# Portable CNC - Leaving the workbench behind

### Help Support HomeBuiltAirplanes.com:

#### ScaleBirdsScott

##### Well-Known Member
Santa Claus (my wife) authorized me getting an Origin last Christmas but I decided to wait, at least, a year. I think the 'concept' is completely valid but the 'real life' hardware and software are still maturing too fast to *jump in now.

*I learned my lesson when I bought my first CarveWright machine before they we're ready.

>>>random thoughts<<<

I just watched the "pros and cons" video on SBS's link. I disagree with a lot of what the guy said (feeds and speeds calculations being an issue, etc.) but he did say his opinion was biased.

Comparing an Origin to a Gantry style CNC machine is like comparing a table saw to band saw. They can "kinda" do the same thing but they're different tools with different missions.

We've talked about "thumb drive and pile of wood" airplanes before, I think there could be an " 'Origin' and pile of wood (or aluminum)" airplane just as easy. ...design the airplane to match the tool, just like Bud Evans designed the VP-1 for a drillpress and jigsaw.
I think the comparison of tablesaw to bandsaw is pretty valid.

I do think it's an interesting point about speeds and feeds as, yeah, the Origin makes it seem like you just plug in paths and that's it. A stepper-driven machine you do need to use some CAM software and mess with gcode and dial in your speeds/feeds and while it's not hard, it is something you could mess up. Could also screw up your coordinate system entirely with the standard fixed machines. There have been times I've screwed that up.

I'd like to have both tools eventually.

#### FritzW

##### Well-Known Member
Log Member
...There have been times I've screwed that up.
I've never screwed up but I've heard of people who have :whistle:

##### Well-Known Member
Oh my, it's been over 6 months since I got my Origin...

To be honest, haven't really used it too much due to craziness at work and my "homebuilt" time devoted to finishing up my design on paper.

I can't speak for larger work pieces, but it seems dead on for anything under 2'x2'. Not much has changed perception-wise since my last thoughts https://www.homebuiltairplanes.com/forums/showthread.php?t=29725&page=2&p=425515&viewfull=1#post425515

The last software update, which updates automatically anytime you're on the internet, had some huge tweaks such as increased plunge and retraction rates to stop cutting to avoid mis-cuts.

Otherwise, what impressed me the most after purchase was how much thought the Shaper team put into work-flow. The software is built around Autodesk Fusion 360 (which is free for hobbyists by the way) SVG files. The same files can be generated by Adobe Photoshop and Illustrator, which opens up the device to graphic designers. The workflow after you generate the files is to upload them onto your account on the Shaper website, which your Origin will automatically download. Alternatively if you want to keep your files off the interwebs, you can use a flash drive.

Oh, and with the right cutting bits, it'll readily cut aluminum and probably brass. Larger skins might be a bit tedious and a bit problematic due to the tolerance wandering off to ~0.03 for pieces around the 8ft size, but it'll do the job.

HBA Supporter

#### FritzW

##### Well-Known Member
Log Member
Now that I think about it, the Shaper Origin seems like exactly the right tool for something like...
The Origin is a really neat tool (it's probably the next major tool I buy) but I don't think it'd work well for the Mk-2 wing (or any tab and slot stuff).

Radicaldude1234, our resident Origin expert, say's it's "dead on" within about 2' x 2'. There are VERY few parts on a Mk-2 wing that would fit in that sweet spot.

As I understand it (bear with me) when you set your home, zero, location (0,0,0,) the Origin's camera sets the domino dots at zero for that location. As it moves it sees more dots and sets those dots as offsets from the 0,0,0, dots. The location the Origin sets for those dots have some error (even if it's very small). As the Origin moves it sees more dots (each with it's own very small error) and adds those locations (with their errors) to the last dots. (errors are cumulative). Even if the Origin gets back to it's "home" dots, it adds the accumulated errors to the location of the home dots and continues from there. ie. you start at 0,0,0, when you get back to zero the new 0,0,0, is .002,040, 00? and the error just adds up from there.

