Impressions: Duet vs. Smoothie

User avatar
626Pilot
ULTIMATE 3D JEDDI
Posts: 1716
Joined: Tue May 14, 2013 12:52 pm

Impressions: Duet vs. Smoothie

Postby 626Pilot » Sat Oct 24, 2015 1:28 am

This is a comparison between the Smoothieboard and the Duet 0.8.5. I haven't used the Duet 0.6 or any of the other iterations.

Hardware Capabilities
  • The Duet has 96KB RAM, runs at 84MHz, and has five stepper axes (expandable to 9). You can drive up to six extruders.
  • Smoothie has 64KB RAM, runs at 120MHz, and has five stepper axes. There are no official or third-party expansion boards, so you're limited to two extruders max.
  • Both boards have USB and Ethernet connectivity. Smoothie allows you to mount the SD card over USB, but transfer is slow, and an incomplete filesystem implementation means that you have to handle uploads in a very specific way in order to avoid corrupting the file allocation table. The Duet has no support for mounting the SD card over USB. However, you can upload files over the web interface, or via FTP.
  • Both boards let you set stepper current in software, so you don't have to twiddle a jewler's screwdriver in a micropot.

Supported Robots and Kinematics
  • Smoothie supports more 3D printer varieties ("classic" overhead-rotational delta, HBot, "rotational" Cartesian, and SCARA). It also has support for laser cutters, and vinyl cutters are planned as well. Presumably, if you're here, you have a linear delta (Rostock/Kossel-type printer), so either board can work for your hardware.
  • If you want to drive a big CNC machine with its own high-torque stepper drivers, Smoothie has pin connectors that you can use to bypass the built-in drivers, so you can use Gecko drives or whatever you want. Duet allows you to do this as well, but you have to remap the pins to the expansion header. (I previously stated that it wasn't supported at all.)
  • Duet does delta kinematics calculations on the fly, which is the best way to do it. Smoothie, like Repetier and Marlin, uses tiny segments that are re-calculated hundreds of times per second. This is slightly less accurate. I haven't seen any side-by-sides, and you can probably wring more accuracy out of Smoothie by increasing the frequency of segment generation, so I have no opinion on how much it matters.

Acquisition and and Integration
  • You can get the Duet 0.8.5, DueX4, and PanelDue from a single vendor in the UK that will ship to the US reasonably quick. I got mine in about a week. Since all three parts can be had from the same vendor, I consider this a fully-integrated hardware solution.
  • You can get a Smoothieboard - or the Azteeg X5, a clone - from vendors in the US and UK. There is no expansion board. You can use a variety of 3rd-party displays. There is a smaller, cheaper clone called the Azteeg X5 Mini. I like the FETs on the Smoothieboard better.

Documentation
  • The Duet documentation is scattered between several different places (T3P3's site, the RepRap wiki, and various forum/usenet postings), and the main commissioning document assumes you're running an Ormerod (Cartesian) printer, which is a little confusing. You have to figure out what applies to you, and what doesn't, because it isn't generalized. It took me awhile to find a wiring diagram for the board - for some reason, it isn't included in the commissioning document. Fortunately, this forum is full of geniuses, and our own mhackney has drawn up a very good wiring diagram. For delta users, this page is required reading. You should be using the dc42 fork, which has the best support for linear delta (Rostock/Kossel-type) printers, as well as a working auto-calibration system.
  • The Smoothie documentation is totally excellent. It's all in one place on their site, which itself is a wiki - a very smart decision, I think. They have step-by-step instructions for doing all kinds of things, and I never had trouble finding a wiring diagram. You should be using my fork (link in sig), which has the auto-calibration system.

