[MArruda]

Hello,

I've been working on my CNC router, built from a Fine Line Automation kit about a year and a half ago or so. Everything is going great, aside from some backlash issues with the Z axis.

My problem is odd- I hope others have expereineced it and can give some input. The backlash is being caused by the spiral cut motor to shaft coupler. With the Z axis carriage not attached, just the coupler shaft and upper bearing block, everyting turns fine. With Mach3 set to a .001 step per keystroke, everything moves fine with each step. When I attach the acme rod and Z axis carriage, there's .001 backlash- the motor turns with every keystroke, but the bottom of the motor coupler tied to everythin else stays stationary for one step- .001- of movement before it starts moving. With the router and dust collection shoe on the unit, there's a .002 takeup before it starts to move. You can visibly see the top of the spiral cut motor coupler turn and tighten the spiral like a spring before the rest of the assembly starts to move.

The problem with this, is that it's not repeatable backlash. The amount of "spring" being introduced by the motor coupler is variable when the motor is running, depending on if the Z axis is making a fast decel to reverse direction, etc. The problem is not exibited on the X and Y axes- I believe it's an effect of gravity and the weight of the motor and carriage on the coupler.

Any ideas? Would I lovejoy 3 piece type coupler be better for the Z axis?

Thank you,
Michael

[Al_The_Man]

If this is a helical coupler, is it aluminum instead of stainless?
Al.

[MArruda]

The couplers are these:

CNCRouterParts

I believe they're aluminum, actually, but could very well be wrong.

-Michael

[Al_The_Man]

They appear to be aluminum which is not the ideal type for stepper/servo drive, for applications like encoder they do well, but helical stainless would be a great improvement and takes motor flex much more without failure.
Al.

[MrWild]

i bought three like that from linear Motionbearings2006on eBay. Totally worthless. The ones off eBay had .800 (20mm) diameter bodies. Slinkies have better torque load stiffness. I've been looking off and on for affordable stainless ones with a 1' diameter. So far, no love.

[gio666]

All of the Spiral are not meant for torque ,and will get lots of backlash in the aluminium ones, you can turn in your hands and break in half ,best would be rigid coupling ,or this type is the best to use in aluminium and closed to rigid and still be a little flexible coupling you will find

http://x0cnc.com

[wantacnc]

Is your Acme shaft turned down and in a bearing block with angular contact bearings? Could you tighten the retainer nut to adjust the preload in the block? The helical coupler shouldnt have load on it only to transfer and couple the stepper to the rod, and to take up any alignment issue.

I could see it trying to tighten up if there is alot of drag or pressure in the z rails. Does your z slide smooth when moved by hand?

Im not a expert on this but just thinking and throwing it out there for discussion on your backlash
Joe

[RICHARD ZASTROW]

I usually use the stainless steel bellows type couplings. They transfer torque without backlash and still tolerate some minimal misalignment.

Dick Z

[leaveme]

@Michael:

I had similar problem in one of my machines. My couplers are own built helical couplers (L 30mm x D 25mm) from alu. I replaced them by steel but still some backlash is there. Then I discovered it is actually my ballscrew, which is flexing in the bearing. I locked the ballscrew at stepper end (bearing) and backlash is gone.

Just my two cents...

[MArruda]

Thanks for the replies everyone!

In the design, the assembly is motor->Helical Coupler->1/2" drill rod->Bearing block->ridgid coupler->acme screw->Z axis carriage. With the ridgid coupler unattached to the acme screw- thus taking the screw, carriage and router out of the assembly- there is no backlash. When the screw is reattached without a router on the carriage, there is .001" backlash. With the router, it brings it up to .002-.0025" backlash. It's definately the weight of the carriage and router that the coupler can't handle. You can visibly see the coupler "wind" up as the motor turns but the assembly doesn't.

Anyone know where to get bellows couplers in SS with 1/4" on one side and 1/2" on the other?

Thanks again,
Michael

[Connor]

Thanks for the replies everyone!

In the design, the assembly is motor->Helical Coupler->1/2" drill rod->Bearing block->ridgid coupler->acme screw->Z axis carriage. With the ridgid coupler unattached to the acme screw- thus taking the screw, carriage and router out of the assembly- there is no backlash. When the screw is reattached without a router on the carriage, there is .001" backlash. With the router, it brings it up to .002-.0025" backlash. It's definately the weight of the carriage and router that the coupler can't handle. You can visibly see the coupler "wind" up as the motor turns but the assembly doesn't.

Anyone know where to get bellows couplers in SS with 1/4" on one side and 1/2" on the other?

