I was running the original SD4 printrboard so to me the upside was DRV8825 drivers and a 2nd hotend mosfet (in addition to fan mosfet) for possible future upgrades. And the fact that it's ARM32 wasn't hurting.

The downside was the fact the my board was version 1.0 and required a little fix to 5V supply and the fact that smoothie developers aren't exactly happy about the clone, as I've later found out.

But as I said, the board had been on the bottom of my drawer for half a year maybe before now putting it on, and first impressions are good. Lots of options to do stuff in addition to extended functionality. I like.

And I love messing with new stuff mainly !! :-D

Hear hear! Good luck!

Very informative and well detailed post right there, I'm standing up, and applause silently !!

I personnaly went for the smoothie little sister, the azteeg x5 mini, which is just what I need for the SD3. It gives no possibilities to upgrade the machine though, no 2nd extruder possible.

If you ask me why I did that, well... It's a hard question to answer... Maybe I just wanted to upgrade my sanguino, and the vikki 2 screen is a real plus.. And I love messing with new stuff mainly !! :-D

Smoothie Pros:
* Firmware current control of steppers (this actually makes them more difficult to tune by ear; I like using the trimpot!)
* Ethernet interface (just a serial-over-ip interface; the web thing is no replacement for octoprint)
* Configuration is flexible and clear, this and firmware itself (!!) can be updated on the fly via USB/SD card. No arduino IDE / flashing procedure (this is tops)
* Nearly every control pin is broken out to a header
* SD card is provided/integrated/necessary, works straight out of the box (supposedly the most reliable way to print from this board)
* Extra grunt reduces chances of non-time-critical things like LCD getting in the way of stepper/PID control (these issues pop up every now and then on 8-bit boards).

Rumba Pros:
* More logical board layout (slightly, definitely a subjective opinion!)
* Removable pseudo-standard driver modules
* Cheaper (just)
* Marlin firmware is more 3d-printer-specific, and development is livelier, so new/fun features likely to be supported earlier

In summary:

I believe the advantages of the smoothie's performance (32-bit 96MHz ARM) currently do not translate into real value gain for a 3d printer - even though the extra grunt could in theory support better temperature, motion and extrusion control than an 8-bit uC could (16MHz 8-bit AVR), which was my main hope in trying it out, in practice the firmware basically implements the core functions similarly to Marlin (GRBL-based motion control, vanilla PID for temperature control, no (working) extrusion advance). The only present advantage is the ability to run higher combined steprates (e.g. for running a couple of 256x ustep external drivers, for example) - which is usually a little outside typical needs.

Get whichever you can afford and like the look of, as either will likely serve you well. But hoping that the smoothie firmware starts taking advantage of the hardware is a gamble that has kept a fair few people waiting for a couple of years already!

Just gave the Smoothie a look, and while they certainly packed a lot of features into one board, I can't see paying for something I probably won't use 1/2 the functionality of.

Then there's the integrated stepper drivers.  Not a big fan of that.  These are 4982 steppers, and they charge basically $20 a pop.  That's pretty steep.

I don't know.  You can do better for that money I think.

I agree.  Disappointed that there isn't a Smoothie option to use standard stepper modules.  However, there is are 4 pin connectors on the board with the controls for each driver enabling reasonably easy use of external drivers.  X&Y are really the only channels I want/need to upgrade so I am thinking once I start pushing the speed of my printer, I will replace those with external drivers on a separate breakout board.  It is a little more expensive, but I expect to spend a LOT of time on this project so I'm not going to let $50 get in the way of it being what I want.

Thanks guys!

Just gave the Smoothie a look, and while they certainly packed a lot of features into one board, I can't see paying for something I probably won't use 1/2 the functionality of.

Then there's the integrated stepper drivers.  Not a big fan of that.  These are 4982 steppers, and they charge basically $20 a pop.  That's pretty steep.

I don't know.  You can do better for that money I think.

I completely agree, as far as usable features I still like the Rumba the best.  I have 3 Rumba's and have never had a problem with any of them.  I know some users have had issues with the 16U2 chip (USB).  I have even had a few of those users ship their boards to me.  Thus far I have been able to recover all but one.  That particular board belonged to AZERATE and the first time I fixed it and sent it back working.  Within a few days it failed again and the second time around I wasn't so lucky.  I tend to lean towards operator error more so than the chip being faulty. 
The point is for exactly all the reasons pirvan stated you can get many more useable features and a board with replaceable drivers (a huge plus) for the same or less money.

Then there's the integrated stepper drivers.  Not a big fan of that.  These are 4982 steppers, and they charge basically $20 a pop.  That's pretty steep.

I don't know.  You can do better for that money I think.

http://shop.uberclock.com/collections/smoothie

This craftware thing sounds like it's worth a look... Is the config easy enough? Would love to see whether it really is generating G3/4 commands in the output... "Vector" sounds a bit vague and internal, not sure it means it's outputting arcs. If the input is STL, then the curves are already segmented: is it actually detecting the curves based on the segments? That would be quite an achievement!

The main reason something like a smoothieboard will do arcs well is both the gruntier ARM processor to calculate the movements, and that it's aimed at CNC routing applications too, which usually do use arcs, so the user base will be regularly using and testing G3/G4 movements. For Marlin, which is basically only aimed at 3D printing and g-code from STL slicers, the G3/G4 commands might work but won't get nearly the same amount of testing as nobody will be actively using them!

According to the Craftware change-log, the software is now vector driven by default, and I don't think they would have made that change if the software wouldn't have worked on the majority of printers on the market.

I too have a RAMPS, and I'm pretty sure it will work just fine.

Downside is it won't support 24V so heating a larger than 8 x 8 heat bed directly requires an SSR...and it certainly won't support the G2/G3 arcs as mentioned above.

I hear the RUMBA is a great board as well, if you can get your hands on one...not sure if it incorporates an ARM processor.