Kern Microtechnik: Accurate CNC Machines [video] | Hacker News Reader

Kern Microtechnik: Accurate CNC Machines [video]

youtube.com

134 pointsSQL221967 comments

Once you get down to below 1/1000 of an inch or so, you start to notice that metal is really like playdough. it squishes out in all directions when you clamp it, it springs back when you cut it. it is never exactly the same shape after it has been handled.

Just another thing that adds to the difficulty of super high precision work.

What do you use to observe such phenomenon? Optical microscopes?

Often using a dial indicator. A rack and pinion converts the linear to rotary motion, and gears magnify the rotary motion:

https://en.wikipedia.org/wiki/Indicator_(distance_amplifying...

tubalcain has a good video showing how they work:

https://www.youtube.com/watch?v=Rp6D3YTx7iU

You can use an optical flat. Interference fringe patterns will tell you how a surface that was cut perfectly flat on a machine becomes bowl shaped when you unmount the part.

https://en.m.wikipedia.org/wiki/Optical_flat

Yup. I often refer to aluminum as “rubber” when chasing that last few thau while machining. I do this to remind myself and others to have the right perspective and expectations when approaching things at this scale. Just because the machine displays four digits after the decimal point it doesn’t mean you can actually cut at that level.

For a slightly more accurate CNC lathe and grinding machine (1 micron) that can be made "at home": https://www.youtube.com/watch?v=sFrVdoOhu1Q

Yeah, that seems to be for values of "at home" that only includes the homes of expert CNC machinist with access to someone to write custom software for it. Otherwise, the accuracy is impressive, though it doesn't seem to approach the level of automation seen in the Kern machines.

Yes, this is not any ordinary person in any ordinary home. His home also happens to be a world class prototyping facility.

I don't think Dan is interested in more automation though. It typically only pays off for big production runs, especially compared with how fast he is.

Dan Gelbart's other videos on prototyping are also definitely worth watching.

I've been pretty bummed on "tech" lately because software is such a shit show so this was impressive and inspiring, both the machines and the seeming quality and focus of the employees.

Yeah, everybody is rooting for advancements in hardware/mechanical until they see the salaries this field pays.

I'm surprised the field is advancing at all seeing app kiddies can walk home with more $.

This was my industry for 40 years and I got to watch it die.

You made me curious. According to the U.S. Bureau of Labor Statistics (bls.gov):

- Metal and Plastic Machine Workers make $17.35 per hour or $36,080 per year

- Machinists and Tool and Die Makers make $21.61 per hour or $44,950 per year

- Software Developers make $50.77 per hour or $105,590 per year

Employees at a top notch company like Kern Microtechnik probably make more than the machine worker average aboves, but then software developers at FAANG make more than the developer average as well.

It's sometimes astonishing that there isn't more correlation between skill and craftsmanship and how much you get paid.

The imbalance is not just in take-home salaries but in the barrier of entry as well.

You can't just get into the field by being a self-taught mechanical or electrical engineer without investing the time and $ in attending a proper university but you can call yourself a software "engineer" after a boot camp and start making $.

The only advantage of these fields though is that you can (still) find work in your 60's as a gray beard while if you're in web dev, hopefully, you have moved far enough up the ladder or made fuck you money until your 60's because I don't expect to see many graybeards doing React in the future.

In Germany (home of Kern microtechnik) the gap isn't that huge. Yearly avg. salary of a software developer is around 70k while a skilled machinist can make around 50k.

Average software developer pay isn’t 70k in Germany. It’s way lower.

As a machinist of 35 years I can say your pay scale is WAY off. Just look jobs on Indeed.Com.

Trust Me, A Tool and Die Maker makes much much more than you quoted. Problem is everybody wants to learn whizbang instead of getting a foundation of conventional machinery.

What cracks me up is I see CNC programmer jobs paying less than Tool and Die makers when programming requires computer knowledge.

The real problem with the trade is Vo-Tech is dead and kids can't even swing wrenches these days, let alone the strong mental math skill required.

BLS numbers are higher than other poster says for Tool and Die makers, but less than you imply.

https://www.bls.gov/oes/current/oes514111.htm

It's probably a thing where there's people BLS considers tool and die makers that you aren't roping into your expectations.

Could you give hard numbers for what you would consider reasonable pay?

I know a tool-and-die maker that works on high volume plastic injection molds that makes a comfortable multiple of my salary. This is very precious knowledge and having it can save a company many millions of dollars in re-tooling and discarded product.

A number would be nice.

The rest of the world outside web/cloud is still trucking along quite nicely and doing a ton of exciting stuff. :) Even web stuff has a lot of good things going on, it's just that there's a ton of derivative fluff along with it (as always when you're the hot new thing).