It's not really the size (2' x 2') that makes the sweet spot, it's how many times the Origin has recomputed it's location from camera interpreted dot locations (errors). When you cut something like a wing rib your staying in a small area but your recomputing (adding small errors to small errors) the router position a whole bunch of times. ...errors add up

Radicaldude1234 says we're okay to about 24". Think about the 25' (300") perimeter of the spar that takes 4 passes (1,200") to cut. You'd be lucky to finish the cut, let alone end up with a usable spar.

There's a reason, like it or not, that 99.99% of CNC routers are the gantry type. ...they take up a lot of room but they work.

#### Vigilant1

##### Well-Known Member
Maslow, Origin Shaper--it seems that many of these units are cutting based on where the controller "thinks" the cutter should be (based on counting previous dots, how far the suspension mechanism has moved--since it was last calibrated, not counting cable/chain stretching, etc). It seems better to make the cutter much smarter about where it actually is, in real time. Feedback on the actual cutter location (three "dumb" omnidirectional light sources of different colors around the workspace and a "smart" high-resolution angular sensor on the cutter unit?) would go a long way toward enhancing accuracy. I'm sure this "closed loop" approach is old hat in the high- units.

#### FritzW

##### Well-Known Member
Log Member
There are two very similar threads running at the same time: Portable CNC and Handheld CNC ...so:

Bottom line, (IMHO):

Maslow has the mechanical aptitude of a fruit fly. He's a blooming genius at getting equally "mechanically challenged" Patreon supporters to fund his 'man bun' but if he dropped his Snicker Bar he wouldn't know why it fell toward the Earth. People who haven't got a clue about 'defection in a structure' keep thinking "...on the next software update, steel cables and bicycle chains will stop deflecting" aughhhhhh!!!!!!

...yes, he has a 'man bun' but he still hasn't got a clue about why his machine doesn't work. ...yes, it'll fit against the wall of your garage and yes, it only cost $10 ...but it doesn't freaking work!!!! WTFU people! (Wake The heck Up people) The Origin has a lot more promise and it has a legitimate business structure (they tell you what they're going to sell you and they sell it). But it's a looong way from being as practical a regular 'ol gantry style CNC machine. ...it's system of knowing where it's at is still flawed. ...it needs still needs some time. .....IMHO #### proppastie ##### Well-Known Member Log Member yes, it'll fit against the wall of your garage and yes, it only cost$10
It does work....just not as good as other machines. Its great for making signs, chairs, tables, bookcases, or anything that does not need a tight tolerance. The best features are as you pointed out, and playing with it is better than watching TV.

##### Well-Known Member
Well, to be fair I only say I know for a fact it's good to 2'x2' because that's the biggest work area I've tried it on. I've not yet made full use of my awesome assembly table that I built a while back https://www.homebuiltairplanes.com/forums/showthread.php?t=29724&highlight=. I've only heard that it's not as accurate for larger pieces but have not experienced it.

I am quite certain that the there's an algorithm in the Origin's software that self corrects so that the origin (0,0,0; heh) doesn't migrate. When I was testing it out with the US map, it could retrace the state lines repeatedly, with an engraving bit, after almost 2 hours of cutting.

Knowing where you are has been the problem with machining from the beginning. Stepper Motors do it by counting angular steps, and Servos rely on a separate feedback sensor. Shaper does it by stitching together the work area in machine vision using the domino stickers, which are well understood and fixed (by the way you can in fact print your own domino stickers, though you won't be saving much money http://browncowww.pythonanywhere.com/). How Maslow does it with moving chains (each link of which has movement and therefore compounding error) and claim those tolerances is beyond me.

