PDA

View Full Version : Rebuilding terribly engineered laser cutter



threedee
20-06-2019, 09:51 PM
I need help.

I am no engineer, tinkerer with penchant for cnc/electronics/rc more like, but this thing that i bought threw even me.
I am at a point where i'm willing to just angle grinder the interior out and put in a proper XY
in place of dogs breakfast equivalent i have here.
I can deal with electronics and control, but i've never done linear motion (just fixed/diagnosed some).

What i have here is a big'ish laser cutter (80W, 600x900mm bed) which is driven by 2 steppers, one for gantry (Y) and one for head (X). Both NEMA17 sized.
Head X motion is fairly ok, runs on block bearing, tensioner/idler fork needs replaced though.

The Y is the dogs breakfast i mentioned. Dual-shaft motor runs shafts to sides of ONE round linear bearing, 12mm round chrome rail and one of whatever the hell that is - just a plastic wheel running on flat side of angled aluminium extrusion. Both are pulled by timing belts, by non-coplanar pulleys, because dual-shaft motor is mounted off centre and coupled with spiral cut flexible aluminium couplings. Motor cant be moved or adjusted because some bright spark decided to weld the bracket into case, by also not leaving space for work table subframe to fully rise.

The machine will do engraving just fine, at speeds of up to 400mm/s in X axis. But as soon as Y motion is engaged in any way(besides moving a scan line down) everything goes to **** - terrible resonance, repeatably bad cuts, i suspect due to windup because of couplings, etc. It cannot cut/draw a straight circle at 25mm/s. Waving cuts all over the place at certain travel angles irrespective of speed (10-200mm/s). Trying to draw anything over 250mm/s stalls the motor, probably due to windup induced racking. Which excludes any small detail work where lots of stops and direction changes are present.

Machine is absolutely useless outside of scan engraving. Angle grinder came to mind.

What i need to figure out is how to lay out and plan for parts, LEARNING about the parts needed, how to square things with no reference points, etc, etc.

I have tons of photos of extraordinary, fairly expensive cheapness to show off if anyone interested.

What i want to do is fit the motion with one bigger motor (nema23?) for Y and plain old NEMA17 for X, remove all the Y motion nonsense and fit something like a block track/bearings on both sides, driven by single shaft, that is driven by belted transmission (nema23 to Y drive shaft) behind it.

I'd appreciate any pointers on where to start.

routercnc
22-06-2019, 03:37 PM
I need help.

I am no engineer, tinkerer with penchant for cnc/electronics/rc more like, but this thing that i bought threw even me.
I am at a point where i'm willing to just angle grinder the interior out and put in a proper XY
in place of dogs breakfast equivalent i have here.
I can deal with electronics and control, but i've never done linear motion (just fixed/diagnosed some).

What i have here is a big'ish laser cutter (80W, 600x900mm bed) which is driven by 2 steppers, one for gantry (Y) and one for head (X). Both NEMA17 sized.
Head X motion is fairly ok, runs on block bearing, tensioner/idler fork needs replaced though.

The Y is the dogs breakfast i mentioned. Dual-shaft motor runs shafts to sides of ONE round linear bearing, 12mm round chrome rail and one of whatever the hell that is - just a plastic wheel running on flat side of angled aluminium extrusion. Both are pulled by timing belts, by non-coplanar pulleys, because dual-shaft motor is mounted off centre and coupled with spiral cut flexible aluminium couplings. Motor cant be moved or adjusted because some bright spark decided to weld the bracket into case, by also not leaving space for work table subframe to fully rise.

The machine will do engraving just fine, at speeds of up to 400mm/s in X axis. But as soon as Y motion is engaged in any way(besides moving a scan line down) everything goes to **** - terrible resonance, repeatably bad cuts, i suspect due to windup because of couplings, etc. It cannot cut/draw a straight circle at 25mm/s. Waving cuts all over the place at certain travel angles irrespective of speed (10-200mm/s). Trying to draw anything over 250mm/s stalls the motor, probably due to windup induced racking. Which excludes any small detail work where lots of stops and direction changes are present.

Machine is absolutely useless outside of scan engraving. Angle grinder came to mind.

What i need to figure out is how to lay out and plan for parts, LEARNING about the parts needed, how to square things with no reference points, etc, etc.

I have tons of photos of extraordinary, fairly expensive cheapness to show off if anyone interested.

What i want to do is fit the motion with one bigger motor (nema23?) for Y and plain old NEMA17 for X, remove all the Y motion nonsense and fit something like a block track/bearings on both sides, driven by single shaft, that is driven by belted transmission (nema23 to Y drive shaft) behind it.

I'd appreciate any pointers on where to start.

routercnc
22-06-2019, 03:39 PM
I don’t have one so can’t help but maybe posting some internal pictures will get you some ideas

Boyan Silyavski
23-06-2019, 08:58 AM
I don’t have one so can’t help but maybe posting some internal pictures will get you some ideas

Yes, it will be great to see the marvels of the Chinese engineering.. I am waiting for it.



I doubt it will be too difficult to upgrade it. Open its insides and find the drives for the motor, find the power supply for the drives. Hopefully separate drives , not integrated in some stupid board..

I doubt nema 17 could achieve proper laser speeds, only if its servo.


You can use square chinese rails 12 or 15 size for the system, not all of them are bad, i can point you to good ones for cheap, which i have tested, not once. Steel and so.

threedee
23-06-2019, 09:58 AM
Electronics are not the problem - Controller+front display panel (RDC6442S), 2x stepper drivers (0.09-1.13A M415DRV), 24V psu, laser tube HV psu (MYJG100W), 2x nema17 sized motors (cant identify by marks). Motors/drivers can be replaced if needed. For now it works.

The motion is another matter. If anyone is familliar with K40 laser construction, mine is that scaled way up past reasonable limits. K40 is 3020 bed, mine is 9060, runs on same XY essentially. Which is BS.

Problem is there is so little space and no continuous plane to put down reinforced rail on right side.
There are bed lift bearings located in the floor of 70mm wide space could have been used for rail. Same 70mm space houses drive belt, drag chain and said lift upper bearing blocks.

Any way to put down reindorced rails are to raise them off the deck somehow.

Robin Hewitt
23-06-2019, 02:44 PM
I think you should stick with the small steppers, you don't need power you need speed and at 200 full steps/rev those little motors are quite zippy.

Boyan Silyavski
23-06-2019, 02:50 PM
Obviously this gantry is racking.

There is enough space to fit 20x20 aluminum profile and on top of it 12 or 15 size bearings. or if not, , size 20 rail in the air with no support and 1 bearing block. It will be way better than it is

threedee
23-06-2019, 02:54 PM
I'm sure they are. My problem with them is i cannot be sure of their specs. Model/serial numbers do not exist on "Smooth" manufacturer website. Nor anywhere on t'intertubes for that matter, aside from what I'VE posted...

Drivers came preset to 0.5A which suggests weak motors. Looked about the net for replacements, seeing nema17 sized motors up to 2A, 1A being more or less norm. So why 0.5A on mine ?

Right side is completely unsupported. The only things holding in "in position" are one plastic wheel pulled by gravity and drive belt. I can lift the gantry off the "rail" by 5mm. Right side motion is effectively restrined only on the left rail. So yeah, racking. But not to the point of stalling. I'm getting oscillations in straight and curved cuts. Resonance wave. Changes angle from speed to speed, but always present.

Boyan Silyavski
23-06-2019, 03:48 PM
Maybe armed with a plastic glue gun and pieces of plastic if needed, you can fix all on place and there will be no real need to do anything more? I have done that to flimsy 3d friends printer and the "upgraded version" was like 10 times better than the original

magicniner
24-06-2019, 09:08 AM
Right side is completely unsupported. The only things holding in "in position" are one plastic wheel pulled by gravity and drive belt. I can lift the gantry off the "rail" by 5mm.

