We are going to start filming our Smoothie v Arduino printer comparison today.

We are going to start filming our Smoothie v Arduino printer comparison today.
We have 4 printers (2 deltas and 2 cartesians), one of each on arduino electronics and one of each on boards running smoothie.

The first comparison will be a side-by-side time lapse printing the same gcode (some start gcode will be different for the different leveling options, but the actual printed part will be identical gcode).

Other things we plan on showing:

  • Differences in seams & retraction
  • Differences in firmware modifications
  • Differences in acceleration over long distances (both in straight lines, and on parts with curves and complex geometry)

What else would you like to see (if possible)?

We haven’t decided on all the models yet, but a simple 50 x 50 calibration cube will be printed, as well as Morena’s ubiquitous Treefrog.

The idea is to print things that people are familiar with and can thus focus on the specifics of the results rather than details of the design itself.

We’ll likely be printing the Make torture test as well, but feedback/suggestions are more than welcome.

I’m not sure what we’re going to end up on as far as settings go. We haven’t decided if we want to show “ideal” settings for each machine, or show them configured the same based on the manufacturer’s settings.

We already know that we can get significantly better results from Arduino boards by tuning the acceleration and jerk (typically way down from most mfgs). We also want to avoid the conversation that starts with, “Yeah, but they weren’t configured the same…”

For the time lapse, they will be configured as identically as possible to show the difference in actual printing speeds.

The rest is still up in the air.

Sorry, I see I wasn’t clear. We definitely will have at least 2 “showdown” prints that will have the machines printing their best (or our best) - running different accel, jerk, and other settings.

The main reason we want to show some prints side-by-side running the same settings is because, frankly, I’m tired of hearing/reading “there’s no noticeable improvement or difference” between Arduino and Smoothie/ARM boards. Showing them running the same settings will effectively dispel that nonsense.

« Showing them running the same settings will effectively dispel that nonsense. »

Eh, no, guys, be objective, scientific and stuff, don’t make assumptions :slight_smile:
Maybe in your specific setup there won’t be any differences, I’m pretty sure that’s possible ( though rare ).

Fair enough. We will be as scientific as is possible given our setup.

Though we do already know the outcome of many of our tests because we’ve been using these machines for quite a while now.

:slight_smile:

The difficulty I see is to what do you attribute any of the subjective differences. Is it the silicon, firmware, rest of the supporting circuitry? Without a control group and a defined testing model it’s really a marketing exercise and not an engineering exercise. There are just too many variables to pin the difference on one specific thing. It’s good click bait but at this point it doesn’t seem to be engineering based.

The advantage of a 32 bit MCU over an 8 bit MCU in terms of processing power is already a given fact, far longer than 3D printers or even CNC motion have been in existence. Much of the advantage is not going to be available to the naked eye, even under time lapse. To get a better perspective you’ll need to look at the architecture of the platform, how well and how fast it handles certain functions, how efficient the code is on a given platform, those sort of things.

I’ll add you to the list of those who will be very much surprised. MUCH of what we will be filming is clearly visible to the naked eye. I’ll allow the results to speak for themselves.

Yes. The proof is in the pudding. I can only speak to cartesian bots. A couple of annoyances I’ve seen: when anyone complains about dimensional accuracy of a given bot, they just haven’t dialed it in. On the flip side, deltas - I hear - are difficult to dial in dead-accurate dimensional parts for any size part. If true, that can’t be glossed over.

Also, when people show a print on a printerbot that lacks surface quality or appears to produce lower quality than my team here gets, what do you do with that. A printer in the hand of a less-than-expert person doesn’t really speak to what is possible, just what was achieved by that person. That one is tough. Many experts could take any printer- even bad printers and make them sing. How do you qualify the value of the user whose hands it is in?

Ease of use is often overlooked.
Price or the value proposition is often under-reported.
Support or customer service is often overlooked.
Company track record or community reputation is often glossed over.
The relative speed of a print is often misinterpreted by the newbies… Like, “10 hours is better than 9”… Is it really?

And who is really neutral? Not me, I admit. I have a limited scope of experience and a dog in this hunt.

