Giving up on the lasersaur

[kevin.loney]

In June of last year the chinese made laser cutter at our local hackerspace died and the decision was made to switch to the Lasersaur. As a result our laser cutter has been down for the last 10 months. The statement that the Lasersaur itself is at 97% completion is a misnomer, while the hardware might be there the software is lacking in just about every aspect.

The areas in which the Lasersaur is lacklustre at best and catastrophically broken at worst are many and varied:

Firstly there is no proper DXF support even after a full re-write of the library supporting it. This is inexcusable in every sense of the word as there is no absolute dimensional data stored in a .svg file. The conversion from a absolute value to dpi is absurd for anybody desiring precision and repeatability, the two tenants of a CNC controlled machine. 

The lasersaur by default only supports linear moves, G1, in g-code files. this represents such an fundamentally inadequate approach that there is literally no way it could be deemed acceptable by anybody critical of the products that can be produced by a lasersaur controlled machine. expanding the g-code library to include radius, circles and splines would increase the functionality of the controller by an order of magnitude and the fact that these are conspicuously absent should have been a red flag from day one. 

These two reasons alone make the lasersaur controller virtually useless for anything other than neat squares. 

The complete absence of raster engrave support up until recently makes the Lasersaur's usefulness to an artist or somebody desiring the crisp and fast engraving possible with a laser engraver questionable at best.

The front end of the controller is buggy and inconsistent in its failures. 

The backend has sub par logging and the entire process from start to finish is so poorly dameonized that troubleshooting, upgrading and even achieving basic functionality is an exercise in hair-pulling frustration. 

Our adventures in Lasersaur represents 10 months of steady work with nothing to show for it. It represents a $1200 investment on behalf of users of this tool with no return. The decision to wholly rip this controller from the laser and replace it with a 60% cheeper, functional controller with more capabilities should have been made months ago. The complete apathy of those involved due to the constant disappointing performance is the only thing that has prevented us from regaling the lasersaur controller to the scrap bin. There it would join its innumerable brethren of non functioning equipment steeped in good intentions and grand promises.

The lack of a contact email address on the lasersaur website has forced us to post this publicly in the hopes of getting the attention of nortd labs. We're seeking a refund of the purchase price of the lasersaur board. If we can't get a refund we see no justification in leaving our efforts accessible on GitHub and will be deleting the repository.

[Chris Uhlik]

Arcs and splines would certainly make the G-code files shorter, but I haven't had any problems running some pretty complex designs.  Some of them take about a minute to convert to G-code, but they work OK.  My only real complaints are:

1) the BOM sources parts from expensive suppliers --- if you do some scrounging and substitution you can save a bit over 50% of the build cost.  It would be nice if an alternate sources BOM helped participants share this knowledge more effectively.

2) the drive system still has a few resonance ripples so I need to experiment with speeds and run slow jobs if I want really precise, wavey-free edges.

3) better raster support would be nice.  We're working on adding an optical scale and FPGA-based laser PWM control to do high-speed, bidirectional X-axis raster passes.

In my opinion, Lasersaur is a great project for learning about building and debugging a CNC machine.  It's not well suited for unfamiliar, infrequent users in a hacker space who would rather just get something cut than learn about CNC.

Chris

--
You received this message because you are subscribed to the Google Groups "lasersaur" group.
To unsubscribe from this group and stop receiving emails from it, send an email to lasersaur+...@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

[Derek Seabury]

A lesson I have learned in many team projects is that anyone who threatens to leave if they don't get their way should just be encouraged to do so.  So take your toys and go home if you would like.

