Various bits of hardware or a small upgrade to an engraving machine

A few months ago, I published a review of a small engraving machine. In that review, I pointed out the design flaws and mentioned that I planned to fix them.
In this review, I’ll cover the parts for the machine, add a little theory, and show the upgrade process.

First, I should say that this is the first part of the upgrade. I plan to continue, but I don’t know when that will happen yet. The point of this review is to show what’s needed and why. I apologize in advance to the CNC machine gurus, since I’m new to this and some of this was new to me too. So, I found some of the information online, and some comes from my own experience. If you have any corrections, please share them—I think it’ll be useful for everyone.

The title mentions one item, but in fact I ordered three items, one of them in two versions. But let’s take things in order.

I received an envelope packed with a whole bunch of little bags and packets.


I’ll start with the lead screws.
I ordered three of them. They’re all the same diameter, 8 mm, but come in different lengths: two are 200 mm and one is 100 mm. There’s a link to them in the review title.


Each lead screw comes with a nut, so the whole set looks like this.


Screw drives are commonly used to move machine mechanisms. The screw and nut convert the motor shaft’s rotary motion into linear motion of the nut, which is attached to the mechanism being moved.

In the simplest version, it’s just a long screw and a nut. But when I reviewed the machine, I noticed that the manufacturer had used a less-than-ideal type of screw—or, more precisely, thread.
This thread is intended for fasteners such as screws, bolts, and nuts, but it’s not very suitable for use as a lead screw in a mini machine.


There are quite a few types of thread: metric, imperial, coarse and fine (precision). For example, an acme thread is used in vises and presses—in other words, where the force is applied in one direction.
You can see examples of other thread types in this picture.


Trapezoidal or square threads are commonly used for lead screws.
Besides their names and profiles, there’s another small difference: unlike trapezoidal threads, square threads aren’t standardized. In theory, you can make them to your own dimensions, but it’s recommended to stick to the same dimensions and pitch as trapezoidal threads. That’s why I’ve shown them together.

Trapezoidal


Square


I used to think that lead screws with a square thread were the simplest type of drive used in small machine tools, but experience has shown that they fall somewhere between ordinary threaded rods and expensive ball screws.

Speaking of ball screws, in my opinion they’re the best option for CNC machines, but unfortunately also the most expensive.
They use a lead screw with a complex thread profile, which is usually ground as well, since it forms part of a linear bearing.
The nut also has a return mechanism that sends the balls back to the start, completing the loop.


It’s easier to see how it works in motion.



It’s worth noting that screw drives aren’t the only way to convert rotary motion into linear motion; gear drives are also used, though they’re even less common in small machines. A gear drive requires a more powerful motor, but can also provide higher travel speeds.


Here’s what a square-thread lead screw looks like in real life.


The kit included matching nuts with holes for mounting them to the mechanism.




The photo shows the difference between the ordinary threaded rod used before the upgrade and the purpose-made lead screw. I think the difference is obvious.


In my previous review, I also complained that the kit came with ordinary couplings to connect the motor shaft to the lead screw.
On the one hand, they have some advantages, which I’ll get to below. But there’s also a major drawback: for everything to work properly, the motor shaft and lead screw have to be precisely aligned. Even the slightest misalignment gradually wears out the mounting, and the threaded rod starts to slip out.

I couldn’t get one of the axes aligned properly at all, so I had to wrap a bit of tape around it.


That’s why I ordered some purpose-made couplings.
In my case, the motor shaft is 5 mm in diameter and the lead screws are 8 mm, so I chose 5x8 couplings, link to the product in the store; they cost $3.80.


Other diameter combinations are available, from 3x3 to 10x10, including intermediate sizes.
But keep in mind that in my case the shaft diameter matched the thread diameter. Strictly speaking, that’s not right—the ends should be unthreaded—but I ordered what was available.
Just be careful when choosing a set.


There are four types of couplings:
Rigid
This type of coupling is inexpensive and highly reliable, and can transmit high torque, but it does nothing to compensate for misalignment.

Helical
Essentially a spring, this type is moderately priced and can compensate for fairly significant misalignment. However, it cannot transmit high torque, is more fragile, and its springiness can increase backlash and cause resonance.

Jaw
Can transmit fairly high torque, but compensates for misalignment less effectively than a helical coupling. The flexible insert can also wear out.

Diaphragm
Can compensate for significant misalignment while still transmitting high torque, but unfortunately, it is the most expensive type.


At first, before commercially made couplings were readily available, people even used homemade ones made from whatever materials they had on hand, such as a piece of high-pressure hose. These can work perfectly well, but they are less practical to make and don't last as long.


I ordered one coupling for each motor, since I decided to modify all the axes.


