PDA

View Full Version : Engraving issues



D-man
27-02-2016, 09:11 AM
Hey guys, I thought the machine was pretty good, setup wise, until I tried to do some small engraving (100x50mm) but turns out I have some issues!

Can anyone point me in the right direction with this as to what the issue could be?
All nuts and bolts are tight, no slop in any axis and slaved motors are spot on. I'm actually stumped

http://uploads.tapatalk-cdn.com/20160227/837a7f501dc66a9e159f528cc9ca5db6.jpg


Sent from my iPhone using Tapatalk

Davek0974
27-02-2016, 09:23 AM
Cutter moved up in the collet???

D-man
27-02-2016, 09:25 AM
Thanks Davek0975

Apologies I should of made more clear what my issue was. As you can see the curves in the engraving are horrendous! Straight lines are amazing.

The different thicknesses was me trying different depths


Sent from my iPhone using Tapatalk

Davek0974
27-02-2016, 09:31 AM
Ok, flat portions on curves are generally caused by slop in one or more axes, maybe a loose pulley or worn belt (if any)

D-man
27-02-2016, 09:42 AM
An I've never checked the pulleys on the shaft! Will do that now


Sent from my iPhone using Tapatalk

komatias
27-02-2016, 09:44 AM
It could be the gcode too. Which programming you using?

D-man
27-02-2016, 09:50 AM
I'm using Vcarve pro pal


Sent from my iPhone using Tapatalk

D-man
27-02-2016, 09:50 AM
Created DXF in solidworks THEN MOVED OVER TO V CARVE


Sent from my iPhone using Tapatalk

komatias
27-02-2016, 11:27 AM
Created DXF in solidworks THEN MOVED OVER TO V CARVE


Sent from my iPhone using Tapatalk



Not used V-Carve so not sure of the options. Check to see if the arc are made up of straight segments or if they are real curves. It they are segments, then you there may be an option to smooth the curve. This replaces the straight lines with curves that make the program run better.

Failing that it may be mechanical.

Best of luck

D-man
27-02-2016, 11:35 AM
Cheers I will take a look. Mechanical wise everything's looking solid. Could it be a backlash problem? I had measured it but it was fine


Sent from my iPhone using Tapatalk

Neale
27-02-2016, 11:55 AM
Not sure about SolidWorks but certainly exporting DXF files from something like Adobe Illustrator gives a few different DXF file options. I tried using an "early" DXF format and ended up with curves with so many straight line segments that it was almost impossible to import into Fusion 360. Changed to exporting in the "latest" DXF format and it used "proper" arcs and curves - enormous difference.

JAZZCNC
27-02-2016, 12:11 PM
Derek take another pic closer up to curve please.

This looks like end float to me because the top of the E as kink in it just where direction change would occur.? Are the screws moving in the end bearings.? Give it firm shake.

Edit: Could also be how you have SW setup because if the display settings are low it will break arcs into lines for resolution purposes. Not sure if when exporting to DXF it changes anything or just puts out whats on screen. How does it look on SW screen.?

D-man
27-02-2016, 12:34 PM
http://uploads.tapatalk-cdn.com/20160227/5792727da6871ce834ba009f04088b9a.jpghttp://uploads.tapatalk-cdn.com/20160227/f28ab2c6fd09bbadc8d992700079ef82.jpg

Cheers Dean, screws are solid mate

Attached is my setup

http://uploads.tapatalk-cdn.com/20160227/42de43e020708f863664f44406f0e8e0.jpg


Sent from my iPhone using Tapatalk

JAZZCNC
27-02-2016, 12:41 PM
Defo Backlash or Slop.!! If you look every curve as got kink at 90deg segment were direction change occurs.

Edit: Actually looking closer it's just before the 90deg segmant which is strange but still I think it's mechanical rather than G-code. Can you post the G-code file.

Don't know how you had them on bed but you will be able to work out which axis is the problem.

D-man
27-02-2016, 12:48 PM
I've just ran the machine again with a hand on each axis! Now, on the right side of the machine it seems to jump forward ever so slightly I'm guessing this is where the issue lays!!

Backlash - maybe sorting the backlash for a legend like yourself it's easy but this will be the first try for me haha can you give the best way to do this?


Sent from my iPhone using Tapatalk

Davek0974
27-02-2016, 01:12 PM
It all depends where the backlash is coming from, could be loose shaft-end block allowing the shaft to move end-end, slack nut and some other possibilities.

JAZZCNC
27-02-2016, 01:12 PM
Backlash - maybe sorting the backlash for a legend like yourself it's easy but this will be the first try for me haha can you give the best way to do this?

Leg End more likely but what I'm certainly not is Mystic Meg. Can't tell you how to fix when don't know whats wrong.??

I'll make a small wager that jumping is the Rotating nut needs adjusting.!

D-man
27-02-2016, 01:14 PM
I meant sort the backlash haha or best way to measure for backlash.

Cheers Dave, I'm leaning towards stripping everything and giving it a complete going over


Sent from my iPhone using Tapatalk

JAZZCNC
27-02-2016, 01:19 PM
Cheers Dave, I'm leaning towards stripping everything and giving it a complete going over

Don't do that. Find the problem then do that. If not you'll always be guessing what the problem was.

Measuring back lash is easy just need a dial gauge and have the machine move away then return and should come back to same place every time. Search youtube you'll find loads of examples.

D-man
27-02-2016, 01:24 PM
Cheers is it normal for the ball nuts to be magnetised!!?? I'm think it is due to them running up and down the screws all day. I'm sure I've just had a small jolt while checking ball nut screws! Trouble is I'm scared to check again maybe it was the dodgy tenant in my hand...

It's like plucking up the courage to stick your tongue on a 9v battery all over again hahaha


Sent from my iPhone using Tapatalk

JAZZCNC
27-02-2016, 01:28 PM
PUSSY.!!!

Do you have an Earth lead on the machine.? If not then put one on and take back to Star Earth point. It was possibly static.

D-man
27-02-2016, 01:31 PM
Cheers, nothing like being blunt PMSL

Nothing on the machine itself no. Star earth point?


Sent from my iPhone using Tapatalk

JAZZCNC
27-02-2016, 02:07 PM
Nothing on the machine itself no. Star earth point?

You should know this Dman you've been around long enough.!! . . . .Can't just throw this shit together and expect it to work perfectly.

Star point is the Single Earth point in the Control box that you should have for all Gnd's and Shields to return to. If not then your open to Earth loop problems.

phill05
27-02-2016, 02:20 PM
Before you go ripping into it have you checked ALL vectors are closed in V Carve before you toolpath. (closed not grouped) DXF are a swine to mess you about when imported into V carve.


Phill

Davek0974
27-02-2016, 03:24 PM
Whatever you do, get a continuous earth from the body of the spindle (If metal) or the mount back to the star point as mentioned, without this, any stray voltage is going to earth via you ballscrews, nuts and bearings - this is seriously bad news and can ruin balls in short time, use multi-strand wire similar to heavy speaker cable etc - flexible but lots of strands.

D-man
27-02-2016, 03:58 PM
Ok getting to the backlash, is this normal? Seems to bounce back then slowly zero, am I doing something wrong?

This is on the left side of the gantry, bot on the right side it's a different story.

http://youtu.be/GcV8a14q7xs


Sent from my iPhone using Tapatalk

D-man
27-02-2016, 04:01 PM
Scrap that right side is now the same!


Sent from my iPhone using Tapatalk

D-man
27-02-2016, 04:06 PM
http://uploads.tapatalk-cdn.com/20160227/66b25417ad08f0b65a7bc4bb33b3c4ca.jpg

Not sure it's made a difference. That "e" is wrote in VCarve. Could it be the processor settings?


Sent from my iPhone using Tapatalk

D-man
27-02-2016, 04:16 PM
Changing the processor didn't help, well maybe a tiny bit, something is still not right and it's at the same pints all the time (at the black dots)

Left is mach3 processor with ARCS
right is mach3 processor with (MM)

http://uploads.tapatalk-cdn.com/20160227/10db7659aff27033bc2001ca2417a878.jpg


Sent from my iPhone using Tapatalk

komatias
27-02-2016, 04:27 PM
That is lost motion somewhere, either backlash or the screws/bearings are not tight. The blips happen when start accelerating when you change direction as you can see. Also have you changed any setting for backlash compensation in the software?

D-man
27-02-2016, 04:28 PM
Yeah I did have some

http://uploads.tapatalk-cdn.com/20160227/0f95e7b3163a7812b317a43a4b19f4cb.jpg



Sent from my iPhone using Tapatalk

D-man
27-02-2016, 06:47 PM
Well that's me stumped! Been through everything and nothing is loose. I'm starting to wonder if it's in the setting in the servo drives


Sent from my iPhone using Tapatalk

komatias
27-02-2016, 07:14 PM
I cannot comment on the electornics and motor settings.
The video you posted shows repeatability which is a start. If the lost motion is repeatable it will....repeat. It does not show lost motion. Backlash in the screws or bearings tends to be repeatable and that is why in old milling machines and lathes we only cut in one direction.

You may have already done this but for the benefit of other, what I would do to ensure the screws are not moving is:

1)put the dial indicator on the captive ends of your ballscrews
2)run the gantry in one direction only for 20mm. This will take up the slack in the opposite direction of travel.
3)zero the dial indicator
4) run the gantry in the other direction and see if the dial indicator has jumped. If it does, your screws are not rigidly held in place.