But quickly for those interested or concerned:
- SVG does support measurements provided in em, ex, px, pt, pc, cm, mm, in, and percentages.  For values given in px (pixels) the lasersaur front end uses a specific DPI but otherwise the exact value.
http://www.w3.org/TR/SVG11/coords.html
- Division is a fairly standardized operation computers are capable of performing repeatably so conversion of relative measures and paths will be precise and consistent
- The lasersaur mechanically only supports linear moves as all motion is translated into steps of the X and Y motor.  More complex G code support would just be translated to left, right, up and down so there is no functional benefit to additional operations except for shorter code.
- Lasersaur is pretty consistent in describing the machine as a laser cutter.  Not an engraver.

Best of luck with your "cheeper" solution, I hope it is of the kwality you seek.  While I am sorry to see anyone have a hard time and anyone withdraw the code they had supposedly contributed to a community it sounds like you had some fundamental misunderstandings.  If you would like to put your drive board up for sale perhaps someone would buy it, but I doubt you will recoup your $1200 as the board costs less than $400 new.

/Derek

[Peter van der Walt (Gmail)]

Well said!

I agree completely. If you cannot parttake in the community this is not the place for you.

If you have a problem, work on it. Fix it. Develop a solution, then bring it back to the community. No use bitching about it... if you dont develop what you need who will...

I suggest you go read up what the phrases 'open source' and 'hackerspace' means as I get the feeling you don't understand either.

Lasersaur is an empowering technology that brought lasercutting to people who otherwise would never have had the skill to use it. Lasaurapp's simplicity opens doors to 'people' and thats what you want in your hackerspaces! Real people with real ideas and ideals!

Peter
South Africa

[Steve Baker]

kevin.loney wrote:
> In June of last year the chinese made laser cutter at our local
> hackerspace
> died and the decision was made to switch to the Lasersaur. As a result our
> laser cutter has been down for the last 10 months. The statement that the
> Lasersaur itself is at 97% completion is a misnomer, while the hardware
> might be there the software is lacking in just about every aspect.

What?!? The lasersaur software is fantastic. My beef has been with the
hardware - but that's essentially fixed now.

I do precision cutting over the entire 2'x4' bed and not one of the
problems you list below is an actual problem. My #1 lasersaur has been
performing perfectly for the last year - and I don't mean at the kinds of
rates you'll be using it - our machine runs 14 hours a day, 7 days a week
without pause. It's earned us over $120,000 for a $10,000 outlay...and
our #2 lasersaur is catching up fast.

> The areas in which the Lasersaur is lacklustre at best and
> catastrophically broken at worst are many and varied:
>
> Firstly there is no proper DXF support even after a full re-write of the
> library supporting it. This is inexcusable in every sense of the word as
> there is no absolute dimensional data stored in a .svg file.

Most laser cutters support some formats better than others. I spent six
months using a $30,000 ULS laser cutter - and the support for any format
*OTHER* than DXF was appalling. The drivers on that machine were just
abyssmal - the only software that worked worth a damn with it was
CorelDraw under Windows - and switching to lasersaur was like a breath of
fresh air. Now I can use our Windows laptop, my Linux desktop or even
our Android tablets and cellphones to drive the machine - and every CAD
package that can export SVG seems to work 100% perfectly with it.

Sure, I don't use DXF - but SVG is a fine alternative - and I doubt there
are many CAD packages that don't export SVG well enough for our needs
here.

> The conversion
> from a absolute value to dpi is absurd for anybody desiring precision and
> repeatability, the two tenants of a CNC controlled machine.

No, that's complete and utter bullshit.

SVG has an issue with the mapping from "pixels" to "inches" being
undefined. That's not a Lasersaur issue - it's an SVG issue.

But so long as you know the DPI setting that your CAD software uses - the
lasersaur handles precision, repeatability just perfectly.

I'm using Inkscape - I know that it uses 90dpi. CorelDraw uses 96 dpi.
I tell Inkscape that I'd like to work in cm's or inches or whatever - and
I make my designs...and when the lasersaur reproduces them, they come out
to the size I request...to about the precision of the width of the
kerf...which is about as good as it gets.