Wiring
  • The Duet has fewer wires to connect because it doesn't have as fancy a power distribution system. Smoothie requires three DC power connections.
  • I feel the connectors are better on the Smoothieboard. Duet's hot end terminals in particular are problematic - they're tiny, difficult to cram a 16-gauge wire into, and the act of tightening down the terminal screw torques the entire connector block so you have to worry about it messing up your board. They should've used beefier connectors for that, full stop.
  • The Duet lines up the stepper and endstop pins for each axis on a single 7-pin header. Smoothie has stepper drivers at the bottom, and endstops at the top.
  • Smoothie's thermistor inputs are on the right side. The Duet's thermistor inputs are near the middle of the board. Smoothie has 4 thermistor inputs. Duet has 3.
  • The Duet has two PWM ports (for driving fans, LEDs, etc.) and two always-on 12V ports, both near the middle of the board. Smoothie has about the same setup, except that there are no always-on ports. For that, you'd need to connect to one of the numerous 12V wires coming out of the power supply. Smoothie's PWM ports are along the top, next to the power inputs. The separate inputs on Smoothie provide cleaner power distribution, but I don't know how much it matters.
  • Both boards can run your hot end and bed heat at either 12 or 24 volts.
  • Smoothie has a slot for a 5V linear regulator, which you can either solder in or not. I like running it with the regulator soldered in. Duet ships with the regulator already installed. I don't see much point in leaving that out. Both boards have to be powered down by turning off the switch, AND unplugging the USB cable.
  • Smoothie's endstop connectors are 5V tolerant. Duet's are not. If your hot end is wired to +5V, like mine was, you'll want to cut that and take power directly from the +3.3V pin on E0's endstop connector. If you run 5V through the endstop pin on a Duet, you will probably kill your board.

Panels
  • The Duet 0.8.5 has a native port for the PanelDue, a beautiful and highly functional touchscreen controller. It has its own Arduino-type controller onboard. Good: It doesn't use any of the Duet's RAM for menu strings, leaving loads of space available to do other things. Annoying: Coders have to program it separately, and users have to flash updates separately, adding yet another thing you have to do when you update. In the future, perhaps someone will add a scripting engine that would let you bind menu screens with menu items and G-code commands in a file on the SD card. (It does let you run macros off the SD card, but that's as much customization as you get.)
  • Smoothie supports LOTS of different panels. I don't think it supports the PanelDue yet, but that is supposedly in the queue.
  • The PanelDue will let you home the printer and move the effector around. If you have the right firmware, it will also run the calibration for you - but it won't save it to your config file, because the Duet's M500 command doesn't save delta parameters. (I previously stated it doesn't save any parameters, but it does save everything else.)
  • I never used a panel with Smoothie, but it does appear to have a "jog" function for moving the effector. The menus are nicer on the PanelDue.

Web interfaces and file transfers
  • The Duet allows file uploading over FTP, as well as the web interface, although there is a bug that will prevent any file under ~1.5K from being uploaded over the web (which was discovered by our own mhackney). dc42 is aware of this, and will probably release a fix for it soon. The web interface is really, really nice, and even more functional than the panel.
  • Duet's web interface lacks the up-arrow command history you get with Repetier Host.
  • Smoothie's web interface is plain by comparison. The Duet stores all web interface files on the SD card, so it can potentially have a much more complex interface - Smoothie's web assets either have to be served remotely, or be pulled out of the controller's built-in flash, which has limited (and ever-shrinking) space.
  • Smoothie allows you to edit the config files and upload G-code directly to the SD card by mounting it over USB. This is very convenient. However, transfer speeds are slow, and you have to be careful about how you do it. The file system implementation is incomplete, so if you upload/edit stuff over USB and then type M500 ("save config settings"), it will corrupt your files. You have to edit over USB, safely unmount, power cycle, edit your settings (or run the auto-calibration), and then do M500.
  • Duet doesn't allow you to mount the SD card over USB at all, which is an inconvenience. However, you can edit over FTP, or upload over the web interface. According to dc42, xfer speeds over Ethernet are typically around 200kb/sec, totally blowing Smoothie's ISDN-like transfer speeds out of the water. However, on my board, I'm getting 30KB/sec - not at all faster than Smoothie, I'm afraid. I don't buy the argument that "some SD cards are faster than others" because if I take that SD card out of the Duet and put it into a cheap SD card reader, I can drag a ten-meg file to it, and it'll finish writing in a fraction of a second. I honestly think this is down to the device driver on the Duet not being particularly fast.

Upgrading the Firmware
  • Smoothie's firmware is "monolithic." You compile it and upload it to the board in a single step. The entire process takes about 20 seconds, and as long as you don't get an error message, it's impossible to screw up.
  • Duet's firmware is split between the actual firmware, and various files needed to draw the web interface and so on. If you upgrade the firmware image, you also have to copy over the new files. This is an extra step, and it can be a little confusing, since the interface files are distributed in folders that have the names of printers on them - none of them ours. However, because the assets for the web interface and so on are stored on the SD card, precious on-board flash memory (on the controller, not the SD card) is preserved.

