UBW32 - Powerful and semi-free

As a follow-up on Makerbot electronics, the newer version (With future compatibility) could be powered by the UBW32. I know i am going to get hell for heading down the path or Microchip or PICs but hear me out.

UBW has many MANY pins for anything you might want to do. Its similar to the Sanguino in that it is a 'stick' that can be purchased separately from the motherboard itself, but still can be easily incorporated as price falls.
The onboard memory and peripherals are awesome! You could have output for 5 steppers, many analog inputs for all sorts of analysis, and so much more.
It can operate at a cruising 80Mhz, which means it can host through USB for direct unadulterated high speed connections! This thing is capable of many many things.

Bonus, Microchip offers a free command line programmer, C or ASM compiler, full IDE, and a few other little items. Also the chip itself has a USB HID bootloader. No programmer required. If you really want to reprogram the Chip past the bootloader then you can buy a $50 PicKit2 or 3 and do something fancy.

I dont know what people think of it. I felt this device added so much while maintaining the essence of the current Makerbot Reprap motherboard.

The amount of power this device holds in its tiny package could potentially allow for a single board to perform both Motherboard and Extruder tasks.

FYI, if anyone is interested in moving in this direction, i have already ordered a couple UBW32s and plan on giving it a shot. I am no expert though, so if anyone wants to work together on this let me know.


I found this link since i posted this.
[builders.reprap.org]
Uses Pic32, but not using UBW32. Will be doing much reading now.

Edited 1 time(s). Last edit at 05/07/2010 05:17PM by CidVilas.

It’s great! I’ll give it a try. Would you please send me a copy of your code? Thank you so much for your time.
Tonyphan432 at yahoo dot com

Not a bad chip, lack of Factory bootloader is what puts me off.

Microchip's dev tools aren't the best either.

Are there any people interested in this project? I havent gotten started on this, so there is no code available for others from me, and i havent ran across any code myself either. If you are interested, please post here and send me a Private Message. Then we can discuss the details of this being a possibility.

I know Microchip doesnt have the best tools for development, but it would be interesting to see where this goes. Could make for an excellent alternative.

Are there any people interested in this project?

hmmm... CidVilas, this sounds like development.

I'd be honored if you join us in reprap-dev our (super-secret) developers' mailing list.

Since you're a developer.

---

-Sebastien, RepRap.org library gnome.

---

Remember, you're all RepRap developers (once you've joined the super-secret developer mailing list), and the wiki, RepRap.org, [reprap.org] is for everyone and everything!

I like the general idea, but suggest instead the Atmel 32bit micro AT91SAM7XC256B. It DOES have a built in boot loader, and can be loaded through either the serial port (2 USAURTs) or the USB 12.0Mbit port. Further, it has an ehternet MAC built in that will run 10/100. All you need to add is a phy chip, transformers and connector. Looks like an extra $15 for hardware adds an ethernet interface. And we can still stick with Atmel development environment.

For $15 in single unit quantities, you get 256Kbytes flash, 64Kbytes SRAM, 81 I/O pins, 55Mhz, more PWM channels, built in USB transciever with 1352 byte FIFO buffers, DMA, 8 channels of 10bit ADC. And the SAM-BA boot loader and Windows GUI.

Mike

@SebastienBailard: Thanks! Im on it now and am mostly listening in to the conversations at this point. Ill chime in when the time is right. Cheers!

@rocket_scientist: I can't say either way, but Pic32 has most of what AVR32 has except it carries more processing power. Pic32 tops at 80Mhz. My real point is that we should really start looking into overkill as far as electronics are concerned, rather than a conservative approach. I'd rather have a ton of unused memory and ports than evolve and hit the limits of the electronics. I imagine the brain being external to the rest of the components. This way you can carry the brain (Whether PIC32 or AVR32) to a new set of electronics and then download the new firmware, and you are on your way. Its silly that we have to buy a whole new board when the brain can remain and the other components evolve around it.

Its just a thought, but i would really like to make it to the point where the electronics are no longer the limitation, but only software/firmware.

I agree complete, CidVilas. I just have a lot more experience programming the AVRs, and from what I have seen so far, the architecture is more flexible, especially PWM which I use extensively for DC motor control in the under water robot I am still working on. But the basic concept off more processing power, more memory, more I/O ports setup for step/direction stepper motor controller input, adding PWM channels for motor drivers or servos, more analog inputs and built in thermocouple channels, more heater control channels with heat sinks designed in from the start.

Some people argue against putting everything on a single mother board, so that if a single MOSFET or stepper driver burns out, you need to replace the whole thing. Perhaps we can compromise by putting the high risk parts, MOSFETs, stepper motor drivers, and DC motor drivers on DIP headers, so that if one burns out, you pull the header off and plug in a replacement. You can also use the same approach to reduce the early purchase price by not getting all 7 stepper-drivers or only half the MOSFETs. Then, as you added extruders or milling heads, you buy additional headers with the appropriate drivers and plug them in to get the remaining channels fully functioning. This will increase the total cost of a fully packed board, but allow for easier, cheaper repairs. And upgrades and replacements can be made by people who are not comfortable with soldering if someone is willing to mass-manufacture the headers. Hhmmm, sound like a feature to add to a reprap!

I would also rather use screw type terminal for all the high current connections, motors and heaters, and standard ribbon cable connectors for the rest. That way the high current is more easily handled, and the other connections are simple and cheap. Clear marking of polarity on the ribbon cable connectors will be important.

To provide the high current needed on the motherboard for the heater MOSFETS and motor drivers, it will likely be necessary to solder copper wire on top of the power traces, otherwise the standard copper plating thickness would make the power traces way to fat. Even then they should be 200 mills wide to make soldering the current carrying wire on.

All this will lead to a largish board, but it should be less expensive to make a single large board than several different types of smaller boards.

Mike

I think the price diff should be even less than 15$, because the FT232RL is about 8-9 eur to me, and/or the UM232R module was around 25-30 eur; although the new "comm" chip atmega8U2 (on the new mega boards), that one is like 3 eur only.

Judging the price in context, its my belief that using a chip with integrated features would / could be generally cheaper for the overall solution than otherwise (besides just being better).

Has anyone got a working PCBA using the UBW32 ??

We have machines ready to use them..!