ERC TimSav - Cheap DIY CNC Foamboard Cutter

[Dudley1001]

Hey all, I’m interested in building one of these. Is this still a good choice, or are there others I should consider? I’m particularly concerned about needing to run the old .91 version of InkScape. Any thoughts?

It depends on what you you want to build and how many. If its a one off then its probably just as easy to cut by hand. As for inkscape, you can use the newer versions of inkscape if you are processing the file yourself. The scaling problem comes in if you are importing svg files that have been made with inkscape .91. There are some new addons such as "Clean Plans" that make processing the files much easier. Be aware that the newer Flitetest plans cannot be converted without redrawing the outlines but the older plans still work. Are there better machines, of course, but at what cost. The compact nature of the TimSav makes storage easy for those without a lot of room. Mine is setup on a table in my garage. I have also changed out the cutter head so it can be used as a laser engraver/cutter. As for making the gcode, I use a different program called Laserweb4 and a different firmware on the Uno called GRBL 1.1h-servo. This allows me to use the laser and needle cutter.

 

[Foamforce]

It depends on what you you want to build and how many.

I’ve built something like 15+ now and the novelty of carefully cutting each part is wearing off. I see that there is a repository of “Classic” FT plans that are already prepared, so that would keep me busy for awhile. Plus, I’m getting into combat and I’d like to invest less time in each crash. 🙂

A couple parts questions. Valuehobby has 2040 extrusion but not 2020. I was thinking about modifying the X axis (I would think of it as the Y axis, but I think he calls it X in the videos) to use 2040 instead of 2020. Other than having to modify a couple parts to fit, can you think of any problems with doing that? Would the stepper motor need to be bigger?

Speaking of the steppers, how hard are they worked on this machine? Value hobby has one that is just marginally smaller, like .36Nm instead of .4. Think that would be ok? Sourcing the parts seems like it will be the most difficult part since the parts list is out of date.

Thanks!

 

[Dudley1001]

You should be able to get the rails from openbuilds as well as other parts. As for the motors, the ones you list should be fine. The problem with bigger motors is they are too long for the z axis as they will drag on the table. I think the max size of the motor height is 40mm if I remember correctly. I think I tried 42mm and they would not work. I was a copier tech and had a friend who ran a recycle center so harvested motors from a copier.

 

[Foamforce]

You should be able to get the rails from openbuilds as well as other parts. As for the motors, the ones you list should be fine. The problem with bigger motors is they are too long for the z axis as they will drag on the table. I think the max size of the motor height is 40mm if I remember correctly. I think I tried 42mm and they would not work. I was a copier tech and had a friend who ran a recycle center so harvested motors from a copier.

Hey, thanks for the tip about OpenBuild! The rails were hard to source for a reasonable price and they’re much better. I think I’ll start collecting parts soon.

 

[Foamforce]

It’s happening!! My build is getting close to complete. Just waiting on some new wheel bearings because the cheap ones I got were all stuck.

I sure wish Edward still made the lite kit! Gathering all the parts, particularly the screws and washers, was painful. None of my local hardware stores had individual m3 pieces, so I had to buy them in all these little baggies. It cost $38 just for the screws and washers. Also, I couldn’t find washers small enough for the bearings, so I had to grind down the outer diameter of about 30 washers. Even having a lite lite kit with just the screws and washers would make this a lot easier!

But overall the build has gone well. The printed parts all fit nicely and the software tutorials have been super helpful. Hopefully I can try my first cuts in a day or two. This is exciting! Thanks Edward!

 

[Foamforce]

Quick question. My X axis carriage is a little loose. The four wheels aren’t complete snug against the 2040 beam. My Y axis is just perfect, so I don’t think it was a print scaling issue or an issue with my wheels being the wrong size.

Is this a concern? Maybe the tension of the drive belt keeps it tight against one side of the rail? If it’s a concern, how have others dealt with it? Perhaps by drilling out two of the holes slightly larger and the gluing a bushing on top, slightly toward the center?

 

[Foamforce]

D’oh! I faced a setback. When I plugged it into the 12v power supply for the first time, my laptop, which was connected via USB (per instructions) shut down and wouldn’t turn back on. After trying everything else, I took the back panel off, disconnected the (non-removable) battery, and also disconnected the bios battery. Miraculously, it turned back on after that. I’m wondering if it was a coincidence and maybe a thermal fuse finally reset, but it had already been almost a half hour, so I’m guessing it was the batteries.

