Author Topic: Suggestions to improve the Zeus  (Read 3507 times)


  • Newbie
  • *
  • Posts: 9
Suggestions to improve the Zeus
« on: December 26, 2014, 05:25:07 PM »
I thought I would also share some suggestions for improvement to the Zeus printer. These are all things I've noticed in the first two days of printing experiments, and I'm mostly sticking to hardware right now.

1) The feed tube is crimped too tightly by the zip tie holding it in place. It makes it hard to feed the filament through. There should be a better attachment there. Also, the zip tie end is so long it sticks up past the opening. It's a rare bit of amateurism in an otherwise excellent build.
2) The filament gets hung up on the lovely double hinges while printing, pulling out excess and making noises. There should be a better feed mechanism or path for it.
3) The filament guide wheels aren't perfectly aligned front-to-back with the guide hole and the extruder. It makes it a little tricky to feed filament through these three components.
4) I love that there are nice bright LED lights inside while printing but they're too low to do much good. Adding another LED or two at the top of the printer would help illuminate the build. Maybe even put one on the extruder itself!
5) Zeus really needs a sleep function. After a half hour or so, the LEDs should go out and the screen should turn off. It should also have an automatic power-down function after a few hours. Most prints take long enough that I'm no longer around to turn the printer off afterward.
6) There should be a simple "Change Filament" option in the menu. It should park the head and heat up the extruder so the filament can be removed.
7) It's not clear in the documentation when you need to use the calibration mat and when you don't. I guess it's not needed for scanning in the latest software rev but that's not really stated anywhere.

I hope this helps!


  • Newbie
  • *
  • Posts: 23
Re: Suggestions to improve the Zeus
« Reply #1 on: December 30, 2014, 07:07:33 AM »
I agree with all those points.

2) I wonder if having the filament pulled across under the lid and then straight down into the extruder will tend to bend it too much.  So far it hasn't been a problem though.  But the filament rubs against the underside of the lid.  That may make some marks on the lid over time.

4) I hadn't really noticed the LED position being suboptimal until you mentioned it.  Maybe it's optimized for scanning and taking photos?

5) yes it's hard to believe they actually don't have screen blanking.  The screen is running an X server.  So you can open a terminal and type xset q to check the blanking timeout (among other things):
Screen Saver:
  prefer blanking:  yes    allow exposures:  yes
  timeout:  600    cycle:  600

That should mean that it will blank after 10 minutes, but it doesn't.

Of course to turn off the LEDs they will need some kind of daemon.  Might as well have a proper screensaver daemon then... it could show something cool for a configurable number of minutes before blanking.  Something better than just a bouncing AIO logo I hope.  Of course the screensaver shouldn't obscure the progress readouts when the machine is actually doing something.  Probably the reason they don't have a screensaver yet is that prints take hours, and it's nice to be able to see the progress bar the whole time.

7) I'm not having good luck with scanning yet.  There are always extra "wings" and stuff.  Maybe I will write another post about that.

Additional points:

8 ) The worst thing that happened so far: the glass build plate came loose from the metal tripod thing which holds it on the turntable.  A print got stuck a bit too well to the build plate: I applied a liberal amount of Uhu stick, and the print had a large flat area.  Afterwards I couldn't get it off, so I tried soaking the whole build plate in a sink full of water, to dissolve the Uhu.  It did, but I still had to pry a bit to get the print off.  Fine... live and learn... if it has a large bottom surface area, then I should not use glue, or use glue which has already had some prints done on it so that it's not so sticky.  Later I did another print though, and when trying to pull that one off the build plate, I instead managed to pull the build plate right off its mount!  I can see that it is attached with a thin clear adhesive pad, like a sticker that is sticky on both sides.  Apparently it also has water-soluble glue.  That's not a good idea, because probably everyone will need to wash their build plate now and then.  It's actually a great feature that the build plate is removable, for just this reason.

9) Speaking of build plate mounting, I'd like to have a lot more options for different types of build plates.  So, I think AIO should sell that metal tripod thing separately, or include an extra with the machine.

