[Marinus de Beer]

Wanted to add, this doesn’t occur when I open the SVG and convert it to DXF and set the pen up and down tolerances to 1.0 and 0.5 in the dialog prompt. But converting it to DXF every time is somewhat of a difficult workaround for a lot of people I guess.

[Marinus de Beer]

Changing back to the master version fixes all the issues I’m having with the dev version. I’m not sure what the difference is, but all I can say is that the dev version absolutely does not work. The dev version does however have a correctly working rotary encoder and I really like the SDFAT system. If anyone has some advice on how I can ‘fix’ the dev version so that the motors work again, I can finally start testing the dev branch. Am I in the right place? Should I put my comments in the github? It’s kinda quiet here 🙂

[Marinus de Beer]

Okay rotary encoder in the dev branch seems to work really well, however, the right motor seems to suddenly be working the other way around and the home position is not working anymore. Turning around the cable for the right motor seems to fix it a little bit, but the machine now starts printing up way too high.

Do I need different settings for the new dev firmware? I cannot print with it the way it is because either the carriage goes to the top or it rams into itself. Homing is disabled and I have made sure all the settings are identical to the latest master firmware that works fine on this machine.

Oh and some advice for the file menu, SDfat works really well but it would be nice if the list was ordered in such a way that the latest file would be on top since now you can’t scroll to the bottom of the list by moving up with the rotary encoder.

[Marinus de Beer]

Hello, I finally managed to fix the servo issue. It turns out the blue plastic geared servo motors do not work with this hardware and firmware combination. But I got a friend to borrow me a couple of metal gear servos to test and I managed to find one that works.

No clue how fast micro servos wear on a vplotter, but fixing the servo naturally fixed all of the other issues I was having except the rotary encoder of course.

I just discovered I’m using the latest master and not the latest DEV. I’m terribly sorry. I never understood github that much and I will try out the DEV branch later today!! Pretty ashamed to say I only found out about the DEV branch today after years of using github haha. I never clicked the ‘branches’ tab and always assumed the main page would display the latest branch. This is how it works with Marlin for example.

[Marinus de Beer]

Okay, so I have found out three things:

1. The servo speed is not working correctly. I have compared it to my other plotter that is running with Marlin and the servo speed is really slow on the Makelangelo firmware. I am not sure what the issue is but I’d like to have some tips on how I can troubleshoot this. I will convert the carriage soon so that it doesn’t lift the carriage but instead only lifts the pen.

2. The ‘spiral’ option in Makelangelo software creates spirals that slow down the plotter. My guess is that the circles have too small segments. Whenever I print anything with curves the machine slows down and speeds up in predictable intervals. Once again, I am not sure what the cause is, but the machine has no trouble working through any other generated file.
The spiral file is so heavy on the hardware that even the rotary encoder becomes unresponsive.

3. The rotary encoder knob is very hard to use, I have tried three different panels that all work fine on Marlin-based machines. So there must be a setting in the firmware to tune this. Any hints on where I can try out different values for the rotary encoder knob settings?

The machine now has perfect dimensions and all lines are straight so props to the designers of the firmware. I am really surprised to see such high definition in the prints that the machine makes even though my machine is still in development.

I have created one file with gcode that draws a template onto the base plate with all paper sizes outlined in both landscape and portrait mode. This works beautifully. The base is made out of a metal plate that is glued to a wooden plate with spray adhesive by 3M so paper can be fixed to the machine with small flat magnets.

Another improvement I made is to add a cooling fan to blow on the heatsinks of the Ramps board. I didn’t have issues with overheating yet but this is will lengthen the lifespan of the DRV8825’s. The drivers are working really well however, I have two spare TMC2208’s and am planning to swap them out some time in the future. I will have to find out how to map the pins for these drivers.

Eventually I decided I want to turn the machine into the best vplotter ever but I just got scammed yesterday and lost all my money and savings so I cannot invest in any new parts for the coming year I guess.

[Marinus de Beer]

Here is the video of the latest success!!

Now the only thing left to fix is the responsiveness of the rotary encoder on the LCD panel (any way I can set the sensitivity?) and fix the servo being absolutely slow as hell. Would you know a fix?

[Marinus de Beer]

Okay nevermind, I fixed it.

I don’t know why, but I tried to upload the firmware again, did another D11 and this time the only thing I did differently was that I ran M101 to check the values. And somehow now the new values stuck in the memory and it drew a PERFECT A4 hahaha.

