printer hot to melt through the foam instead of milling

Next project, foot release bungee

Re the ratty trailing edges at 21:20 and 27:08... there is still noticeable flexing between the support tabs. Instead, butt the trailing edge into a single full length tab, the flex would be virtually eliminated. After cleanup and sanding, one knife cut would give you a nice, straight, clean T/E. You wouldn't even have the remains of the tabs to fix up the way you are doing now. Not sure if it would work on the leading edge...

I'd be curious if the room temperature is at play, maybe the tool is producing just enough heat for the foam to ball up when getting cut, and a cooler day was enough to keep it below that point My next idea would be that maybe the thicker foam has larger cell size? But that is generally manufactured by slicing off from massive blocks, so yeah it might just be random chance on if you get a good part of the block or not

@iforce2d

2 жыл бұрын

I think it's more likely to just be differences between batches. Looking at the other blocks I have, they are slightly different colors. The sides of the slabs where they were saw cut seem the easiest way to distinguish them, the one that machined poorly is a little more 'hairy' looking.

@Frazer.McLaren

2 жыл бұрын

@@iforce2d Sometimes extruded foams can have a grain to them and will machine differently in x and y. Next time you do a test note which way it came off the sheet.

there is a vid on covering foam planes with Kevlar and crack filler that could help seal fill and strengthen

@PiefacePete46

2 жыл бұрын

CROG: There are videos on this channel covering similar processes. ;o)

It could be the quality of the foam that you are using, it might not be as dense as it should be for it to do a nice clean cut if it's a looser density you'll get all kinds of tear out.

Really love your videos. How did you generate the tool path? What software did you use?

@iforce2d

2 жыл бұрын

Fusion 360, more info in description

The finish on those cores is excellent! How does hot iron laminate or packing tape react to that kind of surface? I reckon it could help with weight savings on the next build.

@iforce2d

2 жыл бұрын

I think laminate would stick well, just gotta be careful because the heat has nowhere to go when the iron is pressed against such good thermal insulation :) Packing tape sticks ok but for best results would need a little sanding I think, the surface is a little too porous as is.

I wonder if you tried a heat gun or blowtorch to warm up the surface you just machined would it give you a better surface (more like hot wire cutting) to then fiber glass on ?

i dunno about never flying it....... it lokos like it would be the perfect sloper, as for the trailing edges, i'd not be worried about it cus i'd be slicing them off for balsa elevons anyway :D

5axis wire cutter :)

Thanks for the video mainly for the explanation and closeup of your limit switches. I have the same switches but have struggled for weeks trying to work out the connection to the bob. I am powering the switches with 12v and have 1.5k resistor to bob input but when I benchtest output side of resistor I get 12v, not 5. I have wiring diagram, just missing brainpower (I don't have much electrical knowledge although I've built the cnc and wired everything except these sensors, might have to use microswitch out of desperation. Would you care to do an in-depth video with closeups of your setup, I get it if you are too busy. cheers.

@iforce2d

2 жыл бұрын

My bob doesn't need 5v inputs so fortunately I didn't have to deal with resistors. If you're checking the voltage on the output side of the resistor without it being connected to anything (other than 12v on the other side) then naturally it will be 12v. Don't you need two resistors to make a voltage divider? Regarding closeups there was one at 0:41 but I guess without any labels. Brown is 24v, blue is negative, yellow is signal. The two cables on the left go to the sensors. They are NPN normally closed. tbh I don't understand fully how it works but for these the key to success for the series connection was realizing that the signal of the first sensor should be connected to the ground of the other one, not the positive.

@dtanco

2 жыл бұрын

@@iforce2d thanks for replying, my wiring diagram shows only one resistor and gives the (resistor) rating for different supply voltages but I'm still missing something. I'll keep scratching 🤣

I guess you need higher RPM on the spindle for this material and these feed rates.

I wonder if the round screw head is also causing a problem with detection.

@PiefacePete46

2 жыл бұрын

Do RC: I wondered that... add some nuts to the screw, or use it to hold a piece of steel maybe.

Do you have you final design on your build that can be shared Great videos. Shout out from Oxford Ms

@iforce2d

Жыл бұрын

You can find my final FreeCAD file here, not sure how useful it would actually be though: www.iforce2d.net/tmp/cnc2-noAngle.FCStd

I don't quite understand the thinking behind wiring the limit switches in series, doesn't that mean if you trip the first one they both turn off, and therefore the machine can't distinguish between just the first sensor tripping and both tripping at the same time? I suppose it wouldn't even matter if you're only using them to prevent crashes and not home the machine 🤷 and the more I think about it, if that's what the machine is expecting it could easily cope for homing too

@iforce2d

2 жыл бұрын

Well the assumption is that tripping switches both ends at the same time is not possible. The homing procedure starts with neither triggered, and the machine knows which direction it is moving, so there is no ambiguity about which switch was tripped.

in my oppinion the foam height is adding toppling effect(vibrations) which are causing the rough finish, or I can be anything😂😂

Does the foam have a grain to it? That is it will cut better one way than the other.

@iforce2d

2 жыл бұрын

No grain, that's one thing that makes it nice to work with.

Maybe your depth of cut is what's affecting the tear out

@iforce2d

2 жыл бұрын

All settings were the same, and same tools.

@DoRC

2 жыл бұрын

@@iforce2d was the thickness of the stock that you were going through the same?

So, now you can make bug curved surfaces……how about an Avro Vulcan?

I think you may want to check the tram of your spindle, i don't think its climb vs conventional that is showing the difference, i think it may be the angle of the nod or tilt.