Config files
  • Smoothie uses a file called config to store the main configuration. This is a really nice system that uses human-understandable variable names. Anything that needs to be changed is saved to a file called config-override, which contains G-codes. To save your running config, you type M500, and it saves config-override. The next time the board is powered up, your changes are still there.
  • Duet uses a file called config.g. It has no human-understandable variable names, but it does have comments. There is no override file, and when you type M500 to save your config, it saves almost everything to EEPROM. It doesn't actually change the values in the file. Delta parameters are not saved, although dc42 tells me he may add that soon. He also tells me that the config gets wiped every time you update the firmware because the EEPROM area containing those values is not preserved. This is another area of weakness. It shouldn't be bothering with memory that gets wiped every time there's an update. It should write overrides to a file on the SD card instead.
  • One bonus: You can edit config.g over FTP and press the E-stop button on the web interface, and it'll reload the config. With Smoothie, you have to reboot the board, losing your connection in the process.

Calibration
  • Both the dc42 fork of RepRapFirmware, and my fork of Smoothie, have functional delta calibration systems. You can learn about mine from the link in my sig. Mine corrects more variables, and it also features depth map-based Z correction for whatever Z error wasn't corrected in the calibration. If I think it's necessary, I will port some of my work over to RepRapFirmware. At this time, I'm able to get full-bed prints with my fork, but dc42's calibration has no Z correction, and I can't get full-bed prints with my Duet.
  • Again, M500 saves the config (including calibration results) on Smoothie, and does nothing on Duet - you have to manually copy and paste g-code into config.g.
  • The Duet calibration relies on a generated g-code file to tell it where all the probe points are. This is a very rigid way of doing it because if you don't want to use one of the two supplied probing radii, you have to figure out how to generate a new file yourself. This may seem like a minor issue, but I've seen significant changes in calibration quality from only a 10mm difference. It's nicer when you can give it any arbitrary number, which my firmware does allow. It takes the probing radius and generates the points itself.

Running
  • I haven't been able to get my Duet to extrude yet. It has some weird thing about T0/T1 commands. I'll figure it out soon, I think. It has been a nuisance I never had to deal with on Smoothie.
  • The default PID parameters in RepRapFirmware seem to be fairly good for the heated bed and my hot end, which is an E3D Cyclops. I don't know how it'd hold up under actual printing conditions because it won't extrude yet. Smoothie's PID autotune is sort of good-ish, but I got better results tuning by hand using this.
  • I haven't hooked up the expansion board yet, but I'm looking forward to it!

Development
Smoothie's firmware is much more well-developed, and is very modular. You can, for example, have two calibration strategies that do the same thing in different ways side-by-side, and switch off between them by changing a value in config. RepRapFirmware, which runs the Duet, is not modular at all. Smoothie uses and is a great example of loose coupling. RepRapFirmware has a lot of tightly-coupled code, all intertwined and spaghetti-like, probably because it evolved in the awful Arduino environment. Unfortunately, Smoothie wasn't available at the time of the Duet's original release, and they wound up basing their firmware on Marlin, which IMO is worse than Repetier firmware and doesn't hold a candle to Smoothie.

Why I left Smoothie for the Duet
It has more RAM, and drives more axes. I hated all the time I had to spend optimizing RAM use on Smoothieware. It cost me days of productivity. Duet isn't much better, but it arguably has more breathing room - 96K instead of 64K. I need a minimum of three extruders for my application, so Smoothie is out. Next year, they'll have a better controller with exponentially more RAM - hundreds of megabytes, not tens of kilobytes - that will leave the Duet in the dust. I just don't know whether there will be a nice 3+ extruder option for it. In the meantime, Duet has what I need. If I'm sufficiently impressed by the Smoothieboard 2, I will be going back. The firmware is FAR easier to wrench on.
Last edited by 626Pilot on Fri Nov 13, 2015 5:26 pm, edited 7 times in total.

