The Mini Mobile Robotic Printer

[Jarrod Roberson]

The slightest input onto the table surface is going to throw this thing off more than it can account for quicker than it can recover from.

This is just a big FAIL.

I fear that you have created for yourself a strawman (or, if you prefer, a red herring; a MacGuffin). I recommend that you take a chill pill before reading on 

Mouse sensor can't detect or measure YAW, only relative X/Y translation speed and direction.
And it will never have absolute positioning to recover from YAW errors. Thus it has no feedback loop for correcting YAW.
Just the normal motion of this thing will introduce YAW errors that can't be recovered from.
It isn't as simple a problem you make it out to be, this is not going to be solved at the price point they are targeting, if it could be, it would have already been done.

[CanadianAvenger]

For yaw to be an issue the center of rotation would have to be concentrically aligned with the optical center of the sensor [an unlikely scenario in this case]. Having said that, yaw can be detected, and measured by using 2 mouse sensors, as any rotational component will show up as a differential between the two sensors.

[scientist]

They may also be using gyros.

[Duane Degn]

They've clarified how they are going to solve, pardon me, "have solved" the yaw issue.
On April 20 they posted:

We solved the problem of printing two lines in parallel by designing the drive system in such a way, that the robot has only two degrees of freedom, it is only able to move along the X and Y axis and cannot rotate around the Z axis at all.
Also there is no backlash because the wheels are connected directly to the motors.
All of the above, in combination with high resolution optical flow sensor, allows us to control the printer position with high accuracy.
Here's a video we posted bellow from a few months back showing multi line printing (you may have missed it)-
http://youtu.be/-tcYm_WWFys

This "optical sensor" is also mentioned at the bottom of the FAQ section. The last question asks "How does the printer keep the paper stable?" To which they answer:

Combination of the printer's weight and its Omni-wheels, also, the stepper motor make micro and steady steps, so there aren't any sudden accelerations. We're using a high resolution optical laser sensor in order to control the movement, speed and placement of the bot.

So if they're limiting the robot's movement to two degrees of freedom what is the "high resolution optical flow sensor" sensor? Detecting missed steps in the stepper motors?

Is there such a thing as a stepper motor with the precision this would require? They say there's "no backlash" so it must not have a gearbox.

I'm really at a loss to figure out if they really think this will work. I'm not inclined to think they're just going to take the money and run but from my experience of building small robots, I don't see how their two degrees of freedom strategy has a remote chance of working except to possibly print lines of text which don't require multiple passes to align precisely.

[PlainName]

Damn, I should've patented my idea

I have no doubt that their system will work, but it will probably work in a flaky way: given a optimal setup it will do the biz, but use the wrong paper or the wrong surface, etc...

[Jarrod Roberson]

maybe the machine can't move in the Y axis, but the slightest bump on the surface it is "printing" on can induce an open loop yaw and it will never detect it.

To prove this just put your mouse on any semi-unstableized surface and bump the surface and see if the mouse registers the movement.

This the open loop I mentioned before, if they can't move in the Y axis then they can't compensate for the Y axis.

[Duane Degn]

I don't see how their two degrees of freedom strategy has a remote chance of working except to possibly print lines of text which don't require multiple passes to align precisely.

I think my prediction has come true. I think it's clear the printer will only be printing lines of text which don't require multiple coordinated passes.

Not surprisingly . I imagine the transitions from paper to desktop and back to paper would cause an alignment issue for the robot.

Since the robot needs to stay on the paper, they are making the robot even smaller than the one shown in the KickStarter video. This has meant they had to modify an existing cartridge in order to make it as small as possible.

A lot of backers are upset by the change to a proprietary cartridge but I don't think ZUtA Labs had much of a choice in the matter. The margins shown on their video are very large. If their robot were any larger, there would be even less room for text.

I still think this KickStarter campaign was misleading.  The "story" section of the campaign includes five photos of printed pages. I think it is very unlikely the robot will ever be able to duplicate a single one of the "example" prints.

I do think whoever ZUtA Labs got to build their latest robot did a great job. IMO, the printer is working as well as could be expected given the severe limitations of having only two degrees of freedom with the robot's motion.

I doubt many of the backers realize the wide margins in the update video are mandatory. The backers seem more upset about the proprietary cartridge than they are about the small printable area on a page.

Edit: Fixed links. This forum uses a different format for links than what I'm used to. Thanks electr_peter for pointing it out.

[electr_peter]

They have a prototype (polished from outside) that prints something, that's good.
However, print quality is not shown; printing speed is pretty slow; those margins are a bit wide; it uses proprietary small cartridges.
Even the linked video shows two different prints of the same page (look at left margin in the middle and the end of the video) - I hope that is not CG or repeated attempts to get something printed...

[sunnyhighway]

They have a prototype (polished from outside) that prints something, that's good.

Where do you see it actually print?

The clever montage of the individual shots give you only the impression something is being printed.
At no point in the video you can see a single letter being printed on a spot that was not printed at the beginning of that shot.

All I see is a robot that moves over a pre-printed piece of paper.

I call Shenanigans...

[Duane Degn]

Where do you see it actually print?

The clever montage of the individual shots give you only the impression something is being printed.
At no point in the video you can see a single letter being printed on a spot that was not printed at the beginning of that shot.

All I see is a robot that moves over a pre-printed piece of paper.

I call Shenanigans...

While I also think plenty of shenanigans are going on, I'd be surprised if they didn't have a robot capable of printing a few lines.

At 0:26 in the video, a new line of text appears as the robot moves to the right.

I do think it's very suspicious the top three lines of the text are so different when seen at 0:32 and 0:40. It sure looks like 0:40 page is different from the 0:32 page.

I don't see why they would fake the print but then leave the misaligned text shown at 0:32.

[electr_peter]

~0:20 - paper is very shiny, like a plastic, and thick. I have seen various type of paper, but I would not call that an ordinary paper for printing purposes. If it is paper at all...
Whole video seems to be heavily edited - most likely to hide actual limitations of printing action and inconsistent printing (or whatever action it tries to imitate) quality.
I still think that some printing action is going on. Printing demo (that is, single word was printed in one go) was demonstrated (if I recall correctly) few times before with a barebone robot.

I found a real alternative for printing wherever you are aficionados. Brother makes PocketJet printer - it can print on thermal A4 paper. Not too bad except for the special paper.
Or you can use HandJet EBS 250 portable printer (laber printer for outdoor use).

[PlainName]

That Handjet is brilliant. Wish I had a) an application for it, and b) the funds (they don't give a price, which normally means "We're afraid to tell you because it's so expensive").

[Mr.B]

That Handjet is brilliant. Wish I had a) an application for it, and b) the funds (they don't give a price, which normally means "We're afraid to tell you because it's so expensive").

About USD3500 each.
Standard ink about USD75
Quick dry ink about USD100

Not involved with the product in any way, just saw it at a trade show recently and thought it was so cool.
I don't know the price of the EBS-260, but that model looks better again.

[PlainName]

Oh dear. That's almost affordable 

[Duane Degn]

Here's why I don't think the video is a complete fake.



If they were faking the video, why would they leave the above image in it?

It's clear from later in the video (0:40), the top three lines are supposed to be equally spaced. I also think they were intended to be parallel.