Now, if you flip back and forth between Corel and Inkscape - you'll get
into a horrible mess...but that's not a Lasersaur issue - it's an SVG
issue. So be smart...if you intend to flip between one CAD package and
another, use DXF or ai or something - then at the end, when you're ready
to cut, export to SVG - and when you import the SVG into the lasersaur, be
sure to pick the right DPI setting.

Your beef isn't with Lasersaur - it's with SVG. That's justified - it's
crazy that the SVG design committee never thought to put the words "For
'real world' output SVG uses XX pixels per inch" into the file
specification. But it's manageable so long as you understand what's
going on under the hood - and the benefits of SVG are HUGE because it's
rendered correctly in web browsers. That allows me to keep all of my
designs in a MediaWiki - nicely indexed, categorized and documented and
accessible from anywhere on the Internet. If my wife calls me to ask
what some particular part is called - I can use my phone to bring up our
private Wiki, use the search facility to find the part - and I can view it
right there on my phone with standard software.

Repeatability and such is not a problem. I turn out thousands - perhaps
tens of thousands of each of several hundred part designs - and they all
have to fit together - and they do. We have zero problems mating a part
made six months ago to one we made today from a different SVG file.

Contrast that with the ULS machine I used at my local hackerspace - I
either had to use CorelDraw (closed source - and not Linux - so no
way...but the only software that seemed to ever drive the machine
properly) - or I had to import my SVG files into Corel. Sadly, Corel
doesn't have a DPI setting on it's importer - so there was no way to tell
it that I have 90 dpi files - and so each drawing had to be rescaled by a
factor of 96/90 before it could cut my SVG file properly.

Notice that Lasersaur has a setting for that! Yeaaaahhh!!

So - if you want to use a ULS machine, I strongly recommend that you
adjust your work-path to DXF - and if you want to use a Lasersaur, then
SVG is the way to go. I'm sure there are similar issues with other laser
cutters. The answer is that you learn how your tool works - and use it
intelligently.

You may need to find a better way to convert DXF to SVG...but once you've
done that, and you know what DPI setting to import the file with - you
should be OK.

> The lasersaur by default only supports linear moves, G1, in g-code files.
> this represents such an fundamentally inadequate approach that there is
> literally no way it could be deemed acceptable by anybody critical of the
> products that can be produced by a lasersaur controlled machine. expanding
> the g-code library to include radius, circles and splines would increase
> the functionality of the controller by an order of magnitude and the fact
> that these are conspicuously absent should have been a red flag from day
> one.

The interface to the lasersaur is not intended to be G-code. It's SVG.

You put an SVG file in with any set of curves you like - and the resulting
cut has all of those curves made perfectly.

Your claim that you can only use the lasersaur to cut squares is complete
and utter bullshit - please - let's not lie to each other.

Sure, inside the guts of the machine, the SVG curves are turned into short
straight line segments - but that's done at a precision that's better than
the machine can move - so the result is essentially no different than if
the low level G-code handled curves.

> These two reasons alone make the lasersaur controller virtually useless
> for anything other than neat squares.

That is the most complete bullcrap I've ever heard on this mailing list.
Take a look at the products my business makes - there are curves of all
kinds *EVERYWHERE*.

The biggest issue has been not software - but:

a) The resonance problem (largely fixed in v14, and not a software issue).
b) The belt tension problem (present since Mitsumi changed belt suppliers
- and also fixed in v14).

Those problems are why the lasersaur hardware isn't described as 100%
complete.

> The complete absence of raster engrave support up until recently makes the
> Lasersaur's usefulness to an artist or somebody desiring the crisp and
> fast engraving possible with a laser engraver questionable at best.

I agree that the lack of raster support has been disappointing - but that
was well known 10 months ago when you got your machine - so you went into
that with open eyes. If you thought you needed raster and you weren't
prepared to implement it yourself why the heck did you buy a Lasersaur?
(I'd have thought that any hackerspace worth it's name would have been all
over that - since everything in Lasersaur is OpenSourced and
OpenHardware...but whatever).