The engineering issues there are the poor rail and bearings and the belts not being synchronised properly, the plastic wheel is a good engineering solution if the rest were adequate.
I think you should get someone to look at this for you as you are considering changes which, at best, may not help whilst not addressing the primary issues.

threedee
01-08-2019, 04:47 PM
This is what i'm battling:

https://photos.app.goo.gl/32mNoD2Wcsz9pnzEA

Note letter W in my text speed tests. Even straight angled lines getting the wobble.
Circles (or spirals in my tests) exhibit same vibration patterns radiating out from the centre of ALL cut tests.
The accelleration has been knocked down in half from factory defaults, lowering them any lower makes machine sluggish and controlling tube power becomes an issue (sharp corner overburn, slow decelleration/accelleration induced).

What has been done to day:
1. Replaced super-thin and warped table raising brackets to more substantial and 2 point mounted ones.
Initially, the table was held up by a single m4 screw in a hole drilled for m5 and covered with wider washer to cover size difference. Original brackets got bent on tightening washer nuts. Table subframe had m5 holes drilled from underneath, for TWO mounting points, but only ONE access hole on the other size. As i said, only mounted by one undersized screw.

2. Replaced original 1.5mm aluminium head mount plate to 2mm steel, laser cut by my friend at his workplace. they cut steel for screeners/heavy machinery, plasmas and lasers. Laser in my case as i had small mounting holes and some that needed tapping. Plasma cant do fine detail, hence laser. Had to pull the part, caliper it around and draw out the part for cutting (used to work with autocad a lot, so no biggie). Helped, but only marginally. Wobliness didnt go away.

3. Tested for Y rail vibration transmission. Moved the head to work position in the middle of the table, loaded a 100x100 test pattern (spiral), lightly clamped Y rail to the frame - most of wobliness went away.
Didnt want to wreck existing rail by clamping/bending it, so didnt do any more tests after that. Its pretty damn convinced me to replace whole Y "assembly"...

4. Bought SBR20 rail and double cartridge bearing (95mm long). Cutting the rail was a bit of a challenge, as NONE of local engineering companies had abrasibe cutting machines. All of them mostly work with mild steel, so bandsaws or reciprocating saws only. The hardened rail would have wrecked anything with teeth... I cut off a piece for gantry height fitting by an angle grinder. Disk wouldnt reach all the way through 20mm, had left about 3mm, wrecked 3 bimetal blades in my hand hacksaw to finish the job... Yeesh...

5. Designed and laser cut mounting plates for connecting gantry to bearing. Asked to cut in all available thicknesses from 8mm to 2mm as exact measuring is tricky, all of the machine case is welded, caliper wouldnt fit. So i decided to do a "sandwitch approach" - pick most suitable thickness of 2 plates and weld them before fitting (when it fits correctly.). The idea was to put one plate on the bearing and one plate on gantry, without disturbing existing hole patterns on gantry, which required me to split mounting plate in two due to existing gantry holes. Shim it as necessary. All of this is being done to prevent any disturbance of mirror 1 and the tube placement. If i move one of those - lots of pain of readjusting the whole beam path, squareness issues.

6. Removed all MXL belts and pulleys and replaced them with GT2 10mm belts/pulleys. I highly suspect Y pulleys fitted are GT2, which is a mismatch, as MXL is 2.032 and GT2 is 2mm pitch. Original belts were also too short, at slackest (tensioner all the way inward) they ring. New belts will be about 10mm longer for this reason.

7. Replaced MXL belt with GT2 10mm on X motion (head on the gantry). Had to replace motor as it had pulley pressfit onto it and too short a shaft to fit GT2 pulley with set screw on. Also, while doing tests i have noticed that X motor was getting too hot in XY operations, but, interestingly, NOT in engraving where its moving the most. For some reason, doing vector engraving, while both motors run equally as much, only X was getting overheating. Y, while carrying the whole gantry/head/Xmotor assembly would barely rise in temperature at all... go figure.

8. As someone suggested i swapped over both stepper drivers just to see if overheating issue would move to Y axis - it didnt. Both drivers are functioning. However i still cannot identify them. They are marked as M415DRV, but there is absolutely nothing about them on the t'intertubes... Bought DM556 drivers, for now they sit on shelf until i finish Y rail and gantry assembly.
Cannot do same swapping test on motors, as Y motor is dual shaft and both ends needed to drive Y, while X is single shaft.