PS: I looked in the Shaper forums and the company's CEO had this to say:

[FONT=&quot]The accuracy numbers the author is reporting may not be crazy off. However, the repeatability number he’s reporting does not make sense. There seems to be a natural tendency for people to want to slot data into a neat little matrices in order to compare performance with conventional gantry-based CNC routers; e.g. [/FONT]What is the accuracy? What is the repeatability?[FONT=&quot] We understand this desire. But keep in mind that a human operator is still involved in the cutting process with Origin, so it’s not necessarily simple to say unequivocally: performance = [/FONT]this[FONT=&quot], [/FONT]always[FONT=&quot]. Experience certainly makes a difference, particularly in cut quality. I’m sure many users who already have Origin in hand notice that their cut quality improved rapidly after making a few cuts. This is why anytime when we describe Origin’s performance, we prefer to say that the tool is [/FONT]capable of[FONT=&quot] doing such and such.[/FONT]

[FONT=&quot]Before diving in further, let’s make sure we’re all on the same page w.r.t definitions. We see/hear a lot of people use the terms accuracy and precision interchangeably. They are not the same. Accuracyrefers to the difference between the dimension you expect to see (the intent) and the dimension you actually see after cutting (the result). Precision refers to the variation you can expect to see from cut to cut; e.g. if I cut out the same part twice, can I expect to see the same result each time? If I cut multiple passes on one part, can I expect to see “steps” or variations at each layer? Or can I expect the paths to be exactly on top of one another? Oftentimes, the terms repeatability or reproducibility are used in conjunction with precision.[/FONT]

[FONT=&quot]With Origin, the best way to think about accuracy is in terms of a percentage of target. Origin is currently capable of achieving ~99.7% or 99.8% overall accuracy. For small projects (e.g. ~6 inches or under), expected accuracy of cuts performed by a reasonably experienced user may be measured in single-digit thousandths of an inch. For medium-size projects (e.g ~1 to 3 feet), expected accuracy of cuts may be measured in hundredths of an inch. For larger size projects (e.g. ~4 to 8 feet), expected accuracy may be measured in tenths of an inch. As community member @Brock has noted, system accuracy is dependent on a number of variables, including ShaperTape application, scanning accuracy, cutting from one orientation, etc. He also brings up a good point that it’s actually quite easy to not measure things accurately. The really good news here is that we make continuous software-related improvements to several of the variables (e.g. scanning accuracy) that directly impact cutting accuracy. So we expect that the overall accuracy Origin is capable of achieving will actually improve over time.[/FONT]

[FONT=&quot]Origin is inherently precise; i.e. with care, cuts are remarkably repeatable. One of the major advantages of Origin’s positioning system is that even if [larger scale] cuts are dimensionally off by some small amount, an experienced user can expect to achieve the same results over and over and over. This is pretty wild to consider, and extremely compelling for most applications we encounter. This is how you’re able to return to a taped workspace even days later and shave another tiny sliver off of an existing cut. And this is why the 0.01" number the author claims in this article does not really make sense. If it is true in his experience with the tool, we would like to better understand how he is measuring it. If he is indeed seeing repeatability variations as poor as 0.01", it may indicate that there may be some issue e.g. that the spindle is not clamped properly. We do not consider 0.01" repeatability to be good. If true, it would indicate that for repeated passes on the same path, a user may expect to see a 0.01" step deviation during a depth change - we definitely do not experience this if we are careful and approach cuts the same way for each pass. Or if true, it would indicate that if a user were trying to remove 0.001" from a previously cut mortise, that she might actually end up shaving off 10x that amount - we definitely do not experience this either. We believe the author may be off by an order of magnitude or more in his reported number.[/FONT]

[FONT=&quot]I hope this explanation helps describe the performance you may expect today with Origin. For anyone on the fence, my best recommendation is to try it out to see if Origin is right for you and the applications you have in mind![/FONT]

Last edited:

#### FritzW

##### Well-Known Member
Log Member
It does work..(Maslow scam).
No It doesn't. Not when we're talking about practical "part making". It'll cut out low quality "pretend" parts ...eventually. But other than making a pretend, sloppy ass "see, it works" 'wine rack', it's useless. ...give me a sharp rock and the same amount of time and I'll make make better/faster parts than a Maslow ever could.