So now the Arduino, which I had previously successfully flashed GRBL to, won’t respond, I assume it’s fried. The On, rx, and tx lights also light up, but don’t respond to anything, even though the computer still recognizes the USB device.

The most obvious suspicion is that my soldering from the vin pin to the power positive pin crossed the 5v circuitry somewhere, but I can’t find it. I checked continuity between the vin pin and the 5v pins on the shield and didn’t find any. On the Aruino side, when I plug in the 12v, I read 6v coming off the 5v pin, and I also read 6v coming out of the USB port. So that definitely seems like a blown voltage regulator at least, and could maybe explain why my computer went haywire.

So I’m fairly sure I need a new Arduino (this was a cheap knock off one, btw), but I don’t know about the CNC shield. I don’t see any immediate evidence that it’s not working, but it was apparently getting 6v.

I think what I’ll do is try another Uno and hook up the shield to it, then connect the 12v, and test the 5v output to see if it’s 5v. I’ll also test the USB to see if anything is going there. Any suggestions?

 

[Foamforce]

Success today! I borrowed another Uno from a friend and I used two new stepper drivers, since my kit came with four. I used the same CNC shield. I think what happened was when I was connecting the wiring, I used the trick of moving the X and Y axis to see when the lights lit up. I read somewhere else that this can send too much voltage back through if you move it too fast, burning up the stepper motor and the Arduino. Either that, or the 5v regulator was defective on the Arduino, and the first time I plugged it into 12v it fried everything.

So it’s working now! It’s cutting roughly, but I have a few things I’m going to try. I need to tighten up the Y axis wheels first. I think I’ll just reprint the top plate and place the holes 1mm close together. I think I’ll also slightly shorten the needle and sharpen it more. Lastly, the scaling seems to be off a fair amount, so I need to redo that, alright oddly it seemed fine with the 20mm test. I’m thrilled that it’s more or less working now though. 😀

 

[Foamforce]

So I had it working fairly well for simple sheets, but I would still occasionally have the problem where it would stop dead in its tracks with no error message. In the GRBL Controller, it will just sit on the last command sent. Hitting stop won't respond and typically I have to disconnect and reconnect the USB to get GRBL to respond again. Of note, this seems to happen only after it's been running for awhile. Specifically, I've been testing with sheet 1 of the Mini Guinea which has quite a few cuts. It seems to almost always stop somewhere between 75-100% complete, but not on the same instruction. It sounds most like EMI problems based on what I'm seeing, but it seems like EMI problems usually throw some sort of error, which I don't see. However, it seems very significant that it usually happens late in the process.

Various things that I've tried:

• I've tried a few different USB cables. The shorter ones <might> be slightly more reliable, but I still only have about a 40% success rate of finishing a sheet.
• I put a fan on the shield and Arduino, thinking it might be overheating.
• I tried setting the vref. Fwiw, I put it at .9v, although the calculation says I should be at 1.2v. Could being too low cause errors?
• I put electrical tape over the top of the USB port on the Arduino because it seemed to be touching some contacts on the bottom of the shield.
• I moved to a different electrical circuit in my house, further from the furnace.
• I replaced both stepper drivers.
• I loosened one of the toothed belts, thinking that I might have been overloading the steppers.
• I tried moving my USB cable so that it wasn't touching any of the power cables, again trying to reduce EMI.
• I tried removing the jumper wire and powering the shield directly with the 12v, also moving the power for the servo to the ESC.
• I tried a different 12v power supply.

Tomorrow I have a brand name Arduino coming. It's worth $28 to eliminate the cheap Elegoo clone as the potential issue. Regarding happening late in the process, I wonder if there's some bad internal memory on the Arduino that tends to get hit only after a certain number of instructions have been run. If that doesn't work, I'll try a new CNC shield.

Does anybody have any other things I could try? This is a vexing issue!

 

[luvmy40]

I have this happen once in a while. Not often. May once every 10 or 15 sheets. On at least 1 occasion the problem was a corrupted GCode file. However, it was stopping at exactly the same point every time.

I simple reboot everything and start over when it does happen. With that one exeption, this works. Frustrating, but not the end of the world at $1.25 per sheet.