I’ll add a video soon since I can’t edit my posts I’ll do that in a new reply, sorry for that.

[Marinus de Beer]

Thanks, I had already gone on and found out how it worked (the post was waiting for approval for 4 to 5 days so I decided to go ahead myself and figured it out eventually 🙂 ).

Indeed, I have disables HAS_LIMIT_SWITCH and also removed the ‘go home’ from the menu because even though limit switches were disabled, it still displayed the option. Now, my rotary encoder really doesn’t work nicely in the firmware so I sometimes by accident clicked on the ‘go home’ option causing the machine to crash into itself a bit.

Right now the machine is slowly improving. This is where I am now:

The maze outline is supposed to be an A4 rectangle. The base paper is an A3. So you can clearly see it’s not correctly set up.

I have tried lowering the home position by a lot but then it shrinks down the entire image by a lot and still has the bulging top. And the image stays way too small in the x-plane/horizontal axis.

What I did:
– set correct dimensions in firmware and software
– set correct pulley teeth (16)
– correct 200 steps/rev
– set 1/32 microstepping in both FW and jumpers on the ramps
– tried lowering the home position manually to get a better picture

The centers of the motor shaft are 958 mm apart.
Home position was started at 217 mm down from the center of the shafts and was moved down to see if it improved anything.

Maybe I misunderstood the D11 settings. I have changed the settings in the firmware to match my own machine:

void makelangelo5Setup() {
  // if you accidentally upload m3 firmware to an m5 then upload it ONCE with this line uncommented.
  float limits[NUM_AXIES * 2];
  limits[0] = 479.0;
  limits[1] = -479.0;
  limits[2] = 479.0;
  limits[3] = -479.0;
  limits[4] = PEN_UP_ANGLE;
  limits[5] = PEN_DOWN_ANGLE;
  adjustLimits(limits);

  calibrateLeft = 1025;
  calibrateRight = 1025;
  saveCalibration();

  float homePos[NUM_AXIES];
  homePos[0] = 0;
  homePos[1] = limits[2] - 217.0;
  homePos[2] = 50;
  setHome(homePos);

}

After uploading the firmware I run D11 through the serial of Arduino to install the settings from the firmware in the EEPROM (I am assuming it works this way).

Another issue is that I have set the servo speed between 100 and 1000 in the software, but it still moves really slow whereas when I use ‘pen up’ and ‘pen down’ it just switches really fast. Can it be there is something buggy in the firmware, or can it be the machine runs differently with timing pen up and down when drawing (and moving the stepper motors)?

[Marinus de Beer]

Sorry for replying everywhere 😉

I prefer physical endstops! But I also have experience with the TMC2130 drivers. However, I prefer the TMC2208’s in my machines as I do not have a use for StallGuard at the moment.

The vplotter I’m building is meant to be made from leftover materials so definitely can’t buy anything for it. Got it working pretty wel so far. But the dimensions are way off. I’ll reply in the topic I started myself to explain everything.

[Marinus de Beer]

Hi Ian, I know this is an ‘old’ thread, but still I wanted to let you know my thoughts about the heavy board you’re using.

Actually, using a lighter board will make the vibrations more apparent because it takes less energy to vibrate the plate. A heavier board has more mass and therefor usually dampens the motors more. The ideal solution would be to have a stepper motor mount with a rubber dampener in between. These are pretty common in the machining world, especially when it comes down to medical equipment and a lot of commercial printers have these as well.

I think they are sold on aliexpress for just a few bucks per dampener so it’s good to check there as well.

Honestly I am also using DRV8825’s with 1/32 microstepping on a Ramps 1.4 board and am having no issues with the sound. Can it be you have set the vref correctly? This setup is way quieter than using A4988’s with 1/16th ms.

Right now I’m working on a selfbuilt machine and after solving all the other issues (took me over 30 hours to fix all the issues I have encountered with this self-built machine) the only problems I’m having are dimensional accuracy issues and I want to add endstops because the manual positioning is starting to get pretty old pretty fast.

It would be nice to see an update of your work as well seeing we’re almost a year ahead now.

[Marinus de Beer]

For me I still can’t connect with Makelangelo-7.21.0-with-dependencies.jar

In Arduino it shows up just fine and I can upload the firmware etc.

[Author]

Posts