[Carniflex]

I am planning a 100 eur budget CNC laser burner. The current plan is to get the parts cut from one of the local machine ships with a laser cutter out of 2 or 3 mm steel or aluminium. Probably aluminium as with that budget I'm not gonna get very strong steppers. Due to budget contraints I would have to use regular threaded rods as lead screws, however, I would like to minimze the backlash somewhat. So would something like this:

work? Idea is two disks, bother together with M3 bolts and having 2 tabs that prevent the larger bolts from turning. Spring between the nuts to apply tension. It should work .. based on my google searches, but better ask somewhere where people probably have tried something like that already.

[awerby]

That doesn't sound like a great idea. The problem with using cheap threaded rod with sharp threads rather than the trapezoidal threads of an acme-style screw is that it is resistant to turning in a nut, with some parts a little out of spec, so it tends to hang up rather than feed smoothly. Using double nuts would make this problem twice as bad. With the weak steppers you plan to use, that pretty much guarantees that this machine will lose steps - a much worse problem than backlash. For 100 euros, you'll be lucky to get a machine that works at all, let alone perfectly. (Let me guess, you plan to use the cheapest Chinese drivers off of ebay, and the crappy steppers they sell with them, right?) Save the anti-backlash nut idea for the next build, when you decide to spend a little more, use acme screws and better drives, and get a machine that actually does what it's told.

[aarggh]

As Andrew has pointed out, you have severe problems with your budget. If you plan on using the cheap TB6560 boards on Ebay you'll also have a system that will likely never work as these are absolute garbage, and the steppers they supply usually have such a ridicuously high inductance that makes them useless.

Check out the driver links from pmimno or massmind if you can't afford a Gecko driver, they are the best you'll get for lower price range drivers. The TB6560 are simply a waste of money.

Given the steppers and drivers are the lower portion of the cost of a cnc build though, you'll be struggling to make anything that works satisfactorily to any degree for less than around $750-$1000.

cheers, Ian

[crazydiamond]

Why not scrap the idea entirely and use a cheaper method of power transmission? A simple belt driven axis will work perfectly for a laser driven machine (as there are no "cutting" forces, and a low moving mass...most 3D printers use belts and timing pulleys). When setup correctly, you can achieve very little backlash, super fast rapids and great repeatability, and its relatively cheap. You just need to size your motors appropriately and think about gearing.. Just my 2 cents. Also the guys are right about those TB6560.. Way too hit or miss in my opinion

[elfrench]

You can make a zero backlash nut for threaded rods very inexpensively using the method detailed here: Making Acetal leadscrew nuts the easy way. Most threaded rods have rolled, instead of ground, threads, so they're not very smooth. I'd recommend lapping them to make them smoother and possibly more accurate.

Here's a before shot of some 6MMx1 allthread: Attachment 263200
And the same allthread after lapping with a cast polyurethane lap:Attachment 263202

I still wouldn't recommend using threaded rods because you just can't drive them fast enough for a laser.

[Carniflex]

Thanx for the useful information.

The steppers I was planning to use are CNC Nema17 Hybrid Stepper Motor DC 5V 2 Phase 2 3kg cm 4 Lead 1 8 Degree 3D | eBay - which promise approx 0.225 N.m after converting the number to SI system. Now it's entirely separate question if they can provide that in reality as well. It did sound like enough to just move x and y axis around, but as I'm still in CAD phase with the machine I have not done the math if that kind of torque is sufficient.

The budget is for just the machine part - the plan is to drive the machine with linuxcnc directly from parallel port. I have still reading up to do in this regard to see what is feasible and what is not. Budget is picked that low because this is more sort of proof of concept for me, not a machine which I intend to use professionally. That does not mean, however, that I have to be sloppy with what I want to achieve. The machine would be relatively small - somewhere between 300x300 mm to 500x500mm working area. At first I would use just a 500 mW 808nm diode and if it works out well enough might upgrade to a stronger diode, say up to 2W or so.

For stepper motors I was hoping to use H-Bridges. L9110S Dual DC Motor Driver Controller Board H Bridge Stepper Motor for Arduino | eBay The L9110S seemed sufficient for it.

I did consider a belt drive briefly but at first glance did not find enough information in regards of these so that is why at first glance it seemed the regular M threaded rod might be better option on tight budget. However, that plan is not written in stone. If I would be able to get away with a belt drive it would be certainly be a viable option as well.

So in a nutshell DIY anti backlash with regular M-threads is bad idea as far as I understand from the replies so far and something like a belt drive might be more reasonable approach on such a tight budget.