10) The plate for scanning should be separate.  If it needs a calibration target, that could be permanently attached, or even etched into the glass.  My calibration target doesn't sit quite flat (it bows upwards a little in the middle) so I think that may cause some inaccuracy.  And now that my one-and-only plate has gotten detached from its mount that one time, it may be sitting a little off center anyway.  I should have bought the extra build plate for $69 I guess.  I was thinking at the time, what if the glass breaks one day, then thought nah... but now I see there is more than one reason to have an extra.

11) Actually the main alternative type of build plate that I want is a heated one.  It's IMO a pretty big deficiency that a machine in this price range cannot print anything other than PLA, because it doesn't come with a heated bed.  It could at least be designed for upgrading: there should be a FET-controlled power plug for plugging in the heated bed.  Of course that might require a bigger power supply.  But still, at 24V it doesn't take many extra amps.  It's OK with me if heating up the bed takes some time, to reduce the wattage requirement.  I built my own printer while waiting for this one to arrive, and I managed to run the whole thing (including heated bed) from a 19V laptop power supply.  It just fits within the power supply's max rating.  The Zeus already has a bigger supply than mine does.

The thing is, the build plate doesn't need to rotate, as long as they are not planning to use the rotational access as part of the printing process, for printing smoother round objects.  (That would be an excellent idea IMO, but then it would need to be able to rotate a couple orders of magnitude faster than it does during scanning.)  So, I think the heated build plate should be rectangular and should attach to the Z table by other means.  It should have been designed with an alternative mounting method already: just some additional tapped holes in the Z table would be a good start, and the alternative build plate could then be attached to screw-in standoffs.

12) Regardless of that, I hope AIO plans to continue researching the necessary parameters to print with some other types of plastic that can be printed without a heated bed.  ABS is nearly impossible without one, but Taulman and others have been coming out with filaments which don't shrink quite so much as they cool.  I may try that myself (and I hope I will not be told that trying different filaments voids the warranty or something like that.)  Do we know if the extruder can actually handle higher temperatures?  Oh that reminds me... I really want to try some flexible filaments like NinjaFlex and such.  But I don't think the extruder looks like it will handle it well, because there is enough space between the capstan/pinch roller pair and the actual hot end.  There should be more support to make sure the filament is forced through.  But I haven't tried this yet with any extruder.

13) As has been pointed out in another review I read somewhere, the attempt to wipe off the excess filament on that rubber wiper thing before starting a print is cute, but not 100% successful.  Usually it knocks off the filament on the first pass across the rubber, but then it rubs its nose in it some more and sometimes manages to get it stuck on the brass surface again.  I always wait with a pair of pliers to grab whatever excess doesn't fall off, just like I do with my own printer.  Have to be quick about it since there's no pause.  But I found the startup gcode file, so fixing this is probably easy.  It should be enough to extrude some filament and then go across the rubber only once to knock it off, I'm guessing.

14) I want to have an extra filament feedthrough, on the back of the machine at the top, I think.  The reason is I made a filament rack which sits on a shelf above the machine, so that I can keep several colors and types of filament handy.  Since the internal feed path is so unpolished anyway, I don't really think a simple hole to feed in from the outside would be any worse.  And it would save time compared to the current process of making extra spindle adapters for every spool, pulling out one spool, putting in another one and running the filament back through that tube again.  I can of course fabricate a plastic grommet to make a smooth feedthrough and prevent any binding, kinking or pulling around a tight radius.  (Would drilling a hole in the case void the warranty?  Hope not.)  I figured I will think about optimal placement for a while before actually trying it.

15) the ribbon cable to the extruder does not inspire confidence.  (Is it really OK to run high current for heating the hot end through a thin ribbon?  I normally think of ribbon as only for signals, not power transfer.)  I already managed to pull it out of the printed mount on the Y axis once, accidentally.  It was no problem to put it back though.

Shall I even get started on the software UI in this post?  It's a great start, really.  But I could think of a pretty long list of things to improve.  Little stuff mostly.  Maybe I'll post more later about that.

