Hello everybody!
I need your help fighiting a strange z-axis behavior on my MPCNC converted to an Embroidery machine.
I just get the Z-axis motor and mounted it on an old sewing machine. I designed and 3d printed a hoop which is mounted on the former MPCNC Z-axis (which is currently without a step motor and just holds the hoop). Here is how my setup looks like:
I set Z-axis steps per mm to 8960 (DEFAULT_AXIS_STEPS_PER_UNIT) which gives me full rotation in 5mm. And I set Z-axis max speed to 3.0 (DEFAULT_MAX_FEEDRATE). All the configuration is made in the firmware Marlin-MPCNC_Ramps_T8_16T_LCD_32step which I uploaded to the machine.
I was pretty happy with my embroidery machine and here result I got from it:
But the second time I ran embroidery, I heard terrible noise from the Z-axis motor. Here is the video: https://youtu.be/hkFr5mrA-6Q
When I try to move Z-axis manually – everything is fine. I set Z=10 via LCD Menu and get 2 full rotation of the wheel. But when I run g-code Z-axis doesn’t move (X and Y are ok) and instead of the move I hear that noise.
@vicious1, looks like a new category in the gallery.
So you move to the end and then cycle it a few times to secure it, then you go to the next endpoint? I guess this is pretty manual CAM?
Let’s look at the Z speeds:
8960 steps/mm * 3mm/s = 27,000 steps/s. That’s a bit high.
On a standard MPCNC Z, we have 400 steps/mm * 10mm/s = 4000 steps/s. That is conservative, but it is a good point of reference.
If you were trying to stay close to 4000 steps/mm, you’d need closer to 0.5mm/s max speed.
If you are using 32nd microstepping, then you can double that (but don’t change it). If you use a 24V supply then you can probably get double that speed, maybe more.
Looking at the video though, it doesn’t sound to me like it is skipping steps, unless it is just hard to hear.
check the grub screws on the motor shaft
slow it down
consider a 24V power supply
consider changing the ratios for fewer motor shaft turns, since steppers are stronger at lower turns.
It would be great to get this to work with a nema 17 (and it looks like you’re close). But this might be a case where a beefy driver and a nema 23 would handle this very easily, and since it isn’t moving, the weight isn’t an issue.
This is really neat though. My mind is kinda blown.
So you move to the end and then cycle it a few times to secure it, then you go to the next endpoint? I guess this is pretty manual CAM?
It is pretty automatic. The reason I started this project is that I found the Inkstitch which has G-CODE output option. And it is pretty advanced and supports many settings like max stitch len and satin columns generation. But generated G-CODE is pretty simple, yes: several rotation for initial thread securing and after that sequence of X,Y movement with Z 5mm increase at a stitch end.
I took an old industrial machine (pre WW II, 22 class). It is pretty hardcore from modern perspective, because it can sew through several layers of leather, or denim. And it is LARGE, so I can use really big hoops. But it brought a lot of troubles for me, because I don’t know how to sew and I faced many problems like thread breaking in a middle of embroidery.
I plan to throw all 3D models I used to thingiverse along with description of needed G-Code modifications (not much: M92,M211,M203 and that’s it). I was running everything on a pretty old version of firmware and there were problems as well (I couldn’t disable virtual endstops and Z-axis could not run beyond 200mm, for example).
But now this thing is up and running.
But it is SLOW. Small embroidery takes 1 hour. I am thinking about making the wheel smaller. But actually for speed I would need to use DC motor to spin the z-axis without stops, and add optical sensor to make XY moves while the needle is up. And it is overcomplication for me. So I guess I will stick to use it as is.
A solution like this is really valuable. It won’t be the fastest, but you probably aren’t making thousands of polos into thousands of polos with logos. Instead, it opens the door to a technique you didn’t have before. Hopefully it will open the door for others as well. The MPCNC has done this for a lot of us, and a few have managed to turn it into a business that was profitable enough to justify the larger investment in a production machine. This could do the same for someone. Someone who wouldn’t be able to justify the cost of a large machine until they had the customers, and couldn’t get the customers without a machine can now open that door.
So, can you make patches like in the inkstick gallery? How does that work? You just switch the colors a few times and draw the pattern, then apply some kind of adhesive to make it rigid, and then the machine can do the border and cut it out all at once?
Watching the video again. It is so cool. Does it seem like it has plenty of torque? Maybe a larger pulley on the motor would help with the speed. Or a larger motor could speed it up. Still, I think it’s awesome as it is. 1 CNC embroidery machine is far better than zero CNC embroidery machines.
My daughter and I were just talking about this a couple of days ago… and now Konstantin shows up here with another really interesting take on it. I’ve been messing around with V-slot extrusions so much lately, I have most of the stuff on hand to build the Openbuilds version… they’re quite simple hardware-wise. The software is what has me bum-fuzzled most ATM… so I’m really interested to see how Konstantin is handling that part of it…
So, can you make patches like in the inkstick gallery?
I’ve just made it running with your help. So I can’t do much currently.
My next step is to switch to finer needles, to put less damage to fabric. So theoretically I can do everything (with switching colors manually). And I know that there are similar projects with pretty good results.
