Gerber 408: Y-axis error when travelling at speed

[Mikexx]

Where the carriage moves with up arrow, ie away from me and the keyboard it can move at speed without error.

If the carriage moves with down arrow, towards me, I will get numerous errors. I can use "Reset", press down arrow for the process to repeat where the carriage moves, stops with beeping.


It's notable that it doesn't happen at low speed, such as when orienting.


Ideas and suggestions would be greatly appreciated?

 

[Vassago]

Hmm.. I would think that one of the gears or sensors is loose - almost acting as a one way ratchet due to wear. Going by the issue I'd think it would be one secured by a screw as with wear, that could give rise to such an issue.

Another possibility could be if it's not level, combined with a servo or power supply that's on its way out.

Due to the age of the machine, it's very lightly to have issues with electrolytic capacitors failing which could also cause the issue.

If you've got a multimeter, check the dc outputs of the psus for a ripple and check the machine for level, then try to manually move the gantry when powered off to see if there's any movement

 

[Vassago]

Should say ac ripple not a ripple - autocorrect is the bane of my life. Lol

 

[Mikexx]

Hmm.. I would think that one of the gears or sensors is loose - almost acting as a one way ratchet due to wear. Going by the issue I'd think it would be one secured by a screw as with wear, that could give rise to such an issue.

Another possibility could be if it's not level, combined with a servo or power supply that's on its way out.

Due to the age of the machine, it's very lightly to have issues with electrolytic capacitors failing which could also cause the issue.

If you've got a multimeter, check the dc outputs of the psus for a ripple and check the machine for level, then try to manually move the gantry when powered off to see if there's any movement

Thanks for the ideas. I have checked the linear bearings, cleaned and re-oiled them. I feel the carriage moves smoothly and that this ought not be the source of the issue.

My next idea is to take a look at connectors. I think the Y-encoder signals are reversed compared to X and Z signals, otherwise I could swap the connectors. Will check on this point.

After that as you suggest it's a look at the tower and possibly swap the driver cards between X and Y? Possibly swap X and Y cables to the driver.

I have a scope but not sure if it's a Power Supply issue as the other axis would surely be affected?? However ........

Thanks again.

 

[Vassago]

Can be either drivers or psu - faster movement = more power required. When they're getting past their sell by date which this one will be, they tend to loose capacitance so more ripple gets onto the line.

However..that said, as you said it only happens one way - I'd still think it might be a sensor or gear. Aiming more towards a gear.. It's riding over the grub screw dent one way and jamming in the other.

If you can remove the motor, just see if there's any play.

 

[Mikexx]

Can be either drivers or psu - faster movement = more power required. When they're getting past their sell by date which this one will be, they tend to loose capacitance so more ripple gets onto the line.

However..that said, as you said it only happens one way - I'd still think it might be a sensor or gear. Aiming more towards a gear.. It's riding over the grub screw dent one way and jamming in the other.

If you can remove the motor, just see if there's any play.

Many thanks for your reply.

My approach has been to isolate the motor and have removed the belt and toothed pulley.

When removed, and so disconnected to the carriage, it will run from one end to the other without issues only stopping at the programmed end limits. No errors. I feel that discounts any potential issue with the encoder.

The nut does feel a little stiff but can be moved. I'm therefore checking the whole mechanism associated with the nut and nut assembly etc. Fingers crossed I can free it up and it's something simple.

 

[Vassago]

Fingers crossed..

Tbh most things are "relatively" Simple.

The nut should run very smoothly.. Ball nuts have a propensity to collect gunk - quite a few YouTube videos on how to clean and reassemble. Likewise linear bearings work in the same way.

Could be a combination of things where old servo drives can't quite provide enough power to move a stiff nut. If you do disassemble it - put it in a plastic bag before you remove it.. Nothing worse than balls disappearing.. Lol

 

[Mikexx]

Many thanks.

I took the assembly apart but didn't remove the nuts from the lead screw. After a general clean it moves further before error but I still get an error when moving at speed.

You mentioned, "Ball nuts have a propensity to collect gunk - quite a few YouTube videos on how to clean and reassemble". I can't find any for the Gerber Sabre and not familiar with ball nuts.

The carriage moves freely by itself and TBH I wasn't aware the Gerber 408 used ball nuts? I can see a second nut that is under slight tension from a spring to remove backlash, but nothing to suggest either nut is a ball-nut?

 

[Vassago]

Most of the 408s I've seen have been upgraded with ball nuts, some still use acme thread screws - these use a 2 part nut with a spring to remove backlash. It basically pushes both parts apart to take out play..

What would you might have is debris in the nut, or the spring might have lost tension so that it opens up (reduces tension) in one way and closes up in the other, when in reality there should be very little moment as all it's meant to do is take up the very small amount of backlash present.

See if the spring is stuck or if you can move either part easily. There should be a moderate amount of pressure.

Move the axis and see if you can see the nut move. There shouldn't really be any gross visible moment. If there is, the but will tighten up and cause the error you're experiencing.

This type of screw/nut combination is great when there's no contamination. So as soon as you get gunk it can embed itself in either the nut or screw ( which will usually have a coating which gets damaged) - either means a good clean and disassembly.

Older 3d printers used to use a smaller version of this

 

[Vassago]

This shows a similar type of nut to give you the idea of how it works

 

[Mikexx]

Many thanks for your input.

This is the lead screw nut and torsion spring.

This was removed and cleaned. It comprises of plastic parts, not sure what material. The RHS has the same thread as the LHS but can move in and out. The spring provides a low level of torsion to ensure both nuts are in tension wrt each other.

It now seems to perform without error. However, I couldn't see anything wrong yet presumably the two nuts would lock-up when travelling in the backwards direction.

It's strange that I can't find any info on the nut used on the Gerber Sabre 408. For those interested in the type of anti-backlash nuts:

This is obviously a Torsional Anti-Backlash (TAB) nut.

 

[Vassago]

They are very simple things.. Essentially just two nuts with a spring pushing them apart to take up the backlash.. The spring force must be enough to cover any forces exerted by the system.

If the two nuts are forced to move by external friction (gunk, wear, etc), then the locking force will increase - just like tightening two nuts against each other.

I mentioned on another thread that you can get clock spring type screw protectors - they can be very useful to prevent gunk getting into the system to start with.

Ball screws are the way to go though. Just generally nicer..