Thanks again,
Michael

I would look at your bearing block... You shouldn't have any linear movement at all with the ACME screw when it's mounted in your bearing blocks. The coupler should only have to deal with rotational forces.. I had to add spacers in my bearing block to remove the play. Thus eliminating all the backlash. The helix design is to allow for a slight offset or miss-alignment of the motor and screw.

Thanks, Connor

[MArruda]

Thanks for the reply, but I don't think you're understanding the issue- when the bearing block is attached to the motor through the helical coupler, but not attached to the acme screw through the ridgid coupler, there IS NO backlash. The issue is DEFINATELY the helical coupler- you can see and feel it winding up to take up the rotational torque before it starts to move, when there is a weight- the carriage and motor- on it.

Thanks,
Michael

[Connor]

Thanks for the reply, but I don't think you're understanding the issue- when the bearing block is attached to the motor through the helical coupler, but not attached to the acme screw through the ridgid coupler, there IS NO backlash. The issue is DEFINATELY the helical coupler- you can see and feel it winding up to take up the rotational torque before it starts to move, when there is a weight- the carriage and motor- on it.

Thanks,
Michael

The rigid coupler is taking out the backlash or slop in the bearing block and is bearing some of the load, no other way to explain it. Have you tried to move the axis back and forth without ANY coupler connected?

I made the same mistake with my setup.. I had around .002" backlash.. I mounted my coupler it in such a way that the helix was compressed and took out the backlash, but, it came back.. When I put shims between the bearings in the bearing block and re-assembled everything.. The backlash was completely gone. The ONLY other thing I could think of, is if maybe the shaft of the motor has some backlash in it.. Might want to check that as well.

Thanks, Connor

[MArruda]

Axis moves smoothly, no binding no problem. I can turn the acme with 2 fingers to move it. That's not the issue. The assembly is motor->Helical Coupler->1/2" drill rod->Bearing block->ridgid coupler->acme screw->Z axis carriage. When the ridgid coupler is not attached to the acme thread, it's free to spin, so it's not taking up any type of backlash.

Visualize this- you have mach3 set to .001 per keystroke for manual shuttling. With everything connected and the router in the carriage, you tap the button to move the z axis 2 times. each time, you see the motor move, but the rest of the assembly does not- you can visually see the helical coupler turning on one side and not the other, like holding a spring by the ends and turning them opposite each other. On the third keystroke, the axis finally starts to move. This is what I'm experiencing. As others have said, the coupler should NOT be flexible enough to tighten like a spring those .002". There's not other reasonable explanation- that coupler is the only weak point in the system.

-Michael

[CarveOne]

Helical couplers are better suited for absorbing the shock loads from motors starting and stopping when the mass of the load is high. In many applications (such as on compressor motors) the backlash is not as much of a concern as it is in CNC applications. There are brass, and stainless, versions of the helical couplers that don't have as much of this effect, yet still have some minor shaft misalignment compensation capabilities. Many people here use them for CNC, and prefer them in their situation. I use the DumpsterCNC couplers on my smaller machine.

CarveOne

[MArruda]

I've been looking around for another coupler solution. What do you guys think about these Stepped beam couplers?

http://www.rinomechanical.com/pdf/co...3_DEC_2010.pdf

They're still aluminum, but being as they aren't helical cut, I'd think they can't twist like the one I'm using now. I got a quote of $30/ea shipped for these. Seems to be going rate for an inexpensive coupler- Spider, Oldham, etc. I just can't find anything else that will fit and do what I need it to do- it has to be .25 bore on one side, .5 bore on the other, and be no more than 1.3" in diameter.

Thanks,
Michael

[jsheerin]

They do twist. They have a spec for it right in there - torsional stiffness. If you used the 1/4" ones, 380 oz*in motors, and 1/2-10 5 start screws you'd get around 0.000267" positional error under max torque from your motor. That plus the reversing torque rating makes these sound acceptable for typical usage with Nema23 size motors. I don't have any personal experience with them though.

[wantacnc]

Mike, Is your motor in perfect alignment with the acme shaft? If it is I can make you a coupler with 1/4 to 1/2 bores with out any cuts in it to see if it cures your backlash. 1" diam by what ever length you need. Let me know. I have aluminum or stainless.
Joe,

[awhtyp]

Hi all,
I used disk couplings from "zero-max", they are a very nice product and are not to badly priced.
They are made in a large range off sizes, and in single or double disk configurations.

[msimpson99]

I had the same problem with my latest build. Make sure the slats are compressed. This removed 100% of the backlash problem I was having.