Its great for making signs, chairs, tables, bookcases, or anything that does not need a tight tolerance.
Bull Defication!!!! Give your Etch-A-Sketch a shake proppastie!!! The Maslow sucks at all of that! It might do a ****ty job of it it, one of these days, ...if the parts don't need to fit together at all. ...but it sure as f^@# isn't "great" at it.

Send your money to Maslow's Pateron account if you want (he survives on people like you) but the machine is still garbage!

The best features are as you pointed out...
If you think I've pointed out a single positive aspect of a Maslow, my rant has been in vain. Yes, A Maslow will fit against the wall of your garage, so will a dying syphilitic elephant , I'd rather have the elephant.

IMHO: Maslow is a scammer, he's relying on stupid people to make his 'snake oil' money. ...write your flippen check

...but I'm an introvert and I don't want to burden others with my opinions :roll:

Last edited:

#### cluttonfred

##### Well-Known Member
HBA Supporter
Fritz, on the Origin Shaper and the VP II Mk2 wing, I was actually thinking more of the spar routing, which would be a series of cuts within the 1-3 feet/hundredths of a an inch category described above. I was not thinking tab in slot, that's true, but what about a more conventional VP-2 approach? If there were a way, for example, to use the Shaper Origin to scan or trace the rib drawing and then make up wing ribs without having to make any sort of template, that would already be pretty great.

#### FritzW

##### Well-Known Member
Log Member
...If there were a way, for example, to use the Shaper Origin to scan or trace the rib drawing and then make up wing ribs without having to make any sort of template, that would already be pretty great.

View attachment 75797
It's a neat idea but it would be very difficult to do, even with a very accurate drawing. It would be impossible to do with something like the rib drawings out of the plans book because they're only rough sketches of the rib to show where holes and slots go.

Even on a PtP VP-2 wing there are places where a pretty small error could make life miserable (spar slots in the ribs, spar spacing, etc.)

#### proppastie

##### Well-Known Member
Log Member
Bull Defication!!!! Give your Etch-A-Sketch a shake proppastie!!! The Maslow sucks at all of that! It might do a ****ty job of it it, one of these days, ...if the parts don't need to fit together at all. ...but it sure as f^@# isn't "great" at it.
Well I am a "musical snob" I have now found a "dyi home tool snob".....thats what we need here a little more passion.....When I was working professionally, the tolerance of the CNC jig grinder was +- .0002 with the traveling wire EDM at .00005 multiple passes, you could get a mirror finish that way. I guess it never bothered me that I did not have 250K or whatever they cost for one of those.

#### Armilite

##### Well-Known Member
Since more People today use Facebook than look at Craigslist, there is a lot of CNC stuff showing up on Facebooks Marketplace today. Just Type in Key Words, CNC, Airplanes, Snowmobile, Boring Bar, Rotax, Ultralight, etc.

#### dcstrng

##### Well-Known Member
If there were a way, for example, to use the Shaper Origin to scan or trace the rib drawing and then make up wing ribs without having to make any sort of template, that would already be pretty great.
That’s about the only reason I’d bother with a techno-gizmo in the shop – I loathe making ribs for some reason; doesn’t matter whether metal or wood, the unending tediousness doesn’t have the tranquil effect on me as some others. But otherwise, I’m contentedly of the mechanical pencil, hand grinder and O/A torch era – I used to have a nice 2D CADD program abut 30-years ago in the DOS 5.1 format. But they kept adding feature after feature through many upgrades and I went back to hand drafting as the program became too confusing. The computer is just a semi-trustworthy box for email and ersatz simulations, not real life… :nervous:

##### Well-Known Member
Perhaps a bit of choice supportive bias on my part, but interesting application of the device: