What's new
What's new

How to quantitatively measure straightness for a desktop CNC's columns?

rajhlinux

Cast Iron
Joined
Jul 17, 2020
Location
USA
Hello,

I have a mini desktop CNC, how can I quantitatively measure the straightness for the columns?

It will be primarily used to make PCBs using Direct Laser Imaging (DLI) photoresit lithography.
The laser module is super light, possibly around < 227 grams (< 0.5 lbs) and low powered only about 25mW.

Want to make sure everything is perpendicular without using eyeballs, everything done on a five-tenths (0.01 mm) test indicator.
Everything should be perpendicular squared to around ~3 - 4 tenths (~70 - 100 microns).

Want to do this in a low cost manner.

I'm thinking to place everything on a 12" x 18" x 3" granite surface plate.
Super glue the 1" end faces of a 4-sided all precision ground 1" x 2" x 12" parallels (a pair of two matched straightedges) while placed on the surface plate, this would give a 24" precision ground straightedge and use this as a square reference.

Then a tall surface gauge with a test indicator to measure different height points of the CNC columns.

Is this method a doable improvised "Sqaureness Master"?

Anyone have any better ideas?

If you may wonder to why is it that I'm going into this hectic quantitative measurement process, it's because so that I can also learn how to do this meteorology stuff, mainly due to the fact that I'll soon build a much larger precision based CNC made off of high performance concrete using linear rails.

I will first test tolerance accuracy for each aluminium extrusion on the surface plate with a test indicator.
If there are any bows or twists, I'll simply hand scrape them which isn't difficult to do.

Not sure how to check the squareness and correct it for the side cuts of the aluminium extrusion.

Thanks for any advice.

I have attached a picture of the mini CNC:

IMG_3352_2.JPG
 
Last edited:
Your wheels will not chase tenths. Lasers are really close to material, the depth of your material is small, let good enough be good enough. Using a drafting triangle will get you closer than stacked and glued rule and be an order of magnitude better than wheels can hold. I am very pro stepper motor, without reduction you are looking at .001 true accuracy range with that drive - at best. Reduction cost serious money at the level you are asking for. Fine pitch Ball screw drive is the cheapest way i know of.
 
Your 'mini CNC' sure looks a lot like an Ender 3 :LOL:

You've been a member here long enough to know that most of the members like to keep the content focused on professional endeavors and discourage hobbyist projects. Posts like this are ripe targets for pile ons, other forums are more friendly to this kind of thing.

If you can keep the question focused and it has enough applicability to 'real' manufacturing work, you can get good answers, but reading your question(s) I don't think you are going to get the support you want. Can you glue things to a surface plate and get certain accuracy with unknown quality references and limited metrology gear? How can we answer that? Is it a good idea? The fact that you will have a hard time finding examples of such a strategy might answer that unasked question.

As a homebuilt CNC guy myself, scraping anodized aluminum isn't "isn't difficult" nor is it practical. Live with the error, or find a different material. If you are short on money, gluing things to your limited number of reference surfaces seems short sighted and not the best solution. And lastly, I have a vast number of highly accurate metrology devices bought on the second hand market. Cost shouldn't be an excuse for lacking the needed tools to achieve your goals.
 
Make a little bit of an effort first. There's a Moore book, Elements of Machine Geometry or similar. Definitely similar, that's not the actual name but enough to start a search.

Read it.

Then if you still have questions or ideas, propose them. Don't come up with a bunch of harebrained stuff and expect people to waste their time on it.
 
Your 'mini CNC' sure looks a lot like an Ender 3 :LOL:

You've been a member here long enough to know that most of the members like to keep the content focused on professional endeavors and discourage hobbyist projects. Posts like this are ripe targets for pile ons, other forums are more friendly to this kind of thing.

If you can keep the question focused and it has enough applicability to 'real' manufacturing work, you can get good answers, but reading your question(s) I don't think you are going to get the support you want. Can you glue things to a surface plate and get certain accuracy with unknown quality references and limited metrology gear? How can we answer that? Is it a good idea? The fact that you will have a hard time finding examples of such a strategy might answer that unasked question.

As a homebuilt CNC guy myself, scraping anodized aluminum isn't "isn't difficult" nor is it practical. Live with the error, or find a different material. If you are short on money, gluing things to your limited number of reference surfaces seems short sighted and not the best solution. And lastly, I have a vast number of highly accurate metrology devices bought on the second hand market. Cost shouldn't be an excuse for lacking the needed tools to achieve your goals.

Where did you buy second hand Metrology device in the second hand market? Well I can not determine if I bought one is indeed a valid working one or authentic. Nor do I have the time for this cat and mouse game of buying and returning.

Therefore, I simply bought a brand new 5-tenths $35 test indicator which gets the job done with hundreds of reviews from amazon stating it's gucci but not the best, gucci is good-enough for me taking quantitative measurements.

I consider myself an apprentice in the trade.

No need to play with $500 metrology tools that I would damage easily, along the way in the years I would buy good $500 metro tools, if ever needed.

Also the CNC is an Ender-3 V1.

Remember laser works with light (photons, practically have no mass) its not milling anything. Don't be surprised once I soldered extremely tiny and difficult QFNs and FPGA chips thanks to a $250 CNC...
 
Last edited:
Make a little bit of an effort first. There's a Moore book, Elements of Machine Geometry or similar. Definitely similar, that's not the actual name but enough to start a search.

Read it.

Then if you still have questions or ideas, propose them. Don't come up with a bunch of harebrained stuff and expect people to waste their time on it.

The book is called "Foundations of Mechanical Accuracy". I actually have a bootleg PDF copy of it, not an enjoyable read.

However if you can send a fresh paperback copy of it, I would gladly read it. Costs $300-$500. Otherwise I would have to ask in forums and talk about it. I have seen youtube videos of home made cheap CNCs to make PCBs using LDI litho tech with excellent results. However these nerds never talks about it. Just shows off their skills.

Therefore Ender-3 is a starting point on how to learn and build CNCs before venturing into the big players of CNC parts and making precision CNCs. Consider it as an apprenticeship self teaching kit.

The main concept is to properly assemble the CNC, I have bought a granite surface plate (grade B) 12" x 18" x 3", it will arrive tomorrow.
 
Foundations of Mechanical Accuracy by Moore. You can find free downloads in PDF.

Yes I have a copy of it, however it's impossible to read it. The person did a terrible job in scanning it. I downloaded from like 5 different sources and all of them are from the same person who did the scan.

I will never read it. I find it insulting and non-educative to read such a great book in an unenjoyable manner.

Can't find a fresh copy of it.
 
Where did you buy second hand Metrology device in the second hand market? Well I can not determine if I bought one is indeed a valid working one or authentic. Nor do I have the time for this cat and mouse game of buying and returning.

I buy them from internet auctions, but it sounds like you aren't willing to put in any effort to take advantage of such a resource so there's little point in explaining my methods or experiences.

Therefore, I simply bought a brand new 5-tenths $35 test indicator which gets the job done with hundreds of reviews from amazon stating it's gucci but not the best, gucci is good-enough for me taking quantitative measurements.

Good thing reviews on Amazon can't be manipulated and that they have excellent control of the product actually being shipped :rolleyes: But everyone needs to start with something.

I consider myself an apprentice in the trade.

Yeah, we're all with you on this point, although you may be giving yourself a little too much credit. An apprentice has a Mentor, in your case his name seems to be Google.

No need to play with $500 metrology tools that I would damage easily, along the way in the years I would buy good $500 metro tools, if ever needed.

Maybe take care of your stuff. And getting $500 tools for $50 dollars used but in good condition takes a lot of the stress off, but again you seem to think it's better to cheap out on questionable stuff. That's your call, but when things seem out of whack and you ask for help here, and we find out you are using untrustworthy tools and methodology, you are not going to get anything but derision.

Also the CNC is an Ender-3 V1.

I know exactly what it is, easily identified as a first gen by the 20x40 Y rail. It is why I was laughing at you calling it a "CNC" which even by the silliest of definitions it most certainly isn't and never will be. It isn't even a very good 3D printer, and without serious modifications lacks the capability to work well as a precision anything. I'm not sure how you think it will help you on your path to a homebuilt CNC, but you do you.

Remember laser works with light (photons, practically have no mass) its not milling anything. Don't be surprised once I soldered extremely tiny and difficult QFNs and FPGA chips thanks to a $250 CNC...

"Remember laser works with light"? Are you serious? I'm thinking this whole post is a troll. No one is this dumb.
 
Most of this thread is 'sparring' which is neither interesting nor helpful.

Going back to the OP's questions: In your shoes I would attach a dial test indicator in the place of the laser head. Then buy or borrow a cylinder square of the correct size, place it on the table, and run up to it from different directions, at different heights. That would give me some good idea of repeatibilty, squareness in Z, and so on. Then get a small granite square, lay it on the table and do the same. For calibration in X, Y and Z I would do similar experiments with gauge blocks.

I think this might not be a formal calibration, but would give you a reasonable understanding of the repeatibility, accuracy, and squareness.
 
Last edited:
Yes I have a copy of it, however it's impossible to read it. The person did a terrible job in scanning it. I downloaded from like 5 different sources and all of them are from the same person who did the scan.

I will never read it. I find it insulting and non-educative to read such a great book in an unenjoyable manner.

Can't find a fresh copy of it.

PM me here with your email, I’ll send you my copy.

Better check it now that you mention it. But I remember being able to read it, even after a coupla three beers.
 
No that Moore Tools book does not cost $300-$500, you can buy it directly from Moore, it's been awhile, but mine was about $100. Ok you have crappy scans of the book, its a PDF and can be read online if your scans are that terrible. If I needed the information bad enough, I'd do whatever the hell I had to and get it one way or another. Own a half decent lathe? Then just make your own cylindrical square as long as you need that's easily checked for accuracy with a micrometer. Its also a self checking item, and even an inaccurate one will be dead square on at least 2 points of its circumference if there's no side taper. My B&S 558 cylinder square is purposely made with one end tapered so it leans, and its still accurately square at two points. Ok making your own won't be hardened and ground to less than 10ths unless you've got tool steel, a proper oven and at least a tool post grinder, but it should still be good enough for what your wanting to do.

They boot strapped the industrial revolution with hard work, there own intelligence and some basic logic before any of what we now have that's almost at our finger tips were even invented. If Gascoigne, Watt, Whitworth ect were still alive, they wouldn't bother even posting the question. Ask yourself how anyone of them would do it.
 








 
Back
Top