9. By chance came across and bought some Baldor integrated motors, 3 nema17 and one nema23 sized. Not encoded ones, but with drivers built into back housing. Will be using nema23 one to drive Y later as i'm planning to chuck that welded in nema17 bracket out. Need to figure out the 2:1 recduction for Y single shaft. Have to design it first, but thats for later.

10. I even suspected my room floor having something to do with all this vibration in cuts nonsense, soft floor (insulation etc). So i moved the whole machine into bedroom where i have tiled section sitting on concrete slab - no change, same vibration patterns in cuts, so no external vibration transferrence...

At this point i'm focused on replacing original 12mm hanging rail with supported 20mm rail. After that, when i verify problem gone/not gone, replace extraction (one more thing that may have induced vibration as its built into the case, that thought came after i started Y rebuilding), insulate high voltage wiring with silicone tubing, replace both motors with known specs, replace both drivers or use Baldor integrated ones i got by chance. Design/manufacture switch/indicator console for water/air/power meter/lighting/accessories. Clean up wiring, as at the moment power wiring is sitting in the same trunk as signal/control wiring (separated for now, to rule out EM interference, was not the case)...

Oof, long post. Are you seeing photos i shared ? There are much more, or, if you want to see something specific in more detail i can snap some more...

#EDIT1# 11. Replaced rubbish water and air pump outlets that would not hold plugs and spark. Put on proper UK double switched socket outlets with proper grounding etc...

Doddy
01-08-2019, 07:04 PM
Completely random thought, feel free to disregard...

Is it worth placing a red laser in front of the tube and verifying the complete optical path (bit of smoke, video camera) to make sure the axis drives aren't causing vibration in the various path mirrors etc. What I'm thinking is position Y at min/max extremes and rapid transits across X, observing any oscillation in the optical path. Might be worth if only to disregard this effect.

threedee
01-08-2019, 07:12 PM
X is behaving on its own. I did multiple different scan engravings with no adverse effects. No vibration until Y axis engages. If it is only used to move a scanline it doesnt show at all. But try to cut vectors - bang, vibration all over the travel path...

Also, i have checked mirrors for wobbling/rattle - everything is tight and screwed down. All the way along the beam path, from lens to tube itself...

Boyan Silyavski
02-08-2019, 11:35 AM
What is powering all this setup? It must be PSu problem. Buy a proper industrial PSU, even from China, say Meanwell at least. A no name PSU=Crap 100% of the time

ericks
02-08-2019, 12:13 PM
This is what i'm battling:

https://photos.app.goo.gl/32mNoD2Wcsz9pnzEA

Note letter W in my text speed tests. Even straight angled lines getting the wobble.
Circles (or spirals in my tests) exhibit same vibration patterns radiating out from the centre of ALL cut tests.
The accelleration has been knocked down in half from factory defaults, lowering them any lower makes machine sluggish and controlling tube power becomes an issue (sharp corner overburn, slow decelleration/accelleration induced).

What has been done to day:
1. Replaced super-thin and warped table raising brackets to more substantial and 2 point mounted ones.
Initially, the table was held up by a single m4 screw in a hole drilled for m5 and covered with wider washer to cover size difference. Original brackets got bent on tightening washer nuts. Table subframe had m5 holes drilled from underneath, for TWO mounting points, but only ONE access hole on the other size. As i said, only mounted by one undersized screw.

2. Replaced original 1.5mm aluminium head mount plate to 2mm steel, laser cut by my friend at his workplace. they cut steel for screeners/heavy machinery, plasmas and lasers. Laser in my case as i had small mounting holes and some that needed tapping. Plasma cant do fine detail, hence laser. Had to pull the part, caliper it around and draw out the part for cutting (used to work with autocad a lot, so no biggie). Helped, but only marginally. Wobliness didnt go away.

3. Tested for Y rail vibration transmission. Moved the head to work position in the middle of the table, loaded a 100x100 test pattern (spiral), lightly clamped Y rail to the frame - most of wobliness went away.
Didnt want to wreck existing rail by clamping/bending it, so didnt do any more tests after that. Its pretty damn convinced me to replace whole Y "assembly"...

4. Bought SBR20 rail and double cartridge bearing (95mm long). Cutting the rail was a bit of a challenge, as NONE of local engineering companies had abrasibe cutting machines. All of them mostly work with mild steel, so bandsaws or reciprocating saws only. The hardened rail would have wrecked anything with teeth... I cut off a piece for gantry height fitting by an angle grinder. Disk wouldnt reach all the way through 20mm, had left about 3mm, wrecked 3 bimetal blades in my hand hacksaw to finish the job... Yeesh...

5. Designed and laser cut mounting plates for connecting gantry to bearing. Asked to cut in all available thicknesses from 8mm to 2mm as exact measuring is tricky, all of the machine case is welded, caliper wouldnt fit. So i decided to do a "sandwitch approach" - pick most suitable thickness of 2 plates and weld them before fitting (when it fits correctly.). The idea was to put one plate on the bearing and one plate on gantry, without disturbing existing hole patterns on gantry, which required me to split mounting plate in two due to existing gantry holes. Shim it as necessary. All of this is being done to prevent any disturbance of mirror 1 and the tube placement. If i move one of those - lots of pain of readjusting the whole beam path, squareness issues.

6. Removed all MXL belts and pulleys and replaced them with GT2 10mm belts/pulleys. I highly suspect Y pulleys fitted are GT2, which is a mismatch, as MXL is 2.032 and GT2 is 2mm pitch. Original belts were also too short, at slackest (tensioner all the way inward) they ring. New belts will be about 10mm longer for this reason.

7. Replaced MXL belt with GT2 10mm on X motion (head on the gantry). Had to replace motor as it had pulley pressfit onto it and too short a shaft to fit GT2 pulley with set screw on. Also, while doing tests i have noticed that X motor was getting too hot in XY operations, but, interestingly, NOT in engraving where its moving the most. For some reason, doing vector engraving, while both motors run equally as much, only X was getting overheating. Y, while carrying the whole gantry/head/Xmotor assembly would barely rise in temperature at all... go figure.

8. As someone suggested i swapped over both stepper drivers just to see if overheating issue would move to Y axis - it didnt. Both drivers are functioning. However i still cannot identify them. They are marked as M415DRV, but there is absolutely nothing about them on the t'intertubes... Bought DM556 drivers, for now they sit on shelf until i finish Y rail and gantry assembly.
Cannot do same swapping test on motors, as Y motor is dual shaft and both ends needed to drive Y, while X is single shaft.

9. By chance came across and bought some Baldor integrated motors, 3 nema17 and one nema23 sized. Not encoded ones, but with drivers built into back housing. Will be using nema23 one to drive Y later as i'm planning to chuck that welded in nema17 bracket out. Need to figure out the 2:1 recduction for Y single shaft. Have to design it first, but thats for later.

10. I even suspected my room floor having something to do with all this vibration in cuts nonsense, soft floor (insulation etc). So i moved the whole machine into bedroom where i have tiled section sitting on concrete slab - no change, same vibration patterns in cuts, so no external vibration transferrence...

At this point i'm focused on replacing original 12mm hanging rail with supported 20mm rail. After that, when i verify problem gone/not gone, replace extraction (one more thing that may have induced vibration as its built into the case, that thought came after i started Y rebuilding), insulate high voltage wiring with silicone tubing, replace both motors with known specs, replace both drivers or use Baldor integrated ones i got by chance. Design/manufacture switch/indicator console for water/air/power meter/lighting/accessories. Clean up wiring, as at the moment power wiring is sitting in the same trunk as signal/control wiring (separated for now, to rule out EM interference, was not the case)...

Oof, long post. Are you seeing photos i shared ? There are much more, or, if you want to see something specific in more detail i can snap some more...

#EDIT1# 11. Replaced rubbish water and air pump outlets that would not hold plugs and spark. Put on proper UK double switched socket outlets with proper grounding etc...