User avatar
bot
Printmaster!
Posts: 993
Joined: Thu Sep 25, 2014 12:18 am
Location: Vancouver
Contact:

Re: Impressions: Duet vs. Smoothie

Postby bot » Sat Oct 24, 2015 2:22 am

A very nice write-up. The PanelDue does indeed have jog functionality, by touching/clicking on the axis you wish to move, you bring up an adjustment menu. It's not a nice arrow keypad or anything, but it works fine.
*not actually a robot

critical_limit
Printmaster!
Posts: 142
Joined: Thu Jul 10, 2014 11:15 am
Location: Berlin Germany

Re: Impressions: Duet vs. Smoothie

Postby critical_limit » Sat Oct 24, 2015 3:53 am

@626Pilot:

Are you planing to port your callibration routine to the Duet board? This would be great. Love your callibration on the smoothie.

User avatar
626Pilot
ULTIMATE 3D JEDDI
Posts: 1716
Joined: Tue May 14, 2013 12:52 pm

Re: Impressions: Duet vs. Smoothie

Postby 626Pilot » Sat Oct 24, 2015 6:34 am

bot wrote:A very nice write-up. The PanelDue does indeed have jog functionality, by touching/clicking on the axis you wish to move, you bring up an adjustment menu. It's not a nice arrow keypad or anything, but it works fine.

Fixed.

critical_limit wrote:Are you planing to port your callibration routine to the Duet board? This would be great. Love your callibration on the smoothie.

It already has a calibration routine. I might extend it to mess with more variables, as mine does. I don't know whether its results are better than mine yet. I'd rather teach the existing system to solve for more variables than gut it and replace it with my parallel simulated annealing algorithm. That would be a LOT of work.

User avatar
mhackney
ULTIMATE 3D JEDDI
Posts: 5402
Joined: Mon Mar 26, 2012 4:15 pm
Location: MA, USA
Contact:

Re: Impressions: Duet vs. Smoothie

Postby mhackney » Sat Oct 24, 2015 9:05 am

Excellent write up and overview, saves me a lot of time not duplicating your work.

On the web upload bug - that appears to be a problem with a specific browser/version with Safari on OS X. I've sent a network trace to the web interface developer (well, David sent the one I sent him) so they are investigating. But it works fine with Firefox and Chrome on all platforms I've tried (windows, linux, OS X).

I've found that configuration changes are a lot faster process with Duet. The upload speed either over the web interface or FTP is faster also as is the process to make changes. I really like that all config is done as gcode in Duet, very simple and consistent. M500 s a bit of a nuisance but should be easily added in future versions of the firmware I would think. The other thing I like about Duet is it separates out files and macros into separate .g files. For instance, homing is down by homedelta.g and you can do basically whatever you want to get the homing behavior you need. Same with probing, pausing, resuming, etc. This is really powerful.

On the probing radii, I've written a python script that creates the necessary gcode for the probing file. I just need to clean up the interface and then I'll post it. I started with a spreadsheet to do this too.

Sublime Layers - my blog on Musings and Experiments in 3D Printing Technology and Art

Start Here:
A Strategy for Successful (and Great) Prints

Strategies for Resolving Print Artifacts

The Eclectic Angler

WZ9V
Printmaster!
Posts: 65
Joined: Sat Jan 25, 2014 8:49 pm

Re: Impressions: Duet vs. Smoothie

Postby WZ9V » Sat Oct 24, 2015 9:44 am

Nice write-up.

I like the ease of use implied by the many posts about the RepRap firmware but I'm thinking something more like the Replicape on Kickstarter has more long term potential. You should check it out, I think they have an expansion card in the works also. With the new pricing it's only about 20/30 more total than the Duet 0.8.5 solution. Their firmware is obviously not as mature but it's mostly in Python making it pretty easy to update.

User avatar
mhackney
ULTIMATE 3D JEDDI
Posts: 5402
Joined: Mon Mar 26, 2012 4:15 pm
Location: MA, USA
Contact:

Re: Impressions: Duet vs. Smoothie

Postby mhackney » Sat Oct 24, 2015 12:25 pm

That's just the problem - there is always something better/faster/cheaper coming. In my case, I moved to Smoothie and Duet because they are here now, when I need them. Next year, I'll try something new if it makes sense.

One comment though - firmwares written in compiled languages are time-tested. As much as I love Python it is really unproven as an embedded system language. Delta kinematics are taxing on processing power. Not to say that this might just be the bees knees, just saying that - like people from Missouri "show me"!

Sublime Layers - my blog on Musings and Experiments in 3D Printing Technology and Art

Start Here:
A Strategy for Successful (and Great) Prints

Strategies for Resolving Print Artifacts

The Eclectic Angler

User avatar
bot
Printmaster!
Posts: 993
Joined: Thu Sep 25, 2014 12:18 am
Location: Vancouver
Contact:

Re: Impressions: Duet vs. Smoothie

Postby bot » Sat Oct 24, 2015 1:07 pm

mhackney wrote:That's just the problem - there is always something better/faster/cheaper coming. In my case, I moved to Smoothie and Duet because they are here now, when I need them. Next year, I'll try something new if it makes sense.

One comment though - firmwares written in compiled languages are time-tested. As much as I love Python it is really unproven as an embedded system language. Delta kinematics are taxing on processing power. Not to say that this might just be the bees knees, just saying that - like people from Missouri "show me"!


Exactly! Smoothie was the big thing last year, this year it will be Duet, next year Smoothie 2.
*not actually a robot

User avatar
mhackney
ULTIMATE 3D JEDDI
Posts: 5402
Joined: Mon Mar 26, 2012 4:15 pm
Location: MA, USA
Contact:

Re: Impressions: Duet vs. Smoothie

Postby mhackney » Sat Oct 24, 2015 1:14 pm

or....

Sublime Layers - my blog on Musings and Experiments in 3D Printing Technology and Art

Start Here:
A Strategy for Successful (and Great) Prints

Strategies for Resolving Print Artifacts

The Eclectic Angler

User avatar
bot
Printmaster!
Posts: 993
Joined: Thu Sep 25, 2014 12:18 am
Location: Vancouver
Contact:

Re: Impressions: Duet vs. Smoothie

Postby bot » Sat Oct 24, 2015 1:58 pm

BBB, or Alligator, or...
*not actually a robot

WZ9V
Printmaster!
Posts: 65
Joined: Sat Jan 25, 2014 8:49 pm

Re: Impressions: Duet vs. Smoothie

Postby WZ9V » Sat Oct 24, 2015 4:38 pm

I got in on one of the Replicape model B in Kickstarter. Will be interesting based on what I've seen in the github for the Redeem Daemon. If the original Python based system is a problem it might be usable with Machine KIt. I saw someone got the older board working with that. After trying 626Pilot's Smoothie fork I suspect I would have problems with Duet also. It seems my FSR setup has issues with repeatable measurements.

I swapped in one of the IR probes and it did better but with the layout on my V-Slot 250 its offset quite a bit from the nozzle and I need to figure what to do to keep the probing on the bed. Most likely reduce the delta radius.

DSteck
Printmaster!
Posts: 39
Joined: Sat Sep 05, 2015 8:34 pm

Re: Impressions: Duet vs. Smoothie

Postby DSteck » Sat Oct 24, 2015 8:35 pm

So here's a dumb question... Can you still run a printer on a Duet through Repetier?

I've been looking into switching to a Duet in the hopes that print quality/accuracy would drastically improve (single extruder). However, I don't have the time to invest into trying to muddle through firmware and get it running because I'm about to dive into reverse engineering all the CAN assignments in a vehicle platform. :shock:

For 626pilot, is there a plan to release an idiot proof setup guide with an included firmware file for the Duet?

User avatar
teoman
ULTIMATE 3D JEDDI
Posts: 1783
Joined: Sat May 24, 2014 5:43 pm

Re: Impressions: Duet vs. Smoothie

Postby teoman » Sat Oct 24, 2015 9:01 pm

Which vehicle?

I am also diving in to that (hobby project).
When on mobile I am brief and may be perceived as an arsl.

User avatar
626Pilot
ULTIMATE 3D JEDDI
Posts: 1716
Joined: Tue May 14, 2013 12:52 pm

Re: Impressions: Duet vs. Smoothie

Postby 626Pilot » Sat Oct 24, 2015 9:36 pm

WZ9V wrote:I swapped in one of the IR probes and it did better but with the layout on my V-Slot 250 its offset quite a bit from the nozzle and I need to figure what to do to keep the probing on the bed. Most likely reduce the delta radius.

I advise strongly against calibrating a delta printer with an offset probe. You will NEVER get the same quality of results as you would with the probe at effector center. Cartesian machines can get away with that because their coordinate system doesn't have non-linear characteristics, but delta math is very weird and it will throw off the results.

DSteck wrote:For 626pilot, is there a plan to release an idiot proof setup guide with an included firmware file for the Duet?

I have forked dc42's repo. I would like to have firmware and configs specifically for our printers. If I do any work on his auto-calibration routine, I'll publish it there. Right now, I haven't pushed any changes at all.

WZ9V
Printmaster!
Posts: 65
Joined: Sat Jan 25, 2014 8:49 pm

Re: Impressions: Duet vs. Smoothie

Postby WZ9V » Sat Oct 24, 2015 9:57 pm

Unfortunately the V-Slot 250 requires an offset mounting. It has a PG35 direct drive extruder along with a J-head and part fan hanging below the effector. I might be able to move the sensor closer I'll never be able to get it real close. My new project will be better in that regard since I'll be using David Crocker's E3D fan mount. My FSR setup had no offset but kept failing your probe test G29 E, at least the IR probe gives me good results with the G29 E. I'm convinced that a FSR setup really needs to be built into the effector instead of below the bed.

626Pilot wrote:I advise strongly against calibrating a delta printer with an offset probe. You will NEVER get the same quality of results as you would with the probe at effector center. Cartesian machines can get away with that because their coordinate system doesn't have non-linear characteristics, but delta math is very weird and it will throw off the results.

DSteck
Printmaster!
Posts: 39
Joined: Sat Sep 05, 2015 8:34 pm

Re: Impressions: Duet vs. Smoothie

Postby DSteck » Sat Oct 24, 2015 10:59 pm

teoman wrote:Which vehicle?

I am also diving in to that (hobby project).

It's for making a Motec system compatible in 2006-2013 Corvettes as a complete ECM replacement that will still communicate to the BCM, ABS, and TCM.

Koenig
Noob
Posts: 1
Joined: Sun Oct 25, 2015 5:54 am

Re: Impressions: Duet vs. Smoothie

Postby Koenig » Sun Oct 25, 2015 10:00 am

Nice writeup!

Thought I could comment on a few things.

If you want to drive a big CNC machine with its own high-torque stepper drivers, Smoothie has pin connectors that you can use to bypass the built-in drivers, so you can use Gecko drives or whatever you want. Duet lacks this feature - but again, if you're here, you probably only care about running a low-current delta printer such as the Rostock MAX.


You can do the same with the duet using the expansion header and then use an M-code (don't remember which) to relocate the desired axis there.

You can get the Duet 0.8.5, DueX4, and PanelDue from a single vendor in the UK that will ship to the US reasonably quick. I got mine in about a week. Since all three parts can be had from the same vendor, I consider this a fully-integrated hardware solution.


The duet 0.6 is also available from two other vendors, replikeo and reprappro and is still a good option if you are planning to run one or three to five extruders, the 0.85 is the better choice if you plan to run two or six extruders.

Duet doesn't allow you to mount the SD card over USB at all, which is an inconvenience. However, you can edit over FTP, or upload over the web interface. I don't know whether it's faster than USB transfers to Smoothie. Someone else can chime in on that.


I can only chime in about the speed of the duet.
Seems most user get upload speeds around 200-300kB/s when uploading via the WEB-interface, I read someone getting much higher speeds but he used an "overkill" SD-card if I remember correctly.
Myself I get ~240kB/s when uploading via the WEB-interface, haven't tried the ftp-function as of yet.

User avatar
mhackney
ULTIMATE 3D JEDDI
Posts: 5402
Joined: Mon Mar 26, 2012 4:15 pm
Location: MA, USA
Contact:

Re: Impressions: Duet vs. Smoothie

Postby mhackney » Sun Oct 25, 2015 10:22 am

I didn't mention the reprappro version because it has been out of stock for ages AND it was a different version. Note that they don't show the version number, I really it was a predecessor to the v0.6. David knows the details but there is a lot of info on the reprapforums also. Note that the reprappro Duet Shield is different also and I don't know how easy/difficult it would be to interface it.

Sublime Layers - my blog on Musings and Experiments in 3D Printing Technology and Art

Start Here:
A Strategy for Successful (and Great) Prints

Strategies for Resolving Print Artifacts

The Eclectic Angler

User avatar
mhackney
ULTIMATE 3D JEDDI
Posts: 5402
Joined: Mon Mar 26, 2012 4:15 pm
Location: MA, USA
Contact:

Re: Impressions: Duet vs. Smoothie

Postby mhackney » Sun Oct 25, 2015 10:29 am

By the way, I did all my initial Duet work on a Kossel250 but I built mine from components so I could use a Bowden system on it. I use FSRs under the bed with an E3D v6 hotend. There is a lot of FUD out there on FSRs. I have them installed on 4 deltas and now a TAZ Cartesian printer and I've had nothing but good experience with them. I have heated beds on all of these printers and my mounting system (I've posted a few versions on this forum) insulates the FSRs sufficiently that heat is not an issue. The issues I hear about reproducibility and/or inaccurate Z offset - in my experience - can be traced directly to the mechanical setup of your printer and probing system. I have no doubt that an unsupported 9" diameter glass plate floating on 3 perimeter points is going to flex slightly if you ram the nozzle into it at too high speed. I have snowflake backing plates on all my printers except the TAZ and see no flex and have excellent reproducibility. It can be done, just investigate for the source of issues and fix them.

I can mention this since it is public on his Facebook page, but John Oly is experimenting with an effector mounted accelerometer to detect probe contact with the plate. This will eliminate the minimal complexity of the FSR system and allow a rigid bed mount. Stay tuned...

Sublime Layers - my blog on Musings and Experiments in 3D Printing Technology and Art

Start Here:
A Strategy for Successful (and Great) Prints

Strategies for Resolving Print Artifacts

The Eclectic Angler

User avatar
mhackney
ULTIMATE 3D JEDDI
Posts: 5402
Joined: Mon Mar 26, 2012 4:15 pm
Location: MA, USA
Contact:

Re: Impressions: Duet vs. Smoothie

Postby mhackney » Sun Oct 25, 2015 11:48 am

Pilot626, have you set up a dev environment yet? The one thing I have to say is that developing on Smoothie was painlessly easy. It built a complete dev environment for any platform. I was able to start work on my TemperatureSwitch module in less than 15 minutes, no joking. So far, I've spent 2 hours trying to get a dev environment. First on Mac and then on Windows.

The current version of Eclipse is Mars and that's what I've been using for other projects. The firmware build instructions are based off the earlier Juno. I really didn't want to have to install an older version of Eclipse but I can't get the Arduino tools setup for Mars yet. Probably another few hours and I'll figure it out.

Sublime Layers - my blog on Musings and Experiments in 3D Printing Technology and Art

Start Here:
A Strategy for Successful (and Great) Prints

Strategies for Resolving Print Artifacts

The Eclectic Angler

User avatar
626Pilot
ULTIMATE 3D JEDDI
Posts: 1716
Joined: Tue May 14, 2013 12:52 pm

Re: Impressions: Duet vs. Smoothie

Postby 626Pilot » Tue Oct 27, 2015 1:20 am

mhackney wrote:Pilot626, have you set up a dev environment yet? The one thing I have to say is that developing on Smoothie was painlessly easy. It built a complete dev environment for any platform. I was able to start work on my TemperatureSwitch module in less than 15 minutes, no joking. So far, I've spent 2 hours trying to get a dev environment. First on Mac and then on Windows.

The current version of Eclipse is Mars and that's what I've been using for other projects. The firmware build instructions are based off the earlier Juno. I really didn't want to have to install an older version of Eclipse but I can't get the Arduino tools setup for Mars yet. Probably another few hours and I'll figure it out.

I haven't.

My patience for this firmware is already running thin. It doesn't extrude, even if T0 is selected. That, and the fact that M500 doesn't save delta parameters, is irritating the freaking heck out of me. Smoothie never failed to extrude, nor did it make me edit the config file in a text editor because M500 doesn't work right.

