I really wish there was a 32bit controller board, small size,

3d-faq
3d-qa
3d-help
gplus
(Stephanie A) #1

I really wish there was a 32bit controller board, small size, with 4 TMC2660 stepper drivers for under $100

This is a vastly underserved market. A few years ago I started on a similar design but never finished it. Near Drop in replacements for lower cost printers with minimal required features.

(Taylor Landry) #2

Considering 4 2660 drivers retail for around $40, that’s not exactly an easy thing to do.

Yes, you could price out a 32-bit board as described for well under $100 in component costs, but unless you’re making one yourself, manufacturers have to make money, so there’s going to be markup.

An Azteeg X5 mini wifi is $113 with 4 TMC 2224 drivers.

Unless you really need 2660s for some reason, that’s pretty darn close to what you’re asking for.

(Taylor Landry) #3

or, you can order the board without any bundled drivers for $89 and BYO 2660 drivers

(Ray Kholodovsky) #4

The 2660 is too large to put on a standard stepstick, that’s why there is Bigfoot stick, which is on the GT and not the Mini.

(Ryan Carlyle) #5

I’d suggest looking at a cheaper, slightly less awesome (but still really good) driver like the 6128 or 5984 if the Trinamics are too expensive. 90% of the 2660 capabilities don’t get used by most users. For example, I KNOW you don’t need ~2.5A steppers in that Printrbot you’re rebuilding.

(ThantiK) #6

http://www.panucatt.com/Re_ARM_for_RAMPS_p/ra1768.htm + existing RAMPS + Drivers?

(ThantiK) #7

@Elias_Bakken I thought was supposed to have a new uber amazing board coming out for a super awesome price, but I’m not sure if that vanished or not.

(Elias Bakken) #8

It has not vanished! We’ve actually spent forever working on the pricing which is why it’s been so delayed. It will not be under $100 though. I’ve got the first of several blog posts with more information ready, just waiting for go from the other parties:)

(Griffin Paquette) #9

It’s really tough to hit that pricepoint without it being a Chinese clone (Duet WiFi from AliExpress).

Personally I’ve wanted a 2130 board and 32 bit for a while.

(Whosawhatsis) #10

Cost aside, it bugs me that most controller boards are so much bigger than they need to be. You can get the same processor on a breadboardable dev board and use it to lay out an equivalent circuit in a smaller footprint than a lot of these controllers. When you’re building bigger and bigger printers, it might not be a big concern, but these huge controller boards look pretty funny hanging off the side of something like the Cranestyle Mini.

(Ray Kholodovsky) #11

Don’t forget about the heat dissipation requirement.

(Rien Stouten) #12

Just use Klipper.

(Whosawhatsis) #13

Yes, head dissipation is a concern, but you see a lot of boards with a bunch of dead space around flat surface-mount MOSFETS next to a bunch of Pololu-style stepper drivers sticking up from the board, when they could use some of that wasted vertical space to mount a heatsink instead.

In fact, a really smart design might even be able to take advantage of the fact that most machines these days have lots of aluminum in their frames. Granted, it would be tough to do in a DIY-friendly way, but it would be cool to be able to to sink heat from the drivers and MOSFETS all to the printer’s frame (or, optionally, to one big heatsink that can optionally be purchased along with the controller).

I bet if you designed for it specifically, you could get all of the drivers and mosfets into a 40mm square area that you could attach a standard CPU heatsink to. Then if you needed to run more current, you could just get one of those fancy rigs that overclockers use.

(Stephanie A) #14

If smoothie had pressure advance control I’d consider it.

I really don’t like RAMPS. It just looks like a mess, don’t like the connectors. My favorite board so far is the RAMBO mini, but even that has some faults in the circuitry (needs more filtering and ground isolation on the ADC).

A lot of these boards have twice the amount of headers that a low cost printer needs.

The current carrying capacity of the TMC2660 is a nice to have because some of the 0.9 degree steppers I like are 2.3A

Even without the 2660’s there really isn’t anything that hits this market.

(Taylor Landry) #15

Heat dissipation is not the reason they’re the size they are. Or at least it shouldn’t be.

RC ESCs are tiny and have a crapload (metric unit) more current running through them.

12-20a is easy to deal with. There’s just no incentive for board makers to be efficient right now.

(Whosawhatsis) #16

In fairness, RC ESCs are thermally isolated from one another and usually mounted so that they have a lot of airflow (sometimes a LOT of airflow), which will help keep them cool, even if they are wrapped in heatshrink.

Bunching all of the components that need heat dissipation together is bad for heat dissipation, unless (as I suggested) doing so allows you to dump heat into something else, like a big piece of aluminum.

(Ryan Carlyle) #17

@Taylor_Landry1 It’s not really the ~12A through the board traces that’s a heat issue, it’s the stepper driver H-bridges. They have a non-negligible RDSon for resistance through the silicon that is difficult to reduce. For ~2.4A (peak per coil) drivers like the 2660 you’re putting ~1.7A RMS current through several hundred milliohms, times 2 coils times 4+ motors. It’s enough heat that board thermal design and size does actually matter.

(Taylor Landry) #18

@Ryan_Carlyle Interesting. I guess that makes sense. So the issue isn’t the total current, per se, it’s the H-bridges switching that generates cumulative heat?

(Whosawhatsis) #19

ESCs have H-bridges too, don’t they?

(Taylor Landry) #20

@Whosa_whatsis yes they do, but I’m guessing the difference is that you aren’t switching very often (generally just going forward), whereas printers switch constantly

But I’m sure Ryan can clarify