@bgdwiepp

2 жыл бұрын

Maybe also for your spindle cable, sheath it in a bit of black pvc garden/sprinkler pipe to slightly stiffen it a bit so it will maintain the larger bend radius; i noticed when it gets all the way to the opposite side of the bed it begins to bend a bit tightly at the top of the spindle, and when it is on the cable side of the bed the bend is a bit tight at the top of the loop.

@bgdwiepp

2 жыл бұрын

You could also try do your glassing in 2 operations, your first op, before you sand add as much fumed silica to your resin as you can, it will fill all of the holes and should seal the surface while being very light. Then sand it (the fumed silica and resin should sand very well) and apply the glass.+ resin over that. Fumed silica is quite cheap so it might be worth trying.

try to check the density of your foam. could be that it is humid and that's what causing the bad cnc cuts. If it's humid the foam will be denser, you'll need a precision scale though.

@iforce2d

2 жыл бұрын

This is closed cell foam made for thermal insulation of buildings, where it is supposed to sit for decades without taking on moisture. That doesn't mean the interior will stay perfectly dry forever, but from what I've seen you need to have it submerged in water for a long long time before it absorbs anything.

@james-xh6wr

2 жыл бұрын

@@iforce2d I agree with you that the foam didn't absorb that much moisture, but what else could be the cause of two identical sheets cutting differently on a cnc? I'm not sure if moisture would really affect how foam is cut, but I know it affects plastics in the 3d printing industry, and that's why I am suggesting that moisture might affect material on the cnc.

@iforce2d

2 жыл бұрын

After doing a lot more machining of this stuff, I'm pretty sure it just depends on the batch of foam as it was made in the factory. Some boards are good on one side and bad on the other, and it can even change quite suddenly, like within a few inches. The thinner boards tend to be better.

Is there anything like a grain direction, in extruded foam products?

@iforce2d

2 жыл бұрын

Not that I can see. After checking all the samples I have, it seems that some are just slightly more rubbery than others. Softer when squeezed, and they bounce back more.

Hi sir how to set a speed for spindle to fast because if I start cutting the bits is broken, because the speed not fast

@iforce2d

2 жыл бұрын

There are a number of online calculators that might help, eg. www.cnccookbook.com/cnc-feed-rate-calculator

Where is that farm?

Is that cable aluminium core rather than copper? Chinese fake?

@iforce2d

2 жыл бұрын

Listing says copper but it sure looks like aluminium to me. www.aliexpress.com/item/1005002279602656.html?spm=a2g0o.order_list.0.0.21ef1802IXpKbR

@alexscarbro796

2 жыл бұрын

@@iforce2d interesting! Try scratching it to see if it reveals copper. Otherwise, try tinning it with solder. If the solder doesn’t take, it’s likely Aluminium.

what airfoil did you use on this wing?

@iforce2d

2 жыл бұрын

eppler 335

you physically swap out drivers? can't just put both drives in the computer and chose the boot drive in the bios?

@iforce2d

2 жыл бұрын

I could if there were any connections left. I have other drives on the other connections so that all my data is always available, regardless of which OS is running. But that would only be a slight improvement. The main annoyance is that I have to shut down everything I'm doing for work and then can't get back to it again until I revert to Linux.

@fuzzy1dk

2 жыл бұрын

@@iforce2d yeh, I know the pain. On my work computer I've basically flipped the problem around. Run win10 and doing all the linux stuff in WSL, it works surprisingly well

@iforce2d

2 жыл бұрын

I came across WSL the other day when I wanted to share the gcode files with the router computer over wifi, and it seemed like the easiest way to get a web server running. Before that I tried samba on the router computer which wasn't possible because Windows10 rejected the outdated protocol, then tried a file share on the Win10 which failed for the same reason. Wasted about 2 hours and then gave up and used a USB flash disk.... over and over.

@fuzzy1dk

2 жыл бұрын

@@iforce2d for moving data to a linux computer from windows I usually use winscp

@iforce2d

2 жыл бұрын

Aaaaand..... it looks like I will be using WinSCP too. I just spent an infuriating two hours trying to tweak a job to fit just inside my machines work area, before realizing that the files placed on WSL apache were not changing when fetched via the browser...! Nothing to do with browser caching, they simply just never get updated if edited when the file name is the same. FU#@##@#$%&&^%#%$!!!!!! superuser.com/questions/1240107/wsl-linux-files-dont-update-when-windows-files-are-modified There are also many other annoyances like having to set permissions to 755 every time a new file is placed in the www tree for it to be viewable (otherwise 403) and super weird behavior in the 'linux' shell where copying a file simply does nothing (and no errors shown) if a file of that name exists already. tbh, this makes me never want to use WSL for anything ever again.

Ahrhghhhhhh f360 the intolerable crap. Cloud cad with no Internet is not much use either! Freecad 0.20 with path is the only way forward, 3d surfacing I think is still experimental but everything else including 3d pocket is functional.

@iforce2d

2 жыл бұрын

Yeah I should have mentioned that I like FreeCAD's CAM result in general, and will be using that sometimes too. Major problem is that the 3d contour paths like I was using here take a painfully long time to calculate. Like for example, 3-8 minutes vs a few seconds for F360. I could just about turn the computer off, swap my hard drive and reboot to run F360 and get the result quicker. The machining simulation in FreeCAD is also way slower.

Are you using a ball nose cutter for all that roughing? That could explain you finish issues.

@iforce2d

2 жыл бұрын

Roughing is done with the 1 inch planing bit. But same tools and settings were used in both cases, so shouldn't be relevant.

@Reach41

2 жыл бұрын

@@iforce2d Maybe not, but with a normal mill cutter turning at some rpm and fed at some rate, the SFM is a constant, as the OD of the cutter is doing the shaving. With a ball nose cutter, the SFM gets less toward the tip of it. Not a very good explanation, but think about it. That’s what jumped out at me, that you could be crushing your way through material.