Upgrading a Chinese laser for engraving

Hi guys.

Recently I posted something for a friend looking at getting into industrial engraving....which got me into testing some engraving on my shenhui, something I haven't done a lot of since it was new....cutting is the main use as it never seemed too fast or too good at engraving.

If you look back at my posts, you might also see that I haven't been super happy with the cutting accuracy of the shenhui either.....and I still haven't figured that out but just do more QC or use the CNC for the jobs that need the tolerance, I work around it.

Tying the two things together, after upgrading to closed loop steppers to try and fix the accuracy issues (didn't help).... I never tried to use that closed loop advantage to speed up the acceleration for engraving. Well, after posting about the industrial engraving, I tried some more testing and it *did* occur to me to up the acceleration. I took the stock 3000mm/s acceleration value and bumped it to 8000mm/s. no problems....which is to say no problems different than any other time engraving with the shenhui. 400-500mm per second speeds, about .6mm and single direction passes worked well at a lower power setting.

Anyways, it cut the time *way* down, at least 200% faster, maybe more. I will try do a run when tomorrow to quantify the improvement. So, long story short, if you have a chinese laser and want to engrave faster, spending $3-400 bucks to replace the XY with "hybrid" or servo-steppers from china is money well spent. Make sure the cables are long enough, shaft is the right size, etc. but they are pretty much drop in replacements.

Though I would share this as a separate post for those interested and as thanks for all the help when I ask!

Cheers, Greg

Still going to be using for more cutting than anything else, but I'm glad I took the time to experiment.

Very interesting, thanks!
Do you have a link to the supplier oif the hybrid stepper/servos, and the drivers needed for them, please?

ClearPath servos, integrated with the driver

Greg, Thanks for posting and its on the right day!

Very cool. Do you think you'll outrun the tube on some substrates and have to consider one with more power? Did you only have to increase speed in the settings?

Quote:

---

Originally Posted by Matt McCoy

Very cool. Do you think you'll outrun the tube on some substrates and have to consider one with more power? Did you only have to increase speed in the settings?

---

Or maybe the DC tube can't fire fast enough to take advantage of the speed? Dave would know.

FWLIW, when I do tiny wood inlays, anything over about 200mm and I can see that the leading edge of the pocket I am creating is not as deep as the rest of the engraved lines. Seems like the tube cannot ramp up quickly enough. On my Epilog I could run at 100% speed (I don't know just how fast that really is but pretty fast) and the leading edge was exactly the same as the rest.

Quote:

---

Originally Posted by Bill George

Or maybe the DC tube can't fire fast enough to take advantage of the speed? Dave would know.

---

Quote:

---

Originally Posted by John Noell

FWLIW, when I do tiny wood inlays, anything over about 200mm and I can see that the leading edge of the pocket I am creating is not as deep as the rest of the engraved lines. Seems like the tube cannot ramp up quickly enough. On my Epilog I could run at 100% speed (I don't know just how fast that really is but pretty fast) and the leading edge was exactly the same as the rest.

---

Good points about rise time on the tube. It seems there are penalties to be addressed by the stepper upgrade. Rapid speed might be a good place to reduce engraving and cutting times without much compromise.

Thanks for the post Greg.

The acceleration has little, perhaps zero, bearing on the limits of a DC tube. We all know a DC tube isn't as fast to respond or as a RF tube, nor as responsive for fine tuning lower power levels, but that is besides the point.

I was running mine at 500mm/s and 700mm/s for engraving speed. 500mm was ok, 400 mm/s little better, 700mm/s not so great, and others opinions and tube and power supply and control system may make those numbers different. Regardless, the speed improvement is the overshoot portion. 700mm/s no longer needs 3-4 inches to let the steppers get up to speed. The faster you try to engrave, the more that overshoot matters, to the point a fast speed can be counterproductive for narrow jobs. Upping the acceleration makes that all very tight....very much like the trotec videos on youtube. I have no doubt the trotec type machines are still faster yet, but I did make the test yesterday and a small job, actual job oddly enough, went from 2min40seconds with 500mm/s engraving and 3000mm/s acceleration, to 1m11s with same engraving speed and 8000mm/s acceleration. Not noticeably different in quality, which isn't surprising as the actual engraving part is the same. FYI, those numbers are .06mm vertical line to line gap and all passes in a single direction, not back and forth.

Also, side tangent. Clear Path servos are probably overkill and have distinct disadvantages in that there control signals can become compromised.....I had a CNT router that had (same brand, tecnic if I recall) servo drivers on the gantry....not quite like clearpath on the motor, but EMF from the control to the driver was corrupting the signal and adding virtual steps....clearpath would have the same issue near as I can tell and it's a total PITA to figure out the problem. The more cable that is closed loop, the better in my opinion.

I will try to post the details of what I bought, via aliexpress.....or maybe I will link to a US reseller.
Hmm, if I recall, links are a no-no here. Search 5080 at automation technologies site and a suitable kit will come up, just make sure the wires are long enough.

Quote:

---

Originally Posted by Bill George

Or maybe the DC tube can't fire fast enough to take advantage of the speed? Dave would know.

---

I can get mine to engrave at up to 800 mm/sec - but there tends to be some random spots scattered about. 500 mm/sec is as fast as I will use for a large item, but generally 300 or less for "regular" sized things - though I do not do a lot of engraving.

Rise time can be compensated by pre-ionising the tubes, RECI and EFR tubes and psu's have a setting to do exactly this

I know this is a little old, but I was looking for a review and stumbled across this...

One area that is often overlooked is the power supplied to the stepper/servo motors.

The 5080's referenced before can run at 50V, but most Chinese lasers are running 24V systems, thus losing tons of (NZ vernacular here) potential grunt.

Adding stepper coolers and increasing the voltage on the stock steppers makes a remarkable difference to CNC performance.

Depending on the controller in your machine, you could either stick another 24V PSU in series with the existing one (if the controller will take 50V) or just power the motive side of the stepper driver with 48V, leaving the controller running at 24V.

For example, my Smoothieboard runs at up to 36V, but it is happiest at 24V, so it was easier to just take the signals from the board and supply the drivers with 48V.

Which brings us to another point - the controller is very influential in how the XY motion control performs.

For me, my steps to improved motion have always been, after making sure the motion paths are as clean and smooth as they can be: 1) controller, 2) voltage/cooling 3) driver

The controllers supplied with most Chinese CNC gear have very rudimentary motion-control code in them. jerk values like slamming a door, no pre-planning (so each new motion comes as a surprise!), often stop motion entirely while waiting for an event (such as changing state of the laser) and have no tuning available due to not knowing what they have coded in their firmware nor how to access that firmware and adjust settings.

Typically vendors set upper limits to things like travel so while you might tell your encoder (gcode or whatever motion control creator) that you want to move at 5000 mm/min, the upper limit inside the controller is set to 3000 mm/min and it will accept your command, while going no higher than pre-set limits. These limits are compiled-in, so no configuration file change will alter them.

So, the best thing you can do to your cheap (or not-so-cheap) Chinese laser with regards to performance is to put a smart 32-bit controller on it and play with the settings. This could be as simple as a $35 32-bit Arduino with CNC shield or a $250 Cohesion3D.

This is 'trivial' from a wiring/installation perspective, but will have significant impact on your toolchain. If you are a Corel user, you won't be able to just print to your laser any more. You will have to use something like LaserWeb after creating a file compatible (IIRC LaserWeb doesn't talk CRW).

As you may have guessed, I'm not your average hand plane and forstner-bit woodworker. My background is electronics and embedded systems and I came to CNC and lasers because of an inability to measure the same twice in a row or to cut a straight line :)

I blame the left-handedness and a lack of discipline as a child... mostly.

Mike this is an older Thread but not a lot has changed. The DC glass tube can not fire fast enough for good engraving quality above a certain speed, either with steppers or servos or a hybrid. The higher priced machines with RF tubes do in fact use servos, most do anyway.

Quote:

---

Originally Posted by Mike Thornbury

The controllers supplied with most Chinese CNC gear have very rudimentary motion-control code in them. jerk values like slamming a door, no pre-planning (so each new motion comes as a surprise!), often stop motion entirely while waiting for an event (such as changing state of the laser) and have no tuning available due to not knowing what they have coded in their firmware nor how to access that firmware and adjust settings.

---

I am curious as to how you determined this. There are several brands of controllers and many versions of each brand. What you describe does not match my experience.

The Ruida controller, which is a very common one on Chinese lasers, offers a great deal of customization.

While I agree that stepper performance could be increased with a higher voltage, that performance is wasted on a glass tubed laser. With the standard electronics I can already move the laser head faster than the tube will fire.

I just remembered a good example of how the Chinese controllers handle path planning.

I was cutting dozens of spur gears out of plywood, each about 7" in diameter with about 120 teeth. Each tooth was made of dozens of tiny line segments.

There were some minor inaccuracies in the cut gears, when stacked they had to be rotated just right to line up. A friend who worked at a very fancy shop offered the use of their laser after hours. It was a late model Universal, 120W I think.

The difference was that my Chinese machine would cut the entire gear at the set speed, something like 12 to 14mm/sec. The many tiny line segments were no obstacle to it. The Universal would bog down when cutting the teeth, excruciatingly slow - like 2-4mm/sec. My cheap Chinese laser was outperforming the expensive mainstream machine. We spent quite a bit of time trying to get it to cut those teeth at a reasonable speed, we ended up just lowering the power instead. The minor inaccuracies were present as well.

Probably, there was a solution to making the Universal behave appropriately, but we couldn't find it at the time.