Trouble with A axis

[tugrulmatras]

Hi everyone,

I built my kit a month ago and made some trials. One thing that worries me was the inconsistent wobbling. So I decided to make modifications after reading forums "Better nozzle with auto change" At the end, I managed to reduce wobbling to a negligible point. I don't thing needle calibration is needed anymore, this is the good news. The bad news is A axis is not working properly any more which makes the machine unusable. Probably I messed up with tinyG settings or wire connections, or else, It doesn't rotate to exact point. I made lots of trials changing motor4 settings in TinyG, I couldn't make it work properly. The weird thing is When I type for example 90 Deg in manual coordinate input, it rotates to a point and when I type 0 Deg, it doesn't come back to where it started. By changing settings I manage to make it rotate 90 Deg. but when I try 180 or other angles it doesn't go to exact point. And never gets back where it starts.

I would appreciate it if you give some advise to solve this problem.

Best regards.

Tugrul

[JuKu]

> When I type for example 90 Deg in manual coordinate input, it rotates to a point, and when I type 0 Deg, it doesn't come back to where it started.

No matter what the settings are, this should work. Even if your 90 degrees might not be really 90, it should be able to get back to where it was. This can be for several reasons:

- Anything loose? Check the belt tightness and the pulley screws.
- Belt slack? The belt wants to be a circle, so there are some forces on it. Try first going from -10 deg to 0. Note the position. Go to 90, back to -10 and then to 0. This way, the rotation approached 0 position both times from negative direction. If there is slack and you don't get it go away by tightening the belt, you need to enable slack correction for A in the software.
- Rotation resistance: If the tube doesn't rotate freely (the belt and the flat spring cause some resistance anyway) and the motor current is not sufficient, the motor just doesn't have enough power to do what you ask it to do. Increase the current but watch out, the TinyG can put enough current trough the small motor to kill it. Make sure the motor doesn't run hot.

Since you have changed the TinyG settings, you a) can read and understand these and b) might have gotten something off there. For your reference, I uploaded a settings dump from my machine to http://www.liteplacer.com/Downloads/Goodsettings.txt.

[mawa]

More trouble with the A Axis:

I don't know what happened but since a couple of days my A-Axis pipe occasionally gets stuck while probing - that is it won't move all the way back down and the Z-Limit switch triggers when ProbingMode is set to false ...and you all know the rest what happens after that.

I already adjusted the spring and the limit switch.

Should I lubricate the two bearings and what lubricant should I use ? (oil, teflon spray.. )

[mrandt]

My suggestion - although Juha might not like me for it - is to replace the bearings and the tube.

The Misumi LHFC8 ball bearings used in Juha's kit are intended for linear and not rotational movement. The bearings currently only work for this purpose because the tube is not 8mm but only 7.8mm so it does have a bit of play. The play is undesirable on a PnP, thus it is eliminated using the metal spring.

Now, with the motor belt pulling to one side, the spring pushing to another and a bit of play, chances are that the tube will eventually tilt and get stuck.

You may use a bit of lubricant to ease things (sewing machine oil works nicely) but this is not a permanent fix.

I suggest you get standard bushings instead, for example Misumi JFRA8 (cheaper than Juha's part by the way...):
http://uk.misumi-ec.com/vona2/result/?Keyword=JFRA8

They nicely fit the original mount locations:

bushings.png (1.78 MiB) Viewed 7734 times

In combination with those I suggest you also get a proper, precision made, exactly 8mm hollow shaft.

If you want to make it perfect, have M5 threads cut into top and bottom so you can permanently attach a rotational joint to the top and metal luer connector to the bottom like this:

steel_tube.png (605.16 KiB) Viewed 7734 times

I do use a 90° angled rotational joint now, but this gives you the idea.

I can recommend Misumi part RTCNL6-M5 for the joint - has M5 thread on one end and 6mm tube push-in connector on the other:
http://uk.misumi-ec.com/vona2/result/?Keyword=RTCNL6-M5

Misumi could also make the shaft for you, including the threads. Custom part number is FSPJBB-D8-L200-M5-N5:
http://uk.misumi-ec.com/vona2/result/?K ... L200-M5-N5

Takes a couple days (made in Japan), will cost around 18-19€ plus tax but you will get very nice quality.

If you like the metal luer connector, shoot me an e-mail or PM and I'll send you one in the mail for the price I paid (have to look that up); I still have a few parts extra.

Long story short: With a precision shaft and matching bushings, you will have almost no play. It is a bit of a challenge to align all the parts and get everything perfectly straight; but once you manage this you can do away with the metal spring, reduce tension on A-motor belt and use a less forceful spring.

I have never had any problems with the A axis since I made this modification.

[JuKu]

Good upgrade. It would add about 60-70€ to the price of the kit, but might well be worth it. The only issue is as you mentioned:

> It is a bit of a challenge to align all the parts and get everything perfectly straight...

I need to try this.

( BTW: > cheaper than Juha's part...
It isn't, but no big deal, your post is so valid anyway.)

[mrandt]

The only issue is as you mentioned:

It is a bit of a challenge to align all the parts and get everything perfectly straight...

I tackled this task as follows:

- I drilled / filed the mount holes for the bushings in the pnp head top and middle plates slightly larger

lp_bearing_plates.png (64.83 KiB) Viewed 7725 times

- I then took extra care to make the mounting bracket and bottom plate exactly parallel to each other and perpendicular to the makerslide

lp_bearing_mounts.png (20.79 KiB) Viewed 7725 times

I then added the top plate and bushings but did NOT fix them yet. With the enlarged holes you can move both bearings and top plate in X and Y axis until their position is just right and the bushings align well. I inserted the shaft and used a precision water level to got it most upright / perpendicular in reference to ground - this of course depends on your table being (almost) perfectly level as well. My LitePlacer table has adjustable feet for this purpose. I fixed the screws one after another, always checking that the shaft / tube could still move freely (up and down + rotation) and was still upright.

lp_upright_check.png (34.05 KiB) Viewed 7725 times

A nice side-effect of this procedure is that you can be sure that the tube is not tilted vs. the Z-axis, so you avoid a Z-dependent XY-error.

BTW: > cheaper than Juha's part...

It isn't, but no big deal, your post is so valid anyway.)

Sorry, my mistake. I mixed up the lines in my BOM... You are correct, the LHFC8 is actually cheaper than the JFRA8.

Good luck with your experiments, I think this is quite an improvement for the current head design.

Best regards
Malte

[mrandt]

BTW: Why does phpBB always add those naste scroll bars to pictures? Can this be switched off in configuration? It is really annoying

[Pixopax]

To make it 100% exact you should use one of these:

Clipboard01.jpg (52.37 KiB) Viewed 7720 times

Just put it on the table, the sensor pin to the axis, and then drive it up and down, adjust until it is straight up.
This is the only way to get a vertical axis 100% vertical in all directions.

[mrandt]

Yeah, I should really get one of those...

However, one of the ideas of LitePlacer was that you shouldn't need specialized tools to build it

[Pixopax]

I think, it is essential to have that axis straight, if it is not, you will have position errors depending on the height of the needle.
In general, it is difficult to make such a machine without some basic measurement tools. Measuring some basic points and adjusting them will make it much better.
Since this dial gauge is not too expensive (ca. 25 Euro), one should consider buying one.