Custom toolhead camera for Bambu Lab A1 Mini

[sgbaird]

@Neil-YL thoughts on mounting the camera to the toolhead similar to what's shown in makerworld as a "nozzle cam"?

Option 1

Option 2
Similar version

Option 3
Another one

In our case, we'd want to image the print directly above, and we'd move the toolhead to get the camera in the right place.

This is the kind of tripod camera mount we've typically been using: https://www.pishop.ca/product/pro-mini-camera-mount-for-raspberry-pi-zero/ (image below). However, the plate that the camera is attached to is "wiggly" relative to the plate the Pi Zero 2W is attached to.

I lean slightly towards the Option 1 style (use a flexure clamping mechanism onto the toolhead). I'd be OK with replacing the cover (seems more mechanically stable and EDIT: it's a pinch-and-remove mechanism). I lean towards reusing the same hardware (nuts/bolts) that come with the "Pro Mini Camera Mount".

Related:

• https://ca.store.bambulab.com/products/print-head-front-cover-a1-series (EDIT: the print cover seems to be removeable without hardware)

So, how about designing a new front cover that simply has a single 3d printed threads jutting out to be attached with the regular tripod nut? Or this could be the arm extension.

[sgbaird]

[sgbaird]

[Neil-YL]

You work so quickly!

I’ll test this part tomorrow. I am also thinking we can replace one of the AMS feeding tubes with the cable for the camera.

[sgbaird]

We'll see if it turns out 😅 thanks!

For provenance, link to Fusion 360 file: https://a360.co/4jq5PUd

Re: feeding tubes, good point - though maybe instead of replacing it, we use reusable zip ties to fasten it? (I have a bunch of these in my office)

[Neil-YL]

The finishing of Carbon Fibre material is quite coarse, I will print a PLA to see if it is better.

[Neil-YL]

For provenance, link to Fusion 360 file: https://a360.co/4jq5PUd

Nice you've attached the file, I don't know why but the project file saved on Bambu doesn't allow me to change the material

[Neil-YL]

I will dismount the camera next to the robotic arm and to see how it fits to the nozzle cover

[Neil-YL]

IMG_1650.MOV

I mounted the pico camera to the nozzle cover. I have tried to raise the temperature of the nozzle to 240 and the cover is still room temperature so I think PLA would be fine.

Do you want me to start the streaming for the next square patterns printing?

[sgbaird]

Awesome! This looks great. I have a bunch of reusable zip ties in the closet cupboard in my office. Could you grab some and work on getting the power cord attached with the rest of the tubing in the way that allows free movement? Lmk if we need a longer power cable and what length.

Really glad you looked into the temperature. Last night when I was telling my spouse about it, she brought that up, and I realized I missed that key detail 😅 thanks for checking that one.

Don't worry about streaming for now. I'll ping @Jonathan-Woo about getting the camera set up.

[Neil-YL]

Really glad you looked into the temperature. Last night when I was telling my spouse about it, she brought that up, and I realized I missed that key detail 😅 thanks for checking that one.

I thought you chose to print it with CF so it would have a higher temperature resistance. I switched to PLA so I think it is necessary to check. I also know that thermal management failure is quite an issue for many cameras in the market but at lease in our case the nozzle would not bring up the temperature crazily.

I will look into the cabling this afternoon. Since the OT-2 is occupied by the borrowing event test for this afternoon, I would like to have some jobs done on it now.

[Neil-YL]

btw I also fix the wiggling issue by filling the gap (with our printed 0.1mm thick square) between the camera mount and the pico.

[sgbaird]

Oh, nice! Can you share an image?

[Neil-YL]

The printer is printing I would take one after it finishes.

Here’s a video for you to check out first, looks quite stable during the print job

IMG_1651.MOV

[sgbaird]

Awesome! I guess another question: what's our new usable area on the build plate? I'm not sure how far "back" the gantry can go (i.e., if all corners of the build plate are accessible). Even if we lose about 20% of the build plate, I don't think it's a big issue, but we'll just need to keep that in mind for the workflow.

[Neil-YL]

Update:

This is the far end position of the plate:

I think a reduction of 20mm for the plate would be safe to use:

For the printed square pattern, it would be 10/11 = 90.9%, around 10% reduction.

[sgbaird]

Awesome! That seems pretty reasonable. Let's go with that.

