Thinking about designing a dedicated OpenPNP controller board – what's your biggest hardware headache right now?

[Tero Koiranen]

Hi everyone,

I'm an electronics student and I've been deep-diving into OpenPNP through build videos and documentation lately. Coming from a pure hardware and electronics background, one thing stands out to me: most builds seem to be fighting with modified 3D printer boards or generic CNC controllers.

It looks like PnP machines have very specific demands (huge amount of axes, electrical noise from pumps/solenoids, messy feeder wiring) that these generic boards weren't really built to handle.

I'm really tempted to sit down with KiCad and design a dedicated, rugged controller board from scratch, purely optimized for OpenPNP. But since I'm looking at this from a hardware designer's perspective rather than an experienced machine operator, I wanted to check in with the community first.

    What is the absolute biggest hardware headache or bottleneck with your current controller setup?

    If you could change one thing about the hardware market for OpenPNP right now, what would it be?

    Are there specific things (like connectivity, opto-isolation, specific voltage rails, or feeder bus ideas) that you feel are completely missing?

I want to map out what a "perfect" board layout would look like for a modern build. Any feedback, wishlists, or rants about current hardware flaws are highly appreciated!

Best regards,
Tero Koiranen

[Pierluigi Garaventa]

Hi Tero,

 I suggest to check Duet3D Mini5+ Board.

I'm using it in my DIY OpenPnP setup. I have to say that this board works really well for a desktop PNP machine.

I'm able to control 5 axis, vacuum pump, 2 vacuum sensor, 2 led camera light and X,Y stop all with DUET3D board. 

I'm not really sure if it's possible to do something better, but perhaps you can optimize the cost.

However, I suggest to spend some time regarding feeders. It's a really important part of the PnP machine.

It's both a mechanical and electronic problem. I personally feel OpenPnP lacks a reference for a good feeder that can be integrated into any DIY machine.

Ciao,

Pierluigi

--
You received this message because you are subscribed to the Google Groups "OpenPnP" group.
To unsubscribe from this group and stop receiving emails from it, send an email to openpnp+u...@googlegroups.com.
To view this discussion visit https://groups.google.com/d/msgid/openpnp/836ce0dd-36f2-4e90-b801-465c0b368038n%40googlegroups.com.

[paulm]

I'm not an expert by any stretch of the word, I'm building my first PNP right now (for fun) and I have no prior experience with CNCs.

That being said,the only thing you can reasonably optimize is cost. Pretty much every board currently used is open source, including the very much popular Duet. Which (and I'm not saying "for no good reason") is quite an investment for people dipping their toes in this area, so to speak. Doubly so if they're not-making-money off their PnPs. 300 Euros is no walk in the park. Even the cheaper versions will set you back 150 or so. Which is a lot, especially if your build is small and you don't need 10 billion amps drivers and whatnot.

However, I think the BOM for a Duet2 is in the hundred bucks range, if I'm not mistaken (at a glance). If you can even source the parts, that is. So the question would be "is your time ordering the PCBs, import forms and duties, ordering parts, assembling the board worth 50 bucks"? Ok, you can throw away useless stuff like ... temp control? But that's not a noticeable reduction in the final cost. And it's not like there's any real estate cost, PCB-wise. (And you can't say Duets skimmed on the noise isolation or safety).

At the end of the day, the main issue is not the hardware. If you consider building a controller board, I think the issue weighs more on the firmware side of things. How would you approach this problem? It's way more complex than it seems at first glance. When you're buying an [insert board name], you're essentially supporting the developers, developer who, in turn, keep updating the hardware/ firmware. With emphasis on the latter. Not to sound condescending, I think we can both agree neither of us is going to whip up a superior solution to what the guys who live and breathe this particular nieche put out. Not in any reasonable time frame. I was younger than some of these boards when I bought my first beer. :))

If however you want to build one because you're a freakin' engineer and, by God, that's what we're supposed to be doing, then what I would advise you to do is to scout an untrodden path and enjoy the ride. For example, I'm going to borrow a duet from my godfather, God bless his soul, until I have my PnP up and running. After which time, I'll get to work on a RP2350 based board, running grblHAL. The 2350 is surprisingly absent from the things-that-go-vroom scene. I mean, a MCU you can reliably overclock to 800Mhz without cooling, dual core M33s, with FPGA-like CPU-independent cycle-perfect IO capabilities?! And [almost] nobody's using it to drive motors?! Instead relying on expensive MCUs that can barely hold a candle to the RP, at 5 to 10 times the price? Ok, I'm not being entirely fair, I realize that, but in the context of OpenPnP, a 3D printer, a CNC in general, the RP2354 (2350 with built-in flash) is as close to a perfect controller as you can reasonably get, without going the FPGA route.

That's just my 1.75 Eurocents.

Regardless of what you choose, the main thing is to... have fun! And, of course, post your results when you have some!