And here is my current result:
You can see small holes around the embroidery. It is because my needle is way to thick and I forgot to use paper as underlaying layer. Will keep working on quality and will keep you posted.
You just switch the colors a few times and draw the pattern
Right
then apply some kind of adhesive to make it rigid
There is no need to make the patch rigid really. But it is pretty rigid because of embroidery itself adds rigidity and usually there is a layer of paper-like material to avoid fabric moving because of the threads strain. But yes, some glue is applied to the back side just to prevent embroidery to fall apart after cutting.
Patch can be either glued in place after cutting or iron-on fabric can be applied:
and then the machine can do the border and cut it out all at once?
You can just cut out the finished patch with scissors. Glue will hold back of the patch.
Looks like a really cool build! I used to run a multi head embroidery machine in a job long ago. My guess is that you could speed it up substantially by figuring out how to time the XY movements with the needle oscillations.
I had a large embroidery machine for a while, and this brings back some memories! I bet when you play with the speed a bit you can get this pretty dang fast. Even at the current speed you have really impressed the heck out of me, great work!
I don’t know how small is “small” to you, but when I worked for a shop with a professional 2-head embroidery machine, it was generally pretty slow, too.
Like the others, I’m amazed at this machine. Seeing the picture made me instantly want to make one myself, but I really don’t have the time, money, or room for it.
From the limited reading I’ve done on the subject, the challenge seems to be getting some sort of position sensor on the shaft that moves the needle up and down, then being able to tune it for when the needle is really out of the work.
As long as the thickness of the work doesn’t change too radically, I would think you could tune for your thickest possible workpiece rather than having to do any sort of “touch-off” for each job.
Luckily (or unfortunately, depending on your point of view) I haven’t invested much more thought into the mechanical problem since I received my mom’s old Viking/Husqvarna sewing/embroidery machine as a hand-me down. Now, when it gets to the top of my to-do list (which will be a while) I’ll be trying to sort out the digital storage card interface so I can look at playing with Inkstitch and getting the designs into the machine. Unfortunately, it predates USB and I haven’t been able to identify a physical standard for the storage cards.
I’ll be following this thread with interest. Congrats on the progress you’ve made so far.
This solution is clever because it uses a stepper to do only whole cycles. You just need to set the Z=0 at the start where it is clear.
If you wanted to speed it up, you could try a larger pulley on the motor. I have no idea how much torque it requires to move that beast of a needle through the fabric. Or you could increase the Z speed and swap out for a 24V PSU.
Hi everybody!
Thank you very much for your support and advice.
Current situation: the CNC part works without any problem, but I am struggling with the sewing machine itself.
I had a couple of really good runs with thinner needle and thinner threads, but now the machine tears threads, or misses stitches, or just generally misbehave:
I have been fightning these problems this whole winter and now they back. I am even thinking about buying some post-WW2 machine which hopefully will be more reliable.
Tuning the ancient sewing machine is a long story, so I am afraid there won’t be any updates for a while.
I worked a a small training center where the owner was a tech toy junkie, and we ran our own embroidered “swag” doing shirts and hats. Because the company logo had open spaces that made picking out standard stabilizers a PITA, we used a water soluble stabilizer that went over the fabric in the hoop, and the the embroidery got stitched over it. The stabilizer dissolved when the finished product was washed the first time and then sort of glued the embroidery threads in place. We did a couple of patches (for personal projects) and just embroidered the design onto iron-on repair patches you’d use for fixing holes in your jeans. We never needed any additional fixatives for the embroidery.
Dont suppose you do requests? My bro’s wife had a baby today. I want to get a babygro for them which has written on it. “My daddy is actuary, very cool”
A solution like this is incredibly beneficial. It won’t be the fastest, but you presumably aren’t making thousands of polos into thousands of polos with logos. Instead, it opens the door to a method you didn’t have before. Hopefully it will open the door for others as well. The MPCNC has done this for a number of us, and a few have managed to convert it into a business that was lucrative enough to warrant the bigger investment in a production machine. This might do the same for someone. Someone who wouldn’t be able to justify the expense of a huge machine until they had the customers, and couldn’t obtain the customers without a machine may now open that door.
So, can you make patches as in the inkstick gallery? How does that work? You just flip the colors a few times and create the design, then put some type of glue to make it hard, and then the machine can do the border and cut it out all at once?
We had a sticky stabliizer in the hoop and just stuck iron-on jeans patches on to it and embroidered on those, but this was back in 1999-2000. These days you can easily get heat-fusible interfacing in sheets and turn any base fabric into an iron-on patch.
I don’t think you need glue over the top of the embroidery if you use a “dissolving” stabilizer on top of the material (under the stitching). As I understand things, the dissolvable stabilizer is a thin sheet of water soluble glue, so it does that when the item is first washed, then dryer heat “sets” the glue. This is different than the “tear away” stabilizers that are used under the fabric in the hoop. I’ve got a shirt we did from that era, that’s been worn and washed a lot (although not so much recently) and the embroidery looks almost new.
. Amazing!