Big fan of NYCCNC, and have been anticipating this video for awhile since he first mentioned it. Too bad it's not as in-depth as some of their other factory tours. Clearly Kern has some secrets they're not keen on letting slip.

Check out their Starrett tour: it's one of my favorite videos on all of YT.

https://youtu.be/T7JqXbm-Nwo

Thanks. I just spent over an hour watching a variety of videos on YouTube about CNC.

Found out about them through Nick and Josh Hacko of https://www.instagram.com/nicholashackowatch/. They use a wire EDM made by Kern to manufacture watch parts.

If you're into this kind of precision, https://www.instagram.com/robinrenzetti/ and https://www.instagram.com/stefan_gtwr/ are great to follow.

2 um accuracy; tools so fine that tool change air flow will break them; amazing. I could use this for machining mmWave components.

I've been acquiring tools to fix/measure/improve cheap CNC mill and 3d printer, the $100 dial gague I picked has 10 µm as its smallest increment, you can just barely see a 2µm variation. The force the measurement tool imparts on the part and its support structure is more than that.

The next steps are either carelessly blinding myself trying to build a laser interferometer or spending a ridiculous amount of money on more precise gagues.

Sam’s lasers sells some refurbished HP laser interferometers on eBay (HP5517 head+the receivers). He has also worked with someone to supply open source software/counting electronics. My notes on the kit are here:

http://41j.com/blog/2017/02/hp-interferometer-notes/

You can often pick up the laser/receivers/optics quite cheap (few hundred USD). And then just use the software supplied by Sam. The Sam kit is tested which is nice (but it’s ~1000USD IIRC).

Interferometers as used on high end CNCs often.

There is a cool effect where a laser reflecting back into itself will cause interference inside the laser diode and affect the power required to drive it in a measurable way and with it you can extract a surprising amount of information (distance, rotation, angle, speed, etc.)

I was ... hasty ... with my first laser diode and burned it out in about 30 seconds. I like to tell myself the money wasted was spent on learning lessons. The second is coming in the mail.

The upshot with this method is that it can be extremely low cost and give you access to write code that interfaces directly with physics instead of somebody else's device.

The idea for hacking on these things comes from Ben Krasnow https://www.youtube.com/watch?v=MUdro-6u2Zg and several research papers.

This is known as a self-mixing interferometer (as is explained in the video)

https://en.m.wikipedia.org/wiki/Self-mixing_interferometry

The black matrix glass in an LCD is typically accurate to a few microns since it's made in a photolithography process. If you couple that with a microscope/camera (e.g. mounted where your tool usually goes) you can use that to calibrate your CNC mill across its work area with some machine vision software. This is probably going to be a lot cheaper and simpler than using interferometers and for any practical purpose should have better accuracy than what you can expect to ever get out of a regular CNC mill. I'm not sure if you can just buy those somewhere or you'd take a display apart or something along those lines...

Glass scales for DRO are relatively inexpensive. The Bridgeport EZ-Trak I bought at work uses them in the servo system. So getting static accuracy is probably not too difficult, but the dynamic accuracy is a whole different can of worms.

I can’t remember who wrote the white paper, but it described “rounding a square”, and measuring how much tangential material was left over at the flats, as a test of dynamic accuracy.

Even static accuracy is not so easy.

The scales will measure the axis position. But when you move the table over to the other end of travel, the shifting weight tilts the knee, and now your cuts that were supposed to be parallel are not. None of the scales measure this rotation.

Then, as the machine warms up, differential thermal expansion starts to twist things further...

My interests are more in general purpose tools for increasing accuracy, and perhaps using them to increase the accuracy of inaccurate tools through feedback systems.

That is, throw out chasing perfection of rigidity, dimension, and flatness of a mill (etc.) Instead keep adding measurement and feedback until the precision of the machine parts doesn't matter. (or, continue to amuse myself and annoy my partner by making messes and spending my disposable income trying)

It is, in my opinion, no coincidence that backlash and battiness both start with a B. Keeping up with (semi) predictable flexture would be... well, not a doddle as such, but... if you didn't also have to deal with backlash. Sure, there's a point where everything is, to a first approximation, made of rubber, but it's a lot easier to move that point by modifying or improving the tool than trying to trim the end of a diving board with a laser-assisted pole-mounted pruning saw.

My thinking goes, the end of the road is always "everything is rubber" when it comes to precision, so I might as well start off with that reality and see where it can take me.

On the other hand: wear. Even the best machines all experience wear all around and there is a lot of work keeping up with it or recovering from it. Keeping precision despite drift from wear and being able to trivially quantify it are really valuable.

The very best machines don't really ever wear out. The surfaces mate over the full area so even sliding bearings experience hydrodynamic contact. Or, they're built with fluid bearings to begin with.

(Assuming they're handled properly and kept clean, of course).

Having tried this, with some fancy measurement tools and a pretty custom and high-performing control systems, I've concluded that it's a route to madness. Fancy software and feedback turn out to be a very hard way to emulate large, straight pieces of cast iron, and there's no substitute for rigidity once you actually start cutting - vibration is a huge issue, and very tricky to actively damp. Essentially every machine tool that does have crazy accurate realtime feedback (diamond turning for optics, ultra precision milling, etc) starts with a rigid and long-term stable frame and adds the $$$ controls.

Granted, my switch over to "more rigid more better" may have gone a bit too far - I now own a 10ee (3200lbs), and am looking for a good jig borer (2400lbs+).

You're right that one can't compensate for lack of rigidity with software but you can have an inaccurate but rigid structure and compensate on top of that. It's pretty common practice to use some sort of software compensation to improve accuracy (at least in certain applications where it matters).

Metal milling machine need to be super rigid because of the cutting forces... You can build super accurate machines (e.g. with granite) that can't be used for milling but can place a tool with a (sub-)micron accuracy.

Even with metal you can take lighter cuts and trade off some rigidity for accuracy, but then it's gonna take much longer to get anything done...

An inaccurate but rigid structure, unfortunately, won't hold its calibration.

Go for the jig borer, and get them while you still can. It's what Professional Instruments does, and it's what I would do if I had space for one.

It can. Basically a repeatable machine can be made accurate by calibration. It doesn't have to be accurate to start with. If it's not repeatable (i.e. won't hold its calibration) then yeah, it won't ;) But accuracy != repeatability. I've worked on precision machinery (1um accuracy) and we used calibration techniques.

This is also my field. The issue is, the machine will be repeatible on the short term but not on the long term. It will look like it's behaving well during calibration and the next few months of operation.

But if the ways and mating surfaces aren't almost-perfectly straight and flat, they'll experience accelerated wear. (If you're using hydrostatic bearings, they won't work to begin with unless the surface is accurate). Then the calibration is gone. And that's just in the static case.

If your ballscrew has uneven pitch, is eccentric, or any number of other issues, you can calibrate it out. But now to move at a constant speed, your servo controller has to drive that inertia at a wobble, and everything shakes.

The machines I worked on had air bearings. No wear. Basically we're talking about something like an error of 10um over 2 meters. It's practically impossible, or cost prohibitive, to remove that error without calibration. The granite surface the air bearings ride on is locally flat but not perfectly accurate over its entire length.

I agree this setup is almost perfectly straight and flat. But it's still not accurate without the calibration.

Just to be clear, there's weren't milling machines...

What kinds of control systems did you use?

My ideas are in the realm of developing detailed physical models of machines and developing control systems out of them using precise measurement tools to both tune the model and act as feedback mechanisms. (I'm not talking about machine learning)

I don't think my living room floor would support three tons of machine tools unfortunately.

As msds was saying, if you're e.g. looking to cut metal you need to have a rigid structure. Otherwise the tool just bounces around given the cutting forces which really can't be compensated for. You entire structure deflects and vibrates under those forces. Accuracy is a different story, if you have a repeatable but not accurate system you can certainly fix the accuracy via calibration or feedback (feedback can be what gets you the repeatability).

Not only cutting forces; the machine will also warp and deflect as the axes move around and shift its weight. Even when all the sensors read that the tool position is correct, there's 6 degrees of freedom and a lot of coupled error from the stack up of the errors that each sensor doesn't measure. You can't really have accuracy if the machine isn't rigid, or if it has any play. And even if you do a calibration, if the ways aren't accurate to begin with then they'll wear and change.

You can certainly correct for all these errors if you are able to measure them. E.g. if your tool moves 1um down in position (x,y,z) due to the machine deflecting all you have to do is adjust z by 1um at that point. As long as there's no other issues and the only issue is the accuracy of placing that tool you're good. One example is precision ball screws, those typically come from the factory with data for compensating the screw. The screw isn't perfectly accurate on its own but if you apply the correction data you get better accuracy. The screw isn't less repeatable because of this, it doesn't wear down faster, it's just that it's absolute accuracy isn't perfect out of the factory so they just measure that and give you the corrections (let's say on the order of 5-10um over 1m). In usage the accuracy is also a function of temperature, so you may want to measure that and correct for it... or you may use a linear encoder or an interferometer ... point is, the screw does not have to be perfectly accurate, it just has to be repeatable and the remaining error can be corrected...

What is your goal with improving the cnc machine?

I’ve used a capacitive distance sensor to measure vibrations in the micrometers and nanometers. At that scale things get unintuitive pretty quickly. I doubt your motors and linear tables will be able to keep up with much precision beyond what you can measure with a cheapish digital caliper.

Eh.

When I'm following my own drives I tend to build something up from basics and improve it's ability to do X until it reaches a certain level of usefulness and I get bored and move on because my goal was building the thing not ever actually doing anything in particular with it. That section of the journey appeals to me.

Ultra-precise laser range finders scratch a particular itch. I so often have found myself wanting one and being unable to find something suitable to buy. Let's say, 1mm to 50µm precision at 1cm to 10m distances, in that ballpark.

What could I do with it?

* Characterize and quantify improvements in machine precision, most of what I do seems to be wild shots in the dark without some better measurement tools.

* Add feedback mechanisms to machines – work around imprecise parts and close the loop with cheap and super-precise measurement tools.

* Lasers.

The key features are programability and range+precision.

Cool!

Here’s a couple thoughts I have, forgive me if you already have these in mind.

Feedback and Controls: I’ve found that feedback suddenly made a lot more sense and seemed less mysterious once I understood the math behind it.

This is the textbook I used: Control System Engineering - Nise

And these videos were particularly helpful: https://www.youtube.com/user/ControlLectures

Regarding Precise Measurement:

It’s hard, something that might not look moving to the naked eye might be vibrating back and forth several uM. I’ve heard that lasers are subject to error from the atmospheres but I’ve never used them myself.

Good luck!

If you look, there's an excellent paper on an academic team that did a low cost construction project on a small mill for micromachining. They achieved very high accuracy with fairly ordinary components.

Here's one link to the project:

https://www.google.com/url?sa=t&rct=j&q=&esrc=s&source=web&c...

If you want to work on software in this industry then check out https://www.moduleworks.com/join-us/ Moduleworks develops algorithms for many of the CAM platforms that target machines like KERN develops.

Why is it that Germany makes the best industrial machines and industrial manufacturing?

There is a strong market for such machines in Germany.

I am not saying they don't but in German machine shops I have worked a lot of machines were of Japanese origin such as Okuma or Mori Seiki.

IIRC from when I did metal work, that's mostly because the Japanese vendors deliver more machines per year.

German manufacturers generally don't make that many machines, one because they're used for a LONG time (I learned milling on a machine from the 60s in 2015), two because they are complicated to build. You can't easily order 100 high quality mills from a german shop.

Because they have an extensive framework of public policies to foster these kinds of industries.

If you want a cultural difference vs America to point a finger at: low ambitions. Even the most talented don't necessarily shy away from modest markets that could never buy them their own private island. Industrial machines is a high value market, but the total growth potential isn't so big.

With a typical European medium business (this is really not limited to Germany at all), if it achieves exceptional success in its particular field, chances are that it won't try too hard to leverage that success for growing aggressively into other fields, eventually losing focus on their core business. High specialization like Kern is achieved by sticking to what you do best and trying to get even better at it, not by trying to become a billionaire. Americans, by comparison, would be more prone to handing over the core business to the management equivalent of a maintenance coder and moving on to the next bigger opportunity.

Manufacturing in Germany tends to be more high-tech on average than America. America has a lot of smaller job shops that buys foreign machines, but Germany has large, highly automated manufacturing companies (esp car companies) that have both the capital and skilled labor to make use of high-tech equipment like this. Probably in part because Germany has better technical education, and worker protection that keeps jobs in Germany (sometimes).

Hold my bier. https://str.llnl.gov/str/April01/Klingmann.html

http://mooretool.com/jig_grinders.html

Really incredible machines. Additive manufacturing has gotten all of the hype lately because it's new and quickly developing, it's equally exciting to see cutting edge developments in traditional subtractive manufacturing.

I am burned out on software, would love to be a machinist. But I am not interested in huge pay cut. I am thinking about building a machine shop in my garage.

I did this, it's awesome. I recommend it.

I did this. Lots of fun. I switch between software and "hardware" in my garage. I work mostly with wood, but also aluminum and acrylic.

I took a side job in a machine shop over the winter.I was loading parts in and out of CNC machines, doing a little setup, cutting metal, made one fixture, and a lot of floor mopping. I thought the manual labor was going to make me appreciate the cushy desk job work that I normally do, but I absolutely loved working in the shop.

I wrote an email to friends and family about, hit me up via my email in my profile and I'm happy to share it with you.

I'm trying to find a similar part-time apprentice position in NYC where I live now - any suggestions?

I'm just disappointed that TechShop went bankrupt. At one point I wanted to open a facility in the Balt/Washington corridor. I figured there'd be a lot of people wanting to learn those skills, and an accessible location. They opened a franchise in Crystal City, VA (DC) but all closed in 2018. It was a a mega-makerspace with classes to learn all theses technologies.

What do one of these machines cost? What's the "build volume"?

Reply on news.ycombinator.com