They have good ides but their attention to detail is poor. I had to rebuild my 6040 cnc router to make it work properly. And when i upgraded i just built my own machine:)

threedee
02-08-2019, 12:57 PM
What is powering all this setup? It must be PSu problem. Buy a proper industrial PSU, even from China, say Meanwell at least. A no name PSU=Crap 100% of the time

You mean low voltage power supply ? I need to check the name. Checked voltage on it, its holding at 24V while operating.
As for high voltage its Zye MYJG-100W. That has absolutely nothing to do with motion at all.


They have good ides but their attention to detail is poor. I had to rebuild my 6040 cnc router to make it work properly. And when i upgraded i just built my own machine:)

That is exactly what i thought and will do myself. As for having good ideas - hell no, in this case its an astonishing display of cost related corner cutting and nothing else...

ericks
02-08-2019, 01:27 PM
I have also been looking at lasers.....basically also decided to build my own. We have a small trotec laser here at work, they paid $35000 aud for it. Let me tell you there is nothing in it to justify the price

threedee
02-08-2019, 01:35 PM
35k Aud for a small laser, yikes.

But, yeah, there is absolutely nothing in there that justifies the price. Not even tech support.
What is laser cutter anyway... 1 tube, 2 power supplies, 1 controller, 2 motors, 2 motor drivers, some switches and wiring in a box. With what i know now i wouldnt have even entertained the idea of importing manufactured laser. Even if i didnt pay for it 35k, i would have built it for one third or half of money i paid... Bloody annoying... Got a nice case though :D

Doddy
02-08-2019, 01:40 PM
This image says a lot...

26141

It's unclear from some of the other pictures, but this one is particularly interesting as it would appear to show the distortion on spokes projecting radially from the spiral centre, and at a position where the Y axis velocity is tending to zero or accelerating from zero. Is this true of other images? (they are less clear).

Could this be related to torsional elasticity of the couplings/Ali extension bars on the Y-axis?, is it worth trying to drive the two belts separately with two steppers, and removing the length of Ali bar?

Just a random thought

ericks
02-08-2019, 01:40 PM
Yes ours at work also has a very nice rounded case and a small stand:)

Obviously they would come out for repairs/maintenance but it's going to cost heaps!!

threedee
02-08-2019, 01:51 PM
This image says a lot...

26141

It's unclear from some of the other pictures, but this one is particularly interesting as it would appear to show the distortion on spokes projecting radially from the spiral centre, and at a position where the Y axis velocity is tending to zero or accelerating from zero. Is this true of other images? (they are less clear).

Could this be related to torsional elasticity of the couplings/Ali extension bars on the Y-axis?, is it worth trying to drive the two belts separately with two steppers, and removing the length of Ali bar?

Just a random thought

This particular picture needs rotated 90 degrees. The "spokes" radiate in X direction (horizontal).
ALL spiral cuts at ANY speed exhibit this same pattern. Lower the speed, closer the oscillations, making even through cuts difficult in those places due to increase in speed due to vibration. When on cutting speeds (10-20mmps) either kerf gets too wide (by increasing power beyond required for cutting) or not cutting through (power is set exactly right for cut through).

I have replaced aluminium spiral couplers to spider couplers to no effect. The shaft that drives Y is 8mm bars, cant be that much windup in them.

Also, the "spoke" pattern changes angle on changing speeds, but is always there and always radiate in spirals/circles. It makes a strong case for resonance vibration somewhere, i'm convinced that its the Y rail bolted into case. Being only 12mm and 750mm long it can act as an amplifier for vibration from motors... Hence me replacing Y.


Yes ours at work also has a very nice rounded case and a small stand:)
Obviously they would come out for repairs/maintenance but it's going to cost heaps!!

Cost whom ? You or them ? If its you - whats the point for that 35k price :D

ericks
02-08-2019, 01:55 PM
Yes our company will have to pay....
Those tubes don't last forever

threedee
02-08-2019, 03:00 PM
Yes our company will have to pay....
Those tubes don't last forever

Replacing the tube is pretty straightforward, they are not that expensive in grand scheme of things. None of this warrants 35k in my opinion.