I was surprised that the UI doesn't offer any more than 30% infill by default, but after making a custom profile with a higher percentage, I see why: the more plastic, the more it warps.  But sometimes it's nice to be able to print something very strong, especially if it has a shape which does not curl too much when printing.  This just requires more experience on my part.  I do like to try to design objects more like design for injection molding, where there is actually not a lot of infill.  Or if there is infill, maybe it is necessary for strength, so I want a higher percentage.  Some things I have made were weaker than expected (like that one which got stuck - I actually managed to delaminate a couple layers in the process of trying to get it off the plate.)  If it had 100% infill, it might have been strong enough.

I also want to try using Skeinforge to generate gcode for the Zeus, so I have to figure out the necessary parameters.  It would be nice if config files for other slicers were included somewhere, to be stumbled upon by people who dig around a bit.  ;-)

I'm sure it's capable of much higher speed printing, but maintaining quality gets tricky.  I will experiment more with that later.

I wasn't able to use a WEP key with the wifi, it apparently only supports WPA (which would be better anyway, but I happen to have old-school WEP at the moment because of an old Squeezebox music player which can't support WPA).  So I just use ethernet for now.  Probably will switch back to WPA2 at some point.

OK this is sounding too negative.  Things I love so far:

1) Resolution is pretty good.  (Compared to my self-built printer, almost anything would be an improvement.  I don't yet have experience with other high-end printers besides the Zeus.)

2) The default slicing settings result in very few curling problems.  So the out-of-box experience for printing tchochkes (where high infill is not important) should be just fine.  (I'm being too conservative, trying not to print tchochkes like most people do.  I guess I should print at least one silly useless object simply for display.)

3) Just the promise of having an all-in-one machine.  It's nowhere near fulfilled yet, but it's early and I assume it will really be a 3D copier some day when the scanner starts working better.

4) Good full-featured Linux installation out of the box.  SSH works.  I have root access.  Even avahi was already installed!  That means I can change the hostname to zeus, and ssh to zeus.local.  scp is convenient for uploading models anyway (but of course I don't do that as root, I made myself a user account):  scp *.stl zeus.local:/var/www/data/files/  (which isn't a very standard place for putting a user's downloads, but whatever).

5) Nice ARM chip with a GPU, and plenty of flash storage.  The UI could be so much more than it is, because it's got the horsepower to do realtime rotation of models on the touchscreen, and other types of UI animations like flicking through lists etc.  (They just need to be using Qt 5.  I will try that myself some time.)  Also it appears to me that they don't use the ARM for any real-time control: there is a separate Atmel like almost every printer has to control the steppers etc.  (I haven't had a peek at the hardware inside yet, I just gathered that by poking around the filesystem and found an atmega firmware file.)  They probably could have done everything with one processor, but using a separate microcontroller helps to avoid any stutters (and consequent blobs) while printing.

6) The power button is nice: it doesn't take long to shut down, and the color change green-orange-red is nice feedback.

7) Good extruder design, especially compared to the ebay special that I'm using on my own printer.  ;-)  This one really inspires faith.  I hope it turns out to be durable.  At least it never skips any steps AFAICT.

8) Z axis probing.  Sure it got to be standard on most printers by now, and I do wonder if it's really necessary before every print.  But at least we know bed leveling should never be a problem.

9) The build plate mounting.  Apart from the fact that it comes unglued of course.  ;-)  It's very nice to be able to remove the build plate for cleaning and for removing prints.  And the way that it slips in place into the three holes is nice and smooth.

10) 24V universal power supply.  Increasing the voltage (beyond the common 12V that many printers use) is how you get the current requirement down, I know from experience.  And it works in any country.

11) both wifi and ethernet


  • Newbie
  • *
  • Posts: 23
Automatic photos of prints?
« Reply #2 on: December 30, 2014, 07:12:35 AM »
Discussing the lighting reminds me of the camera: I thought there would be a feature to take pictures of prints.  I would like to have it do that automatically every time a print is finished, actually.  Use the date/time stamp as the file name (as the scanner software already does), and save the JPGs sequentially in a directory that is accessible to the web server, for easy downloading from the web UI.


  • Newbie
  • *
  • Posts: 5
Re: Suggestions to improve the Zeus
« Reply #3 on: July 24, 2016, 07:15:28 PM »
@ecloud, in the time since you wrote your treatise of improvements at the end of 2014, where you mentioned you'd be trying a few things, did you actually get around to doing any of them?

If so, how'd they turn out?