I am excited to share “SmartPrintCoreH7x”: my open-source Klipper-compatible 3D printer mainboard designed for ultimate performance and customization! Check it out and let’s innovate together!
https://github.com/BoltzRnD/SmartPrintCoreH7x
I would love to get everyone’s feedback to make this design better for wide adaptability in most machines out there. Thanks.
Congratulations on putting the board together.
I’ve done a quick scan and here are some comments:
As I said, it’s a quick scan and some things aren’t immediately obvious and I may have missed them in my questions).
Thanks a lot for the congratulations and your detailed feedback! It’s really helpful for improving the SmartPrintCoreH7x.
I’m currently the only developer on this project, but I’m looking to bring more people on board as we take on more ambitious projects aimed at mainstream 3D printing.
Here’s a quick response to your points:
Thanks again for your thorough review. Your feedback is key to making this project better, and I’m keen to hear any more suggestions or insights you might have. Really appreciate your input!
Just so we’re clear, I have not done anything resembling a “thorough review”. The original feedback was the result of a quick scan of the schematics and I asked about some things that jumped out at me.
A few questions back:
How do you expect the UVLO to work? You seem to imply that if the voltage is below 18V, UVLO kicks in but, in my mind, this means the output of the Bucks are turned off - the voltage doesn’t pass through them with the regulation turned off.
On the same point, IC1 has a different UVLO threshold than IC2. I’m curious as to why you’re doing that.
With the two USB C connectors, I presume that you would power an rPi using J64 and communicate with it using J34. How are you preventing a ground loop in this case?
Debugging Ports. Sorry, I don’t know what you mean - when I look at the schematic above and to the left of the MCU, I just get the DFU buttons:
On the schematic point, as well as making the PDF searchable, it would be appreciated if you would rotate it 90 degrees counter-clockwise as well to make reviewing it easier.
I’m not doubting that having handshaking ensures more reliable communications, I question it because a) when it comes right down to it nothing that is being done here is that critical and b) as far as I know, there are no standard sensors or hardware that use them.
One of the most likely use cases for the serial port is a Raspberry Pi and it has no built in handshaking lines (which means they would have to be implemented in software):
Finally, who are you targeting and what product space? You have a very premium board here in kind of a funny capability space and I don’t see it being salable for less than $200 USD, in quantity. Your initial builds will be closer to $500 each (this comes from hard won experience).
Right off the top of my head, your closest competitor is the Duet 3+ Mini (which retails for $140 USD with TMC2209 drivers built in). For boards with 5 drivers, most people will be looking at:
Each of these boards can be purchased for well under $100 with 5x TMC2209 drivers. Some of them, like the Manta M5P have features that can make their use, especially with Klipper, quite a bit easier than what you’re proposing.
Sorry for the hard news but I don’t want to see you invest a significant amount of money without a strong likelihood of success.
Hi, Thanks for the feedback,
Regarding your points:
UVLO & Buck Converters: We have three distinct buck converters to suit different needs:
1. 5V 10A (U2): Powers SBCs, LCDs, cameras, etc., with an 8V UVLO to ensure stability.
2. 5V 5A (IC2): Dedicated to fans and servos, also with an 8V UVLO to handle inductive loads that might disrupt SBC operations if shared.
3. 12V 5A (IC1): Used mainly for 24V input systems; cuts off at 18V UVLO to protect components.
These UVLO settings are tailored to the specific demands of each buck to ensure reliable and safe operations.
USB C Connectors & Ground Loops: The TPS2116 chip manages power prioritization, effectively handling potential ground loop issues. Our design ensures stable operations, typically with standard 1m USB cables.
Debugging Ports: The debugging connector, J27, is indeed near the MCU on the board’s layout, providing straightforward access for updates or troubleshooting.
Handshaking & Serial Communication: Including handshaking lines adds a layer of reliability, especially useful with higher-end SBCs like the Jetson Nano and some variants of beaglebones. This feature is optional and doesn’t affect users preferring standard serial communication.
Board Cost and Manufacturing: Yes, the board is truly premium and is designed to enhance the reliability of connections and peripheral operations of the machines it is used in. However, I’d like to clarify that the board doesn’t actually cost a lot. In fact, the production costs for quantities of 10 to 25 are close to $130 per board, including the complete connector set, that too if the components are sourced from distributors like Digikey, etc. Furthermore, costs can come down to approximately $70 with some supply chain management.
In comparison to other boards that are mass-manufactured in quantities close to 500 and even 1000s, the components are sourced from LCSC or other local distribution channels, which, if applied to SmartPrintCoreH7x, could bring the costs down to $45 without changing the BOM. This is very feasible for any manufacturer that is mass-producing the board, Unlike many boards mentioned, which use obsolete components that may have vulnerabilities or supply issues, which is not feasible for open-source designs.
Of course, there will be a difference of approximately 25% of the costs, which is totally justified given the upgrade in reliability, high power capabilities, and advantages offered in the design. All of this was studied thoroughly before the design of the architecture of the board, and I was fully aware of the tradeoffs. Hence, I am confident that this design will be beneficial for the community. In case it isn’t, the design can always be tuned to facilitate the targeted needs of the users. As a matter of fact, I’ve received many pieces of feedback that the motor drivers are not sufficient for new designs, for which I’ll be working on expansion modules to meet these requirements. This is definitely an example of the advantages of open-source designs, and the product is built with you in mind.
Sorry about the comment regarding the power supplies. I didn’t follow through on the schematic properly (did I say something about turning it 90 degrees so it can be read easily on the github page?).
I must take issue with your statement claiming that the TPS2006 “effectively handling potential ground loop issues”. If you’re plugging in a USB C J64 to the power connector on a Raspberry Pi and USB C J34 to the power port - you definitely have a ground loop through the cables’ shield and probably the ground connection.
I don’t know what you’re electronic board manufacturing experience is - I personally have 30+ years with a bit of aerospace. I’m not sure you’re getting that the BoM and soldering operations are only part of the cost during volume production and a small percentage of the cost during prototyping.
When I said $500 per board for initial builds, I’m including costs for placement machine programming, stencils, temporary pass through line setup setup, temperature profiling and so on. Along with that, you have a lot of hand placed parts (I’m presuming all those connectors are PTH) that will require somebody to write instructions for the placement operators and have a designer sit through the initial builds to make sure that everything is correct (If you’re board stuffer is not local, then plan for travel expenses). Don’t forget you need to arrange and pay for storage for the parts before assembly (none of which will arrive at the same times as any of the others). Before you start assembling the boards, you’re going to need to load firmware onto the MCUs. You also need boxes, ESD bags, parts bags for the various locking connectors that are non-standard.
You’ll also need business insurance, an accountant with software resources to set up the company as well as the payment system/CRM - you can sell on Amazon or Ebay without this but if you want a distributor to sell the boards, you’re going to have to have a company that they can work with electronically, can see that you’re insured so they can pass any potential problems onto you and know that their money won’t primarily go to paying for that Ferrari you’ve always wanted.
Once you get that stuff set up with some prototypes to show customers/distributors/thought leaders along with a company they can work with and have confidence in, then you can expect orders which will drive down your costs as you do follow on builds. Regardless, I still think you’re being overly aggressive in what you think your costs are and not seeing the big picture with all the work and out of pocket expenses you have in front of you.
Finally, you didn’t address my last point. What makes this board special and would entice somebody to buy it? I listed five boards, all of which are well known with good community support, are sub $100 with drivers, work well with Klipper and some of them, like the Manta M5P, have features that make people get excited about using it in a printer.
You’re emphasizing reliability and good engineering, but when I look at your board:
That doesn’t come through.
What I do see is a hell of a lot of connectors - at least twice as many as on the Robin Nano V3 and Manta M5P I have sitting in front of me. I don’t immediately see the value of all those connectors - if I have a printer that requires five steppers, it’s probably going to be a bed slinger with with two Z axis steppers for Z-Tilt. There are 'way too many options built in here for that type of application.
Along with the apparent complexity, I don’t see the advantage of the locking connectors for the power in and heater outputs. What I see is that you’ve replaced the need for me to put a ferrule on each line and shove it into a screw terminal block with having to crimp on a custom connector. You’re locking me into this board (pun intended) and that’s not a feature I’m willing to pay for.
If I can give you any kind of advice on the product: it is start small with a single, well defined use case with a feature that will catch people’s eyes and make them consider buying the product. Don’t try to meet the needs of everyone and every printer type; design for a specific type of customer with one use case.
Sorry for dumping all this on you I can see you’ve done a lot of work but I don’t think you understand that it’s a fraction of the work that you have in front of you. I do admire what you’re trying to do but i think you need to really take a hard look at what it’s going to cost you and what it will take you to break even. I would suggest talking to somebody local that is involved in electronics design, manufacturing and distribution to review your plans and give you advice - depending on where you are, there may be a government funded industry development organization that you can reach out ot.
Hi,
Thanks so much for your detailed feedback. I appreciate your insights and the depth of experience you bring from the aerospace sector.
Updating the repository with rotated schematics is on my agenda, but it might take some time as I’m rolling out corrections in phases. Meanwhile, you can use the PDF rotation feature in tools like Google Chrome to suit your viewing preference.
Regarding the power setup and USB configurations, I assure you they’ve been rigorously tested over 300+ hours of printing without any issues like distortion or packet loss. I’m confident in the design’s reliability, but I’m also open to further testing and refinement. If you’re interested in a closer look, I encourage you to arrange a prototype from a local PCB fab service using the design files available on our GitHub. This way, you can personally assess the performance and provide direct feedback, which would be invaluable for continued improvements.
I understand your concerns about production and cost—coming from an avionics background, the jump to consumer electronics was eye-opening for me too. The economics of mass manufacturing can indeed dramatically reduce costs, sometimes in ways that seem surprising from a high-reliability sector perspective.
Manufacturing isn’t my primary focus; my passion lies in design and helping the community advance open-source 3D printing solutions. That said, I ensure that each project is commercially viable and leaves room for effective supply chain management and production by those who specialize in it.
For a detailed breakdown of the features and thought behind our design, I invite you to revisit our GitHub page. Your feedback is crucial, and I’m committed to incorporating community insights to enhance the board further.
Since this project is open-source, ongoing community engagement like yours is invaluable for testing and refining in diverse systems and environments. Together, we can strive to make this a new standard in the 3D printing world.
Thanks again for your review and advice.
Best regards,
Yatin
This topic was automatically closed 60 days after the last reply. New replies are no longer allowed.