D-man
27-02-2016, 07:16 PM
Ah that's a point I will try that tonorrow


Sent from my iPhone using Tapatalk

Neale
27-02-2016, 07:39 PM
If your dial gauge has enough travel, you can also come back to the gauge position and overshoot a few millimetres, then return to zero from the opposite direction, so to speak. This will show up backlash in the whole system, rotating nuts included.

Boyan Silyavski
28-02-2016, 09:37 AM
Hi Derek,

Dont rush to disassemble everything.

That what you see on the dial indicator is the servo correcting after movement and fixing the gantry in position i believe.

If needed mark with lines and marker and check in order both sides :
-the 6 screws that tighten the ball nut to the rotating part
-the big bearings on the rotating nut if they have not slipped on the housing
-if 2 nuts that hold rotating part to nut are not loose

Most possibly this is loose pulley, belt or end support problem. Check carefully all pulleys for slip, check if belts are well under tension/ same tension. Then check end supports if sth un screwed there. If not sure, disassemble motors and cement pulley again.

In fact you should have cemented every single screw on the machine, which i discovered was my big mistake on the first machine i made.

Ahh, and move gantry to one side Check with finger as you are pulling guitar strings if both screws sound with same pitch. If not, tighten all so that is the case. No need to check against tuner. But if you want to go so far, use Linostrobe on the phone. Other will not pick the sound.


PS. And remove that backslash compensation in Mach3. this will additionally mess things up. I have a small machine with backslash and from experience i know its better without it.

D-man
28-02-2016, 10:01 AM
I think I may have just found the problem.

Ball screw mound seem to move slightly

https://youtu.be/XFrAJuSvGoo


Sent from my iPhone using Tapatalk

magicniner
28-02-2016, 11:51 AM
From the brief view in the video it appears that the screw end bracket has no triangulation!

- Nick

JAZZCNC
28-02-2016, 01:22 PM
This is prime example when I say can't just throw these things together and expect to work perfectly.!!

No disrepect ment here Derek but that end fixing is pathetic and no way can you expect to hold fine tolerences done like that. The inertia of the gantry when cutting will rock those mounts like craddle.

The details are what make better machines and if poorly engineered or rushed into service missing these small details these are the results you can expect when get into finer work like engraving.!!! . . . . . The simple truth Derek is your machine isn't engineered good enough to do it.:whistle:

JAZZCNC
28-02-2016, 01:27 PM
PS. And remove that backslash compensation in Mach3. this will additionally mess things up. I have a small machine with backslash and from experience i know its better without it.

Agree 100% on BL-comp but look close Boyan and you'll notice it's not Mach3 Looks like Edding Cnc to me.? . . Still needs removing thou.

D-man
28-02-2016, 01:39 PM
Cheers lads! In my defence the end brackets were simply a temporary fix as I had run out of funds for more machining.

So do you think this is the problem? Yeah I'm using Eding CNC software


Sent from my iPhone using Tapatalk

D-man
28-02-2016, 01:45 PM
Ah now I remember playing with different mounting options!

Back to the trading board!

http://uploads.tapatalk-cdn.com/20160228/6c52baf2b7f569b3c1a0b80dcc7093f4.jpg


Sent from my iPhone using Tapatalk

D-man
28-02-2016, 04:25 PM
Ok so I found some scrap 25mm Ali and cut some end mounts out this shouldn't move!

Loving these single flute bits from toolstoday.com

http://uploads.tapatalk-cdn.com/20160228/d96a1c1baa45786895aac73d1a5d4e7e.jpg
http://uploads.tapatalk-cdn.com/20160228/766de1b30bd3cf78c42cd537ceeb583f.jpg


Sent from my iPhone using Tapatalk

JAZZCNC
28-02-2016, 04:36 PM
Hows that going to help.?

Davek0974
28-02-2016, 04:41 PM
Hopefully moving the screw along and mounting end bracket to it? But what about the other end??

D-man
28-02-2016, 04:43 PM
I thought thicker the better and mounted on the end of the extrusion with 4 M12 bolts would keep it sturdy? The screw will be offered up to it and bolted to the plate


Sent from my iPhone using Tapatalk

njhussey
28-02-2016, 06:34 PM
Just make a bracket from 2 pieces of your 25mm plate and fit where your existing bracket is?

Clive S
28-02-2016, 06:38 PM
Just make a bracket from 2 pieces of your 25mm plate and fit where your existing bracket is?You mean a spacer between the two fixed peices:sneakiness:

njhussey
28-02-2016, 07:09 PM
No, I meant a bracket ;)

Boyan Silyavski
29-02-2016, 09:57 AM
I dont know :hopelessness:...


Derek, you know that ball screw should be under maximum tension right???? Not just fixed there to stay on place. And that means screwed both sides with lock nuts. I thought that was clear to you when decided on the rotating ball nut to move gantry.

D-man
29-02-2016, 10:01 AM
Yes I understand that now. These are not your rotating parts they have been remade and designed


Sent from my iPhone using Tapatalk