Truly unbiased reporting is tough. But I look forward to the comparison. I applause these efforts, b/c my hope is that the community will benefit.

The 3D printing landscape is changing. Old assumptions are being challenged. New approaches are emerging. One good example is the transition Printrbot is undertaking to move to 32 bit arm and completely new (open source) firmware with a special emphasis on ease of use and user friendliness. The statistical comparison of manufacturer-published specs has become less and less useful. The nuance of ease of use is making a slow march to the forefront. Technical specs matter less now. The plateau of the hardware itself is evident and software is the new battlefront, in my opinion. In a year or so, no one will argue that 8 bit Marlin is good enough, or that repetier, etc (in its current form) is ok.

I, for one, am ready to swallow any bitter pill necessary to move this 3D printer thing beyond the maker movement and into the realm of the average consumer. Why? Not for reasons of profit, but to help consumers discover their inner-maker and experience the pleasure of tapping their own creativity.

Stink, that read like a marketing shpeal!

Keep making, on whatever machine you have!

Brook

@Taylor_Landry You’re already approaching it from a bias. You’re interested in proving your point, not providing an objective engineering overview. That’s all well and good but don’t cloak it under some engineering exercise when it’s anything but. What you are proving is for this specific app you are able to tune an 8 bit AVR better than a 32 bit ARM. What you need to do is let Arthur tune the Smoothie.

It’s not so much a question of if a 32 bit MCU is more powerful, it’s a question of if it’s required for any given application. That’s certainly debatable though I think for the bulk of Reprap/DIY or even low end complete Cartesian machines an 8 bit AVR is plenty given the current state of widely available firmware.

Blindly flogging 8 bit will hinder development of machines that will be easier to use and have greater capability. Right now the bottleneck is the availability of more full feature 32 bit firmware/ hardware combinations. With scale you could basically put a smartphone inside a printer. The parts are cheap and once something like that starts happening Shenzhen fab shops will start cranking them out by the thousands. You can get RAMPS/MEGA/Stepstick clones in 1000 qty for under $10. Right now something like Smoothie or Replicape is expensive because it’s small batch. Once 10-100k batches start rolling out the price is going to drop to the point where 8 bit won’t be so attractive.

How about we stick to my original question posed… You really have no idea what my bias is (and I do, obviously have one), nor what we are going to be showing, so conjecture doesn’t help anyone.

I can assure you I’m 100% committed to progress. I was an original KS backer of smoothie, and am genuinely excited for the future of consumer 3D printers.

I literally deal with hundreds of customers every month who want to get into the space. It would serve us no purpose to bash or denigrate hardware for its own sake. Especially since 100% of the machines we carry run on Arduino boards.

We hope to have the filming done this week and with any luck, it will be edited and posted next week.

Please forgive the takeaverse links, but how about these?
They will get you thin vertical fins, spikes, walls, gaps, overhangs, etc.

Also, please excuse me for having redundant options. Some of those might take care of the tests that are done by a number of the other items.

Nice. A couple of those would definitely be good to add to the mix. Thanks

Hang in there. Be as open minded as possible and it should be an interesting conversation. Always cracks me up to see the passion start flying. I think it’s good overall. Those of us that deliver products to customers have thick skin and are well worn on the road of opinions and community engagement. I ain’t scared. Go for it.

And I think both camps have good points. I’m looking forward to showing of Synthetos Tinyg- a 32 bit arm platform that will show speed and quality improvements over anything out there. Mark my words… And take them with s grain of salt :wink: if there ain’t video and in-person Demi’s, it didn’t happen, right!?

Brook

I would add the delta printer calibration print the prints the concentric circles and the lines going towards the center and a similar thing for the cartesians just so that you can prove the beds are levels and all.
Also, try Slic3r, Cura and at least one other slicer. It may turn out that some slicers get better results from different firmware and hardware. Polulu vs mosfet is another variable and it may turn out that those favor certain firmware and hardware.

i think one way someone can run a shoot out like this is to pick a couple users who are familure enough to change out a board and send them a set of boards you want tested. then they can do a shoot out from ease of use to final print product.

hard thing is getting the boards back.