View Full Version : Belt tension

Kev Williams
01-23-2014, 12:01 PM
What's the consensus on how tight the belts should be? My main focus of this question concerns the very long X-axis belt on my 1300x900 Triumph. FWIW, I've always kept the belts on my other 2 lasers quite tight. Not quite guitar-string tight mind you, but I CAN "strum" the X-axis belt on my LS900. I can tell when it needs tightening because some letters, near the point of deceleration on the sweep, start getting a bit thicker than others. My ULS-Optima has a smallish belt so it's not REAL tight but it's not loose. I've never messed with the Y-axis belts on either machine, but they're snug.

As for short belts, the Y-axis motor-to-shaft belt on my LS900 is extremely tight. In contrast, my IS7000 engraving machine uses a short belt from the stepper to the X-axis lead screw, and to me, that belt feels fairly loose. Yet the machine's 'slack' is virtually zero, even when machining aluminum.

Back to the Triumph... I'm thinking I need to cinch up the X-axis belt, as I'm having trouble with backlash alignment and producing lettering with smooth edges. Compared to my LS900 the belt isn't tight at all. I have played with the tension a few times. First time I snugged it up a bit, the belt (to me) still wasn't all that tight, but as the machine was running a 24" back and forth pass at 500mm/s (full speed for this machine), it seemed to produce a high frequency vibration harmonic that caused the machine's steel housing to 'hum'. Not real bad or loud, just noticeable. Could be I was just hearing something it always does, I'm not sure?

Anyway- If I keep the tension a little on the loose side, the long-side of the belt will wobble as it rasters. Pretty sure I don't want that because I've noticed very little movement of the belt causes the lens head to move quite noticeably. Also, I've created my own 'alignment test' which is simply a bunch of vertical lines .010 wide, about 1/8" apart that spans about 4" wide. I use a .3mm gap to run the test. Regardless of how close I can get the backlash to zero in the middle of the lines, the end lines will have around .010" or more 'zig-zag'. Tightening the belt helps reduce the end slack. Right now, I have the belt as tight as possible as the adjustment screws are bottomed out. Not a problem for me to move the belt in the adjuster a tooth so I can cinch it up tighter. As is, it still seems relatively loose to me by feel, and the belt still wobbles slightly during rastering. For what it's worth, the Y belts are MUCH tighter than the X belt..

Side question on the adjusters- there's 2 lock nuts on the adjuster's screws, which can be cinched up either against the lens head, or the belt clamp. I notice if they're cinched against the head, the belt clamp is free to "rock" a bit. If cinched against the belt clamp, the clamp won't rock. Which is correct?

Thanks in advance! :)

Joe Hillmann
01-23-2014, 12:13 PM
On Universal lasers they have a specific belt tension and to measure it you use a spring scale and a ruler. On mine on the x axis I think it is something like 50 grams of pressure in the middle of the belt should flex it 1/2 inch. Have you tried contacting triumph to see what they suggest on belt tension? I would assume when they assemble them in their factory they have some way to keep the tension somewhat consistent from machine to machine.

Gary Hair
01-23-2014, 12:49 PM
I can't tell you what yours should be but I can tell you what mine shouldn't be... too tight. I have been battling misalignment between raster and vector, nothing I did seemed to help - tightened belts, cleaning encoders, changing speeds, cleaning and lubricating rails, nothing. So, over the "holiday break" I took from the 20th of Dec to the 31st and tore into the machine. I replaced both the long X belt and the short belt and tightened them up pretty tight. Like yours, not quite guitar string tight, but tight nonetheless. I cleaned the rails to a point where they were spotless, both X and Y, and cleaned the carriages as best as I could. The machine got such a thorough cleaning it almost looked brand new again! Hopeful, I fired it up and - same alignment problems... So I thought that I'd loosen up the belts to a point where I thought it couldn't possibly work, just barely tight enough that they wouldn't continue to bounce when I press them down - they were l o o s e ! Amazingly enough, the alignment problems were virtually gone! I tightened them up 1/4 turn at a time and saw the alignment get progressively worse. I backed them off to about where I started and the alignment was back to acceptable. Over the years, 7 that I've had the machine, I have periodically tightened the belts a bit when I thought they felt loose. It's likely that they were never loose or stretching, and all I was doing was "solving" a problem where there wasn't one. My advice? Draw a filled circle and then one without an outline that's about .010 larger than the filled one. Raster the filled and vector the unfilled, at the same time, and see how the alignment looks. If it's ok then your belts are probably tensioned properly, if they are off then I'd suggest starting with them loose and test from there.


Kev Williams
01-23-2014, 6:46 PM
The program(s) that came with mine are PHcad and PHsoft. Best I can tell PHsoft is simply a newer version. But I don't use it because it's fraught with bugs, so I use the older PHcad. Anyway, I've seen screenshots of Laserworks(? I think) and it's basically the same program. What I'm getting at is the backlash adjustment is called "engrave reverse offset" (or "carving backlash in the newer program), in 'system settings/crafts parameters' According to the manual, they suggest some numbers that should be close for certain speeds. Actually, let me copy/paste from the manual (love the translations)...

Engrave Reverse offset:
The edges of figure may be unsmooth because of the extending of machine strap when the laser bilaterally scanning the picture. So we increase the reverse interval to for the amendment. There is certain reverse interval under certain speed; Generally speaking, the faster the speed is, the bigger the reverse interval is.

When speed is 200mm/s,the reverse offset should be 0.30mm;when the speed is slower than 200mm/s,the speed is proportional to reverse offset. Thus, when the speed is 100mm/s,the reverse offset should be 0.30*(100/200) =0.15mm. When the speed is 300mm/s,reverse offset is 0.50mm;when the speed is between 200~300mm/s,the speed is proportional to reverse;thus when the speed is 250mm/s,the reverse offset should be 0.30+(300-250)/(300-200)(0.5-0.3)=0.40mm;
When the speed is faster than 300mm/s,the reverse offset is the same as 300mm/s(0.50mm).

--get all that? ;)

Here's the deal with my machine: to get the offset to work correctly I have to put in NEGATIVE numbers. If I put in positive numbers I can get 1/8" or more of misalignment between raster directions!

During my last alignment test, as I noted above the end alignment was worse than the middle, but the LEFT edge was REAL bad. I did the test from the left side of the machine, so when the head was at the far left, the belt was at its longest, with about 47" from clamp to pulley. Stands to reason (to me) that the more belt available that's loose = more misalignment...

I do know a loose belt has it's place. Years ago, severely loosening the X-belt on my LS900 eliminated a banding problem I was having. But the compromise was I couldn't engrave any faster than 4" per second!