These are helical couplings. As you can see in the photo, they're essentially springs. I first encountered one of these couplings while building a 3D printer, and was a little surprised to find that it bends :)


In the original design, the free ends of the lead screws simply hung in the air. That was acceptable with rigid couplings, but completely unsuitable with helical ones.
As I mentioned above, a helical coupling is essentially a stiff spring by design.
That very property causes an unwanted effect: backlash. If the end of the shaft isn't secured, you can move it by ±2-3 mm, which is a lot and would undo all the effort put into improving the design.

I ordered bearings to secure the shafts, link to the product, price $4.20.


These bearings come in different designs, with either a through-hole or a blind hole, and with different mounting options. I decided to order the through-hole version with an 8 mm bore. In my case, 8 mm is the diameter of the lead screw. If the shaft has a reduced end, the bore should match the diameter of that end.
They have sealed bearings, so dust shouldn't be a problem within reason.
There are two screws to secure the shaft.

Testing showed that they don't eliminate backlash completely, though they bring it close to zero.
I'm no expert, but I think bearings intended for this application should have a somewhat different design, since these are better suited to a different operating mode, rather than the axial load they're subjected to here.
Still, they're much better than nothing.


All three sets, as they'll be installed on the machine.


That's the brief introduction to the parts involved. Now let's move on to the modification itself.


First, I removed the shaft that moves the table. The photo shows the original parts and what I planned to install.


Couplings


Nuts.
I'll spend a little more time on this part.
As I showed in the previous review, plastic nuts were used, two per axis, separated by a silicone washer.
This design doesn't last long, but it provides almost zero backlash at a fairly low cost. My initial thought was to replace the plastic nuts with metal ones, but I decided that would cause the shafts to wear instead. Although the kit included three spare nuts, I decided that replacing the whole set would be the better solution.


The thing is, backlash in a lead screw drive is almost unavoidable, so it's reduced in various ways, for example with a special design and additional springs.


Another option is a split nut, which is essentially equivalent to the two nuts supplied with the engraver.
In this case, an extra screw tightens the nut and reduces backlash. Just keep two things in mind: the tighter you make it, the less backlash there is, but the more force it takes to turn the mechanism, and the faster the nuts wear.


For now, I decided to leave it as is and live with the small amount of backlash. I may improve it later.

The new lead screws were slightly longer than the originals: 199 mm versus 189 mm. That isn't critical here.


I started with the table, since it was the easiest part to modify.
I swore quite a bit, because I had to drill out the hole to fit the nut. And not just drill it out—I also had to enlarge it a little by hand. I drilled it to 10 mm (actually 9.9 mm), while the nut has an outer diameter of about 10.2 mm.


The hardest part was installing the nut, especially since the table was in the way. The previous nut had been trimmed on one side.


I used a multitool to cut a piece out of the table :) Then I unscrewed the front section and marked the mounting holes for the nut. It has four mounting holes, but since the previous one had three, I couldn't manage to use all four no matter how hard I tried. Two would probably have been enough, and might even have been better, but I decided to use three.


Judging by the design, the nut is supposed to go into the hole with its longer section first, but I did it differently. Now it faces inward.


As you can imagine, I didn't do this for no reason—it was «for a reason». This gave me a little more travel along this axis.


There were no problems with the coupling: I removed the old one, positioned the new one, and screwed it in place.
The motor shaft has a flat machined along it, so part of its length is cut away. It needs to be mounted so the screws don't land on that flat.


I test-fitted the bearing. The shaft was longer than the previous one, but it worked out perfectly: the extra length came in handy, since the shaft needs to be secured in the bearing.
I had initially thought about enlarging the hole for the bearing, but it turned out to be so large that it would extend beyond the metal frame it sits on.
Just to be safe, I enlarged the shaft hole from 8 to 10 mm.


The rest was straightforward.
I drilled out the hole for the shaft.
I marked the mounting holes. For this, you need to move the table as close to the bearing as possible so the shaft lines itself up in the right position.
Drill a couple of 3.2 mm holes and tap them for 4 mm screws.
Position the bearing so it doesn't interfere with the shaft's rotation, then tighten the fasteners. The original holes are 5 mm in diameter, but I decided to leave some «wiggle room»—it's more convenient that way.
Finally, secure the shaft using the bearing's own screws.


After that, things got both easier and harder at the same time.
Easier because I already had a rough idea of what I was going to do, and harder because I had to take part of the machine apart. Since I'd tried to do everything neatly up to that point, I had to do some of the work without disconnecting any wires. Sometimes the motor has a connector, which makes things easier, but I could only disconnect the spindle motor.


Next, I modified the axis that moves the mechanism left and right. The process was the same as for the table axis, just a little easier.


While I was at it, I raised the Z-axis motor a little to increase its travel. The new lead screw was longer than the old one, too.
I raised it using standoffs from some motherboard kit.


Then I hit a snag. The X-axis was misaligned—badly enough that even the coupling couldn't compensate. Once the shaft was secured, it was very hard to turn.
I had to take everything apart again, sand down the surface under the nut, and put it back together.


The X-axis bearing fit without any problems, just like the one on the table axis.


I ran into a small snag with the Z axis. I wanted to get the maximum travel here too, but didn't think about the bearing possibly getting in the way of the workpiece.
Also, even after raising the motor by 5 mm, the lead screw still turned out to be noticeably longer than necessary.
The second snag was that, for some reason, the hole for the coupler was made 20 mm in diameter (the coupler itself is 19 mm), while the holes on the X and Y axes were 22 mm, so the coupler was an extremely tight fit… A step drill did the trick. I couldn't think of any other way to enlarge a 20 mm hole to 22 mm, and it turned out to be very convenient.


I decided I could saw it off later if it got in the way. In the photo, the Z-axis mechanism isn't at its lowest position; it's raised by about 10 mm.


But once I installed the milling bit, I figured it was fine, especially since I plan to replace the chuck, and then I'd need more upward travel rather than downward.


This is how it turned out. It's not perfect, of course, but in my opinion it's much better than before.

After assembly, I lubricated only the lead screw that drives the “head.” It's better not to lubricate the others, since dust will stick to them and ultimately make things worse.


That wrapped up the machine's “hardware” upgrade, and I moved on to the “software” part of the project.
But I had a little surprise waiting for me there.
It all powered up right away and worked without a hitch, except for one thing. For some reason, I thought the speed would double, but it increased eightfold!
While inspecting and modifying the machine, I noticed that the screws had a four-start thread. You can see this clearly by looking at the end of the screw.
The photo shows the four thread starts.


Just to clarify: a four-start thread is a type of multi-start thread.
Ordinary screws and their nuts have a single-start thread, meaning the thread runs like one continuous strand.
With multi-start threads, there are several “strands” at once. This design allows for a larger thread pitch while maintaining high strength, since several thread turns are engaged at the same time.


So the old screw had an 8 times lower feed rate, but in return it provided 8 times more feed force for the same applied torque.
It was slow but powerful. In my previous review, I complained that the machine worked very slowly. But besides the eightfold slowdown due to the different thread pitch, I also had a twofold slowdown because the stepper motors have a 0.9-degree step angle, compared with the more common 1.8-degree motors.

The photo shows how many turns of the old screw's thread correspond to one turn of the new one.
But the old screw had another advantage: it essentially didn't need the shaft-braking mode. The mechanism could only be moved by the motor; with such a fine pitch, turning the shaft by hand wouldn't move anything.


That's about it, in broad strokes. Here's a short video demonstrating how it works after the upgrade.
You can see how the engraver worked before the upgrade in the review.


Screenshot during operation. For the test, I used a sample file from the PCB engraving software.


A photo of the result. Here you can see one of the problems with engraving. The slightest unevenness causes the bit to cut too deep, or, conversely, pass above the workpiece. And I tested it on a piece of laminated particleboard, which is very flat, unlike fiberglass PCB material. In other words, the material needs to be prepared very carefully. For example, cutting fiberglass PCB material with scissors deforms it, and getting it flat again is quite a challenge.
On the other hand, the mini engraver is great for drilling holes in boards. You do need to replace the chuck with a decent one, but that's for later.


A closer look. The bit wasn't quite sharp, so you can see cutting defects at the top left. Also, the tip of the bit should taper down to a point; in my case, it didn't quite do that.

But milling boards isn't suitable in every case either. If you're making a board for some low-voltage device, everything will be fine. But if you plan to build a high-voltage device, you may run into problems. And if you need galvanic isolation, it's better to etch the board, since a breakdown between copper pours is dangerous.


Today I won't list the product's pros and cons; I'll just talk about it in general.

The couplings, screws, and bearings themselves are nothing special compared with others. But it's pretty hard to mess these up. I can say the screws are straight, though I only checked them against the edge of a ruler.
In my opinion, the option shown here is somewhere in the middle of the options covered in the review. Again, only the rapid-traverse speed has increased, since feed rate depends on the material being machined. So there may be cases where the speed is no different at all from the first option.
I think this configuration is perfectly fine for a small machine, but for anything bigger, a ball screw is a better choice.

And now, here's what I ended up with.
The upside: higher rapid-traverse speed and greater reliability.
The downsides: noticeably more noise, a little backlash, and less torque.

All in all, it pretty much evens out.
For now, I'm planning to rework the bearings and chuck. Once I run out of ideas, I'll probably start building my own version of the machine, this time with ball screws and diaphragm couplings. I think that combination is better suited to precision work, though it does come with a different price tag :(

That seems to be all. I'd be very interested to hear from more experienced folks what else could be used on a machine in the 2040–4060 format (for example), since I'm still weighing my options.

The product was provided by the store for review.