[Neil-YL]

How I fix the wiggle gaps, for later design we can print a camera mount with a stable fit:

Not sure about the quality and how close the camera should go for Sissi later evaluation processing, but it is possible that we can still have access to all the area of the plate:

[sgbaird]

Oh, nice! Hmm.. can we get the camera pointing a bit more perpendicular to the plate? The camera should be pretty close to the build plate (i.e., as close as possible without making the rest of the toolhead crash into the plate). It would only be imaging a single square at a time. We'll keep in mind for later, but for now let's ignore that inaccessible 10% of the build plate.

[Neil-YL]

Oh, nice! Hmm.. can we get the camera pointing a bit more perpendicular to the plate? The camera should be pretty close to the build plate (i.e., as close as possible without making the rest of the toolhead crash into the plate). It would only be imaging a single square at a time. We'll keep in mind for later, but for now let's ignore that inaccessible 10% of the build plate.

I think so, current one is a little bit tilting compared to the one next to the OT-2. I will fix it after the print finished (I started another nozzle temperature print again)

I think it worth designing a new L shape part to mount the Pico and the camera so that we don't have too much flexibility for positioning the angle of the camera

[sgbaird]

Doing a slight bit of image registration shouldn't be a problem, let's leave the tilt angle for now and focus on getting the full workflow locked in. Later on, a new L-shape part could make sense. Also, the board on there is a Pi Zero 2W running a full OS.

[sgbaird]

@Neil-YL could you export the 3mf file (directly from Bambu Studio using File --> export to keep all settings) and attach that here? Then, I'll fix up the submission at https://makerworld.com/en/models/1019952#profileId-1001000

[Neil-YL]

@Neil-YL could you export the 3mf file (directly from Bambu Studio using File --> export to keep all settings) and attach that here? Then, I'll fix up the submission at https://makerworld.com/en/models/1019952#profileId-1001000

A1 Mini Cover with tripod screw v3 v1.3mf.zip

[sgbaird]

Reusable zip ties for power cord and power cord (note: power cord might be a bit short. Feel free to link to one from e.g., DigiKey with a longer cable length)

[sgbaird]

https://makerworld.com/en/models/1019952#profileId-1001000

[sgbaird]

We might need some external lighting for when the office lights go off. Is there a built-in light, and if so, do you think it's bright enough to give us OK image quality? (not worrying about difference between office lights on vs. off, just if we can get OK images when office lights are off)

[sgbaird]

I came across Bright Pi with some nice quickstart instructions which should work in terms of mounting onto the camera, but it requires soldering the LEDs and header pins on and might obstruct some of the camera view. An easy fix for potential obstruction would be to only solder the outer positions (since the LEDs seem to be individually addressable).

(cc @brendenpelkie you might be interested in Bright Pi in the context of the overhead camera tool for Science Jubilee - feel free to unsubscribe from notifications for this thread once you see the message).

Alternatives:

• LISIPAROIWHT-01 Add-On Board, LED Camera Light Ring For Raspberry Pi, White (not individually addressable, but probably doesn't matter for our application, and it's just a GPIO input which is much simpler)
• Adafruit: Cool White LED Ring Light with USB Cable and On/Off Switch - 70mm Diameter 5V seems like a nice, easy option in terms of on/off and voltage, but it will get hot. (EDIT: this will get too hot)
• Adafruit LED Ring Light - 76mm Diameter (9V), but even with the heatsink this might get pretty hot if used continuously. Also not straightforward to mount.
• This RGBW neopixel ring might be a nice option, but again, mounting and heat management aren't straightforward.

---

There might be some obstruction of the laser cut piece the camera is attached to.

Ignoring the red circles, as a ballpark, it looks like the hooks at the 90 degree angle would interfere with the camera.

I lean a bit towards designing a 3D printed piece (as @Neil-YL suggested earlier) that mimics the existing camera mount, thus removing the hooks at the end. This could even be printed directly as part of the front cover, eliminating the need for a tripod screw and nut. We can still reuse the screws, nuts, and standoffs from the camera mount kit.

Ballparking, it looks like there might be enough clearance for the Bright Pi or the LISIPAROIWHT-01 relative to the rest of the toolhead, but it's simple enough to add some extra thickness to move the camera away if needed.

---

@Neil-YL I went ahead and ordered LISIPAROIWHT-01 from Newark. Should ship relatively soon.