Raster is coming...not fast enough...but you knew that at the outset...and
it's coming.

> The front end of the controller is buggy and inconsistent in its failures.

I'm using it with near 100% reliability, (14 hours a day, and about 300
days a year). The only bugs I've had were with using Chrome - which
crashed a few times. I haven't seen any messages from you guys here
discussing your software issues - it's pretty dumb to suffer without going
to the mailing list for solutions.

> The backend has sub par logging and the entire process from start to
> finish is so poorly dameonized that troubleshooting, upgrading and even
achieving
> basic functionality is an exercise in hair-pulling frustration.

What?! I installed the current version - and never had to touch it
afterwards.

I'll grant that getting the hardware sorted out (especially belt
tensioning) was frustrating - but once done, the machine has run
fantastically well.

> Our adventures in Lasersaur represents 10 months of steady work with
> nothing to show for it. It represents a $1200 investment on behalf of
> users
> of this tool with no return. The decision to wholly rip this controller
> from the laser and replace it with a 60% cheeper, functional controller
> with more capabilities should have been made months ago. The complete
> apathy of those involved due to the constant disappointing performance is
> the only thing that has prevented us from regaling the lasersaur
> controller
> to the scrap bin. There it would join its innumerable brethren of non
> functioning equipment steeped in good intentions and grand promises.

You're wrong.

My wife and I have an entire business based around our two lasersaurs. I
have a 'day job' - but my wife's entire income depends on those two
machines - and they deliver, fantastically, day in and day out. They've
exceeded all of our expectations.

Our #2 machine was problematic for a while - but not because of software.
The hardware had a hard-to-diagnose issue relating to the change in belt
supplier and the new belts being stiffer than the old ones. That's fixed
comprehensively in v14 - and the upgrade for my v13 machine was relatively
painless.

If your complaints were about hardware - I'd understand where you're
coming from...but the software?!? That's one of the Lasersaurs' best
features!

Our customers are amazed by the product we turn out with our machines -
and we have a literal waiting list of people lining up to buy it. I see
a third and probably a fourth lasersaur in our future - and from what I've
seen of junk chinese machines and over-priced closed-source/hardware
American machines - or other Open designs with too-small bed sizes and
underpowered lasers - I wouldn't touch another laser cutter for my
business. Having a full understanding of how every single nut and bolt,
mirror and lens, wire, switch and resistor works is a fantastic thing for
any heavy laser cutter user - and not having to pay a small fortune to get
a repair tech out to fix our machines is a hell of a win.

> The lack of a contact email address on the lasersaur website has forced us
> to post this publicly in the hopes of getting the attention of nortd labs.

This is an open project - we discuss things here - and we seek community
feedback on problems.

As a hackerspace member you should be EXTREMELY familiar with how
OpenSource projects work...when you have problems you bring them to the
community - and we work together to resolve them. If your Linux PC
crashes - you shouldn't expect to be able to talk with Linus Torvaulds and
get his personal email address to have him help you out. You talk to the
mailing lists and MANY people are right there to help you. Same deal
here. Don't expect Nordt to be the ones who help you...they started the
project - but now it 'belongs' to everyone.

> We're seeking a refund of the purchase price of the lasersaur board. If we
> can't get a refund we see no justification in leaving our efforts
> accessible on GitHub and will be deleting the repository.

Well, I think you're *WAY* off-base here. I don't know what you're doing
wrong - but it's not the Lasersaur software that's the underlying cause of
it.

I suggest you back down your rhetoric a bit and let's visit your problems
one at a time.

These machines *DO* work - and they work reliably, accurately and well.

-- Steve

[tan ken]

Hey Guys,

Hope this finds you well! Is anyone on this group from/in Singapore, Im looking to do a build myself. 

I have a space to tinker around which we can use to do a build.

Give me a shout on this email address!

Ken 

--