 

[Foamforce]

Yeah, that would be ok but my problem is more severe. It’s failing about 60% of the time.

My new brand name Arduino is out for delivery. I hope that cures this.

 

[luvmy40]

Check and make sure the boards are tight to each other. I actually used some packing tape to keep them from migrating.

 

[Foamforce]

Check and make sure the boards are tight to each other. I actually used some packing tape to keep them from migrating.

Thanks, I tried that. Adding to my list I’ve now tried replacing the CNC shield and the Arduino itself, which is now a genuine Arduino. Same problem.

Would you be willing to try my gcode file and see if it will complete two runs in a row on your machine? I’m just testing with the cutting motor turned off, just moving the steppers, so I’m not wasting sheets. I would have gone through 20+ sheets in the past couple days otherwise!

 

[Foamforce]

Well, a different sheet had the same issue, so it’s not necessarily a gcode issue.

I have two things I’m trying next…

1. I read in a big report that GRBL .9i had issues where if you ran it too fast, it could overflow a buffer. So I’m taking the travel speed from 3000 down to 1000. I’m honestly not sure what they meant by going too fast, but it’s worth a shot.
2. Somebody on the Facebook group experienced this problem and they said it was solved with a different power supply. I did try with two before, but the second one was even cheaper than the first one. I’m going to try the 12v output from my Hitec charger, which I think is pretty good quality.

Hope.

 

[luvmy40]

Yeah, that seems a bit fast to me. I cut at 400 and travel at 1000.
BTW, I'm using the multi cut extension shared by a forum member earlier in this thread in Inkscape to create the GCode.

 

[Flying Farm Kid]

Yeah, that seems a bit fast to me. I cut at 400 and travel at 1000.
BTW, I'm using the multi cut extension shared by a forum member earlier in this thread in Inkscape to create the GCode.

Are there still kits being sold or do you have to source the parts yourself.
I am looking into building one for the same reason and not having to buy all the parts separately would be a great help

 

[Foamforce]

Are there still kits being sold or do you have to source the parts yourself.
I am looking into building one for the same reason and not having to buy all the parts separately would be a great help

You have to source the parts. There was a guy in the Facebook group recently who was selling a kit that he never used for $150. I would jump on that if it’s still available.

Yeah, that seems a bit fast to me. I cut at 400 and travel at 1000.
BTW, I'm using the multi cut extension shared by a forum member earlier in this thread in Inkscape to create the GCode.

I don’t want to get my hopes up too much, but I’ve had three successful cuts in a row with the lower speed, but that could just be luck. I’m going to try to replicate that again tonight. If that works, I’m going to test at the higher speed with the bigger power supply. I could see how going faster might be causing a brownout.

 

[Foamforce]

Victory!

Both of my last tests worked.

1. I slowed the travel speed between cuts from 3000mm/minute down to 1000mm/minute. I ran five runs with no failures. Previously I could get a maximum of one successful run before a failure. Each run took 27 minutes.
2. I returned the speed back to 3000mm/minute but switched to a much better power supply. I used the 12v output (actually 13.8) from my Hitec X2 charger, which is capable of 100w. I ran three runs in a row with no failures. Each run took 23 minutes, an improvement of only four minutes, probably because I optimized my gcode files pretty well.

So my hair-pulling issues seem to have been entirely due to a cheapo power supply. I think reducing the speed reduced the load and allowed it to succeed, but just getting a good power supply was the real trick.

Thanks for the help, and I hope somebody else sees this and saves their own hair in the future!

 

[HMS]

Hala satılan kitler var mı yoksa parçaları kendiniz mi tedarik etmeniz gerekiyor?
Aynı nedenden dolayı bir tane inşa etmeyi düşünüyorum ve tüm parçaları ayrı ayrı satın almak zorunda kalmamak çok yardımcı olacaktır.

3D yazıcınız var mı, parçaları yazdırabilir misiniz?

 

[Dudley1001]

One thing to watch out for is don't power the machine through the black power connector on the uno. You should only power it through the blue connector on the cnc shield. If you power though the uno, you will probably burn up a diode that is only rated at 1 amp. I know that Edward said to power through the black connector but don't do it, you can let the magic smoke out. I also power the servo from the extra pins on the servo tester which gets its power from the esc thus reducing the power draw from the uno's 5v regulator which can cause some brownout resets of the uno