I'm starting to think I'll get this running Just Good Enough to pump three extruders, which is, at present, the one and only reason I'm using this and not Smoothie. If this firmware continues to annoy me, I will just bite the bullet and graft a StepStick onto a Smoothieboard with some ass-ugly jumper wires. The "culture shock" is making me want to bail out. I've already expended ridiculous amounts of effort getting a good calibration system on Smoothie. Having to jump through all these hoops for a new platform is making me balk. I have to force myself to keep going with this, and that's not a good sign. That usually happens shortly before I totally run out of self control, at which point I have no motivation left.

Smoothie is deffo easy to develop for. You run a script to download the compiler, another to give you the build environment, and then you just "make" and then "make upload" and that's it. Easy.

User avatar
bot
Printmaster!
Posts: 993
Joined: Thu Sep 25, 2014 12:18 am
Location: Vancouver
Contact:

Re: Impressions: Duet vs. Smoothie

Postby bot » Tue Oct 27, 2015 1:40 am

If you think this is bad, try running a Dynomotion Kflop+Kstep as the controller for a 3d printer. If there was a good firmware for that hardware, it would be THE goto solution for professional delta robots.
*not actually a robot

plasma
Printmaster!
Posts: 61
Joined: Tue Dec 30, 2014 2:41 am

Re: Impressions: Duet vs. Smoothie

Postby plasma » Tue Oct 27, 2015 10:24 am

IMO Smoothieboard is the best and easiest solution at this time. Great hardware with poorly supported software totally negates the benefits from hardware. Smoothieboard offers less stellar hardware then Duet with a much better software integration and support. Its a win for me until the Smothieboard 2 comes out.

WZ9V
Printmaster!
Posts: 65
Joined: Sat Jan 25, 2014 8:49 pm

Re: Impressions: Duet vs. Smoothie

Postby WZ9V » Tue Oct 27, 2015 11:59 am

Your mounting system would fix my FSR woes. My problem is that when my bed is loose enough to get consistent triggering it is too loose to not flex. If I did something like what you did I could solve that. For now I'm just moving to a removable probe on the effector and hard mounting the bed on the existing three point mount. I'm attracted to DC42's fast calibration but I'm not sure it's worth the investment in new hardware but it is tempting. For now I need to sort out my z-probe issues because without that no firmware is going to work correctly.

In short the problem is not FSR's in general it is my printer in particular (in case someone was confused). That's said I do prefer an effector mounted solution to probing such as DC's IR probe.

mhackney wrote:By the way, I did all my initial Duet work on a Kossel250 but I built mine from components so I could use a Bowden system on it. I use FSRs under the bed with an E3D v6 hotend. There is a lot of FUD out there on FSRs. I have them installed on 4 deltas and now a TAZ Cartesian printer and I've had nothing but good experience with them. I have heated beds on all of these printers and my mounting system (I've posted a few versions on this forum) insulates the FSRs sufficiently that heat is not an issue. The issues I hear about reproducibility and/or inaccurate Z offset - in my experience - can be traced directly to the mechanical setup of your printer and probing system. I have no doubt that an unsupported 9" diameter glass plate floating on 3 perimeter points is going to flex slightly if you ram the nozzle into it at too high speed. I have snowflake backing plates on all my printers except the TAZ and see no flex and have excellent reproducibility. It can be done, just investigate for the source of issues and fix them.

I can mention this since it is public on his Facebook page, but John Oly is experimenting with an effector mounted accelerometer to detect probe contact with the plate. This will eliminate the minimal complexity of the FSR system and allow a rigid bed mount. Stay tuned...

WZ9V
Printmaster!
Posts: 65
Joined: Sat Jan 25, 2014 8:49 pm

Re: Impressions: Duet vs. Smoothie

Postby WZ9V » Tue Oct 27, 2015 12:03 pm

Totally agree. A Smoothie port to Duet would be a wonderful thing if it could be done. I suspect they are just too different to make it worthwhile with other alternatives on the horizon. It would be cool to see DC42's calibration in Smoothie.

plasma wrote:IMO Smoothieboard is the best and easiest solution at this time. Great hardware with poorly supported software totally negates the benefits from hardware. Smoothieboard offers less stellar hardware then Duet with a much better software integration and support. Its a win for me until the Smothieboard 2 comes out.


Return to “Duet”

Who is online

Users browsing this forum: No registered users and 1 guest