[elfrench]

You may find a better value for stepper motors here: Stepper Motor & Stepper Motor Driver - StepperOnline I have received several motors from them, but haven't built the rest of the machine.

You should probably spend some time doing the research on how fast you need to move the laser head for the materials you will be using and then decide on the method of moving it.

Here's a little 322 page thread on a belt drive: http://www.cnczone.com/forums/linear...570-forum.html

[Carniflex]

You may find a better value for stepper motors here: Stepper Motor & Stepper Motor Driver - StepperOnline I have received several motors from them, but haven't built the rest of the machine.

You should probably spend some time doing the research on how fast you need to move the laser head for the materials you will be using and then decide on the method of moving it.

Thanx for the link. Seems pretty good shop. Although at first glance it seems that their shipping costs are significant to my location. A shipping (without the price of the motors, the cheapest option) for 2 motors similar to the ones I have looked up on ebay would cost more than these 2 motors from ebay with free shipping so the total cost difference would be significant. I will ofc consider them seriously within the budget I have arbit<rily picked for myself. Also thanx for the belt drive link. It will be a good sport to start reading up about these

[Carniflex]

I have an additional question. How relevant is friction for the lead screws/nuts? As one of the previous replies mentioned that using DIY anti backlash nut setup would introduce additional friction. I am asking because based on one of the online calculators it seems that this additional friction should not be sufficient to make stepper motor to skip a step in modest speeds. Although ofc I have no idea how fast does the torque drop off when the rpm increases so perhaps I'm am overly optimistic in this regard?

The calculator I found: Lead Screw Torque and Force Calculator

Lets say I would use M8 threaded rod. Which would mean the lead-scew diameter of 8 mm and 8 threads per cm. The stepper I think of using (one of the cheapest Nema 17 in ebay) claims it has 2.3 kg.cm which is roughly 22.5 N.cm - based on that claculator such a setup would be able, in theory, to rise a load of max 18 kg (probably very slowly as torque drops off when rpm increases) that is with max friction coefficient of 0.25 for the steel. Assuming the friction coefficient of 0.5 it would still be able to rise max load of 10 kg.

Such setup would have 1.25/200 ~ 0.00625 mm minimum theoretical resolution (not expecting that kind of resolution ofc in reality, but would still be nice to get, say 0.05 mm or so). At, say, 200 rpm (is this a realistic rpm even for a stepmotor?) I would have a linear speed of 250 mm/minute which is a bit on the slow side even for a ~500 mW laser cutter? At 500 rpm the linear speed would be 625 mm/min but I'm not sure if there would be any torque left at that point. Perhaps still enough left to push/pull, say, 4-5 kg on x-y axis without skipping a step?

So it should be adequate for driving X or Y axsis with modest speeds? With only a 500 mW laser I would probably not be able to do anything very fast anyway?

So that brings me back to my original question - would such setup as posted in the initial figure work as DIY anti-backlash nut? Or am I barking up the wrong tree in there and should just go with the belts for that kind of smallish machine to play around?

Unfortunately ACME / trapezoidal lead screws are far too costly for that kind of budget I have in mind.

[elfrench]

The acetel nut would work far better than your original post. A even better method would be to start with a CoreXY 3d printer and use the laser head instead of the plastic extruder.

[Carniflex]

The acetel nut would work far better than your original post. A even better method would be to start with a CoreXY 3d printer and use the laser head instead of the plastic extruder.

Thanx for pointing out the CoreXY setup. That is quite clever way of doing it it seems. I'll sit and think about it for a little while for sure.

As far as Acetal goes I have no easy access to the tools needed to make one, although I will figure something out if I for some reason decide to still go with regular metric lead screw setup.

[Eldon_Joh]

i have lost my sources again, but i had to go back to 1900 to find real data, not some math whizz's explanation.

basically an acme screw at 29 degrees gets you the same force for 25% less torque than a 60 degree thread of the same helix.

however, the 60 degree screw thread has 40% more surface area of contact... so depending on how proper the lubrication is the screw thread can do better in some circumstances.
for example, if you cut a lead screw nut from sintered bronze, a 60 degree screw thread might last longer, it depends on the velocity. for slow speeds and heavy loads the acme will well better.. you can wear 20 mills backlash into a 1/2-10 acme screw and the 29 degree angle will turn into 35 degrees, but it will still have about the same contact area and it won't strip out or fail underload.

a 1/2-10 60 degree screw will wear and the contact surface area will decrease as it wears. for acme this isn't so much the case due to the trapezoidal form.