An enclosure might help, especially with keeping the bed temp up
Building an enclosure is a possibility, but it will be hard, closing the sides will be no big deal, but closing the top will be complicated given the table I’ve built… Maybe using some kind of transparent tarpaulin on top, would look crappy but it is a simple, cheap and easily removable solution, I’ll see if this is necessary.
Another option I like is to print a wall all around the part, Cura offers this function. A bit more wasted plastic but it might do the trick.
Now what I really need to find is some 5KG plastic reels. For some reason they are very hard to find here, especially at affordable prices. One reel of 5kg is about 1.5 times more expensive than purchasing 5 reels… weird.
A 300mm square silicone heating pad that runs off of mains power is about $30 from Hong Kong to the US. These pads would give you all the heat you want very quickly. 3m high temp tape to the back of an aluminum plate would be ideal, i think a steel plate or even a stone floor tile would transfer the heat well enough. At least you get your filament at a great price it’s 3 to 5 times higher in the US.
My build plate is 800x800mm… So I would need about 9 of them, or to find bigger ones (but they are very expensive, since custom made).
Also, they are not flat (the spot where the wires are coming out is much thicker than the rest of the pad), so it would make leveling complicated.
Add a thick aluminum plate to the cost of 9 pads, it may be well over 200 Usd. I don’t really want to spend that kind of money just for a heating bed, that would be almost half the price of my whole machine!
That was the first thing I thought of, I even bought one 30x30 silicon pad a few month ago. In the end I thought it was too much cost and too much trouble to use it.
Keeping it flat would be a pain in the ass as well. The guys on the seemecnc forums ran into this issue as well. You need a fairly thick chunk of aluminum to keep the differential heating from warping the metal. Even the mic-6 ground aluminum plates would warp. One that I know is working uses three silicone pads to heat it. I think it’s half inch thick and around 18" in diameter.
What about a heat lamp? You could place it under the bed and have it turn on and off the keep the temp of the bed where you need it. Might even be able to get it running from your controll board.
We have something simular at work. But they run off of flukr thermoeters. Which is exensive.
if only i lived in china, have to wait for my parts from aliexpress for 3 weeks to get here to the UK, also that screen mount, im printing that :). Some great idea’s here. keep us updated.
1 Like
Ok, so I finally received my electric tile and tested it on the MPCNC.
[attachment file=40185]
The tile is exactly the same size as my mirror, so 800x800mm. It is pretty much flat, not absolutely perfect, there is a slightly higher spot than the rest, but it is really far better than whatever flatness my table had before, so that is quite an improvement.
The tile is pretty heavy, which is good. The top surface is nice and polished, so there is a good contact with the mirror, which eases the heat transfert.
It takes a while to warm up, about 30 minutes to get to 80% of the max temp, and one hour to reach the maximum. I haven’t made very precise measurment, those are just approximations. I wish it would be faster, but it is really no big deal. The maximum temp I measured was 45 degreeC, but I think it was more around 50C since The sensor I was using was not making a great contact and was partially exposed to the outside air. So, not 60 degree, but close enough.
I made a “small” test, printing the side cover of my electronic compartment, to see if it would warp.
The part is 650mm x 250 mm, 2.4mm height. Took around 3 hours to print… I could have gone way faster but I would need to switch from 1/32 step to 1/16 or 1/8, I was limited by the motor speed. At one point, during a fast travel, the Y motor stalled and I had a hard time realigning everything. I think 1/32 step is not adapted to such large things, it could go way faster as long as the extruder motor is able to keep up.
Result: no warping at all after waiting for the plate to cool down, pretty awesome.
I know it is not a very good test, because the object is not very high, but usually I could see warping starting pretty quickly with such big parts, and this time I had no warping even without using any brim or skirt.
Another good point was that it was very easy to remove the printed object from the plate after it cooled down. It used to be very difficult with the cold mirror and I was always afraid to destroy everything. This print removed with very little effort once the glass cooled enough.
So, I still need to confirm over the next prints, but it seems to be a decent working solution!
[attachment file=40188]
[attachment file=40187]
[attachment file=40188]
The steppers and cpu can handle up to 190mm/s on the x and y, you are limited by the steppers ability of max strength dropping off above 60 rpms, not the actual step rate. If you need faster you have to get larger pulleys.
Yeah, I think you’re right. I’ll try anyway since it is very easy to check and it cost me nothing, just to be 100% sure.
I think I was pretty close to 190mm/s when it skipped steps. It was during fast travel, after cranking up the speed using the screen pot.
I could also go for a higher voltage, replacing the pollolu drivers with some TB6560, the same way I did for the 3040 CNC. I still have a bunch of those drivers from an other project.
Bigger pulleys is a solution, but I worry that the torque will be too low. Torque is not very needed in general, but I’m not sure it would perform well if I set up higher accelerations.
Kind of unrelated but I noticed I peak out speed in speed regardless of gcode when it comes to trochoidal milling. This is due to the acceleration limit in the firmware correct? The circles are awfully small so that’s what I would assume. Was thinking to changing to 16 steps but it’s not necessary I bet.
Yes, could be your acceleration setting and/or jerk settings.
I thought it was a memory thing. Those little loops take a shit ton of code compared to normal milling.
Dui look up a nema 17 torque curve, 60-100 rpms is the max before rapid decline in power. That is 32-53mm/s with a 16T pulley above that and the stepper gets weak fast. A larger pulley will have less power loss than the stepper itself, because you are keep the rpms down.
You can also turn down accelerations, or use a higher voltage power supply. In reality I just don’t think you should be trying to move such a large gantry at 190mm/s it is too low of torque and too close to the aduino limit. No need to push every single aspect of the build in one shot.
The “jerk” in Marlin isn’t the derivative of the acceleration. You might already know that, but I found that very confusing. Any speed below the jerk value doesn’t get the acceleration limit applied to it. So if you have jerk set to 1mm/s, then any movement from 0mm/s to 1mm/s is applied instantly, as opposed to using the acceleration value to smoothly move up to 1mm/s. These numbers are 100% independent of the 1/16 or 1/32 stepping, because the steps/mm will cancel that out.
So if the jerk was set higher than the speed you are cutting at (using trochoidal milling) then the acceleration value would not affect the cutting at all.
Well, another disappointing weekend with plasma. This thing is so frustrating, I really don’t know how to get it to finally work…
A reminder of the story: I tried for the first time a few months ago and the Arduino blew up instantly after the torch fired for the first time. Never had the time or the motivation to give it a try again until this weekend.
What I’ve done this time:
-Linking all tubes together, using some copper braided wire
-Attach this to the ground, together with the plasma ground
-Put the Arduino/ramps in a carton box, covered with several layers of aluminum tape (I had no metallic case available for this around)to create a Faraday cage
-Attach this box to the ground together with the others
-Put the plasma cutter as far as possible from the Arduino, and put the torch wires away from the motor wires as much as I could
Results:
-First attempt: normal configuration: the arduino is powered from the 220V (through the 12V power supply of course), the torch is fires from the relay. It started just fine, so I was super excited, cut a few letters from my test file, then crapped out suddently, the Arduino restarted… No damage to the arduino, I made a few other attempts but none of them was successful, the arduino kept restarting or acting weird
I thought it could be some interferences in the relay, so my next move was to disconnect it and try to control the plasma cutter using the manual switch.
Again, it started just fine, cut a few letters and then crapped out the same way.
So, I thought it could be a power supply issue, and decided to power the arduino and steppers through a scooter battery.
Same result, worked fine for a while, then stopped.
I tried several times using the battery, and at some point I guess something died in the electronics, the screen displays the startup logo and then blinks at a high frequency.
I really need some expert opinion here, I really cannot find out what is going on. I don’t want to give up on plasma, cuz it could be super useful in the shop.
This really puzzles me… Apparently it is not because of the relay, and not because of the power supply. I wonder what it could be… Any help would be greatly appreciated.
Here is the setup, the plasma cutter is on the right, the control box on the left of the machine, connected to the battery:
[attachment file=43097]
The Arduino is inside this box:
[attachment file=43098]
The results:
[attachment file=43106]
It doesn’t look very good, but I think that is because the plate I was using was pretty warped, so a lot of spots were too far. But the spots who were at the correct distance are actually very clean. Anyway, tweaking up to improve the quality is not the biggest concern right now, I need to come up with a way to make this work reliably, and if possible more than 10 seconds…
[attachment file=43107]
[attachment file=43108]
1 Like
I’m not an expert… but …
Shield the stepper motor cables … maybe … if you didn’t do that already 
or use those tb6560 drivers
1 Like
And the LCD cables.
1 Like
The motor wires are CAT 6 shielded wires, but I haven’t found a good way to connect them to ground yet. So basically, they are not really shielded.
For the screen wires, I want to shield them, but it is not really easy to find a proper way to do it.
Anyway, I just found out two things:
-My arduino isn’t dead: the extruder driver died, and it somehow conflicted with the Arduino. I just took it out and now everything works fine.
-I made the stupid mistake of forgetting to unplug all the unnecessary wires. I was using the MPCNC as a printer last week, so I had the thermistor, the hot end, fan, and extruder motor connected to the Ramps, with the cables going around the torch… No wonder why it crapped out, I’m actually surprized it was able to work for a few seconds. I guess it picked most of the noise from the extruder wire, since it destroyed it eventually.
So’, I’ve disconnected every unnecessary things and I’ll try again tomorrow to see if this makes things better (I really hope so). No matter what happens, success of failure, I will ground and shield everything properly to make sure it is reliable.
Fingers crossed!
You should be able to just strip off some of the plastic coating on the CAT6 and just ground it from there.
Putting the ramps and screen in a metal box with a screen cutout would be a better Idea I think.