Community Favorite Alternative to the Lead Screw Upgrade Kit

Please note that my lead screws are stuck in customs and I wasn't going to post this until they arrived and I could take pictures of the whole setup, but I really believe it will help a lot of people out here.

First let me say that this is based on the work that Robo3D and @Marquis Johnson has already done to provide an upgrade. I believe this will solve the issues for most and provide a much cheaper and easier installation that either of those two methods. There is no need to re-wire or any removal of stepper motors involved in this upgrade. As a result of reusing more parts, the total cost can be quite low with judicious shopping from multiple sources. A kit would be nice, maybe someone could assemble one.

What is required:
You will need two 300 mm TR8 lead screws and lead screw nuts. The most common size available is TR8*8, which provides 8 mm of travel per single revolution. Other lead screws will require other firmware modification but the assembly is the same. Next you should purchase two solid motor shaft couplers. We need to eliminate play in the motor shaft interface. This is exactly what the integrated linear steppers (the stepper motor with the built in lead screw) do. Then you will need to print two of the TR8 lead screw adapter model. The adapter model is posted on Thingiverse, http://www.thingiverse.com/thing:1147438.

The adapter was based on these brass lead screw nut dimensions:



Solid motor couplings require a 5 mm bore at one end and an 8 mm bore at the other. I would recommend using the kind that have set screws rather then the compression style as they will hold the piece regardless of slight variances in dimensions.


The lead screw itself will look something like this:


And the adapters once printed will look like this and replaces the Z-Axis Stabilizer:


You will also need 2 or 4 M3 x 8 mm self tapping screws (#4 x 3/8" may also work) to hold the adapter in place.

How to put it all together:
I will not explain how to disassemble you Robo, there are plenty of threads that describe this. It is assumed you will either know how or can search for instructions on disassembly.

1. Insert the brass lead screw nut into the bottom of the adapter. Use the M3 screws to hold it in place. Although the dimensions are such that it should be a friction fit, we do not want any chance of slippage. Use 2 or more screws to secure the brass nut into the bottom of the adapter. Attach the two Z axis microswitches, that you removed from the Z Axis Stabilizers, to the front of the adapter with the existing screws.
2. Attach the solid motor shaft coupler to the stepper motor 5 mm bore down and tighten the set screw.
3. Insert the TR8 lead screw through the hole in the X-Axis Idler and then thread the lead screw nut / adapter assembly underneath.
4. Place the end of the lead screw into the 8 mm bore of the solid motor shaft coupling and tighten the set screw.
5. Spin the brass nut up the lead screw until it is partially inserted