What is the line count of your servo's encoder? It would probably be a quadrature encoder so multiply whatever the line count is by 4 and that should be the steps per revolution.

Dan

It's not only the velocity under Motor Tuning that needs to be right to get the right speed, it is also the Steps per. Also check the pulley settings - set it to 1:1 and set the max speed to 3000RPM.

Dan

Tweakie,

Thanks for taking the time to explain it.

Dan

A left a question unanswered some time back, so here is the answer: For my previous little X2 mill, in order to run the spindle at 7,500 RPM, I simply attached an old Pentium 3 aluminum heat sink to the side of the head. That was enough to keep the beargins in spec. Low tech, cheap, ugly, but completely effective.

Thanks Steve.

Daniel

A good looking one, Tweakie.

How does this etching work? How does it produce the required shade? Doesn't it just melt the acrylic?

Think you can do 3D shapes in a glass cube like those sold in souvenir shops? That would be lovely. But then you need to focus the laser at different heights -  can you do this?

Dan

Not a full enclosure but the table has a guard round  three sides, the front having a removable perspex door, so no oil skins required

I have a DA pencil collet that I was contemplating using as a spindle but that ER is a great price.

Hood, beware of these Chinese collet holders. Bought few BT30 holders (from another Chinese seller though) and accuracy is not something these were built for. The quoted 0.01mm runout appeared to be not the worst case like you'd expect, but the best tolerance they can achieve. Actual runout measured on 5 holders ranged from about 0.02mm to 0.25mm (I didn't miss a zero there)!!!

Did make up a spindle speeder of sorts which was geared via a toothed belt but it got too hot to touch after about 3 mins running and the heat seemed to be coming from the belt friction on the pulleys, was thinking of directing some coolant through the body but not sure how the bearings would like it, would have to filter it and may give it a go.



Hood

Round tooth profile belts perform better at higher speeds. They engage better with the pulley, with less friction, and thus less heat (and noise) is produced. But I assume this is what you already have... What speed are you running the driving pulley at? And what diameter is it? I have 75mm diameter pulley that would run for hours at 3500RPM without getting that hot. I am sure I could run it even at 5000RPM much more than 3 minutes and not get that hot. May be check the tension?

Anyway, the best choice for a high speed application is a V-belt, or any belt other than timing belts. Or put it another way, timing belts are least suitable for high speed applications because their teeth engagement and disengagement with the pulley teeth produce a lot of noise and heat at high speeds. You don't really need a timing belt there - no angular positioning needed and no big moments to be transferred - so may be put a V-belt or even an O-ring belt and see if it helps. It really should, unless there is another source for the heat.

Daniel

A good looking design, Amir!

I think I would use solid plates and bars for the body, rather than profiles - gives more rigidity and better absorbs vibrations. Think it would be a benefit for cutting aluminium.

Dan

Interesting build! Would love to see it in action.

Thanks for showing the pictures.

Daniel

So Guys, correct me if I am wrong, but would it be right to say that the striations are being caused by the lack of resolution in the exported arc ?.

Tweakie.

That was my guess right from the beginning, but I thought you were saying that the striations did not match the Gcode...

Dan

OK, if it's not a sequence of straight lines to reflect the G-code and the striations are more pronounced than they should be if it were the G-code, then my guess is that the machine has some dwell between each line segment and the laser melts the material leaving those marks. Could you see if cutting in "Exact Stop" mode makes them more pronounced? I think that could prove the above point...

Also, may be playing with the Lookahead value will show some difference.

Dan