[yassaa]

Hi, I built a home-made 4axis router couple of years ago with mach3 control, leadshine servo-stepper motors and NSK linear bearings. For breakout board I used purelogic PLC4x-G2 interface. Machine backlash can be as little as 0.01-0.02 mm when nuts are preloaded and set (measured with dial gauge), or 0.2-0.5 mm after many hours of hard work (but it can be set again in mather of minutes to 0.02 and hold it for a long time) . The reason I am telling all this is because PLC4X broke last month, and I replaced it for time being with cheap chinese BoB which I had in reserve if such an occasion happened. At first, I didn't notice any difference between work done with PLC4 and with chinese BoB, but when I started new batch of some enclosures I do for some electronics I noticed that lettering is not good at all! Some letters are crooked, letter O does not connect start point with end point, almost everything points out to backlash of 0.5mm or more where is none. When I started cutting openings in enclosure, I noticed that circles are not round, but actually diagonally oval. Those are three holes for LED diodes which are 5mm diameter, which should be circular and fit to diodes but are so crooked that they have become unusable as pristine condition enclosures.
Yes, I read all posts regarding circle not round question, search the internet in three languages, but nowhere I found problem similar to mine so if I missed it forgive me.

The facts are:
- No backlash (0.02mm, negligible)
- Circles used to be true couple of months ago when I was doing exact same work, so I know it is possible (even with more than 0.02 mm backlash)
- The same happens regardless of the working speed - when set to exact stop or in CV mode, even with CV dist. tolerance set to 0.1 and motor acceleration set up to 1000 mm/s. After many combinations of accelerations I managed to make the circle somewhat round, but it still isn't as it should be.
- Even with that 'round circle' settings, letter O which is the same size as the holes (5 mm wide) looks like spiral, every angled straight line looks like arc
- The only change on the machine was BoB, all other settings are the same.

The assumptions are
- Y axis lags behind X axis. Why would it do that? No clue..
..or - Chinese BoB is f.ing me from behind while hitting me in a head with one arm and pulling my ears with another!!

I want to know, is there a possibility that BoB is doing all that mentioned above? Everything is cleaned and lubed, minimum deflection, position following error on drivers not changed since it started working and even after hours of working it still hold the position and repeats every move perfectly, always!! The problem is that the moves (arcs?!) themselves are wrong and parabolic.

Help me, please! I want to upgrade to pokeys57cnc or smoothstepper, but I need to figure out what is going out right now before I do that and curse smoothstepper for something which is not its fault.

Thank you in advance, Jasmin.

[komatias]

Jasmin

Yes, it can if what you are feeding it is not great. Some BOBs work on a low voltage from the parallel port (around 3.5V) ok others do not and need 5V. Get a Smoothstepper to ensure you get good pulses. I use cheap chinese BOB's on my tools to.

Regards

[wizard]

A couple of things to consider :

1. Fonts are not a good way to test a machine because there is no assurance that the "O" character is actually round.
2. Your best bet is to have the CNC controller draw a circle from manually entered G-Code. If you can't do that use a cad cam system to create a circle to be machined out. It would help to do a series of shapes, circles, rectangle and squares.
3. Don't dismiss the possibility that something has come loose. Pulleys and couplings can rock on their shafts creating hard to diagnose problems.
4. A faulty break out box is always possible, however I wonder what happened to the old one? It is possible that the port itself has issues.

[yassaa]

Actually, these particular fonts and holes are from a job well tested and done at least couple of thousands times before, so they are pretty good reference point, and I also tested drawing circles with several g codes, but those are pieces of advice worth to remember and learn by heart.
Anyway, I found the problem last night after disassembled the gantry to small pieces. One of the four channels in one of the four linear bearings got stuck so hard I needed hammer and piece of steel wire to get the balls moving again! I managed to get back almost all 200 balls back, and I noticed that with one ball missing in each channel gantry moves smoother and at least 20 percent faster now! The machine is precise now, no measurable or visible faults in work, so I am sorry to all hard working Chinese people involved in manufacturing of my BoB :wave:, they were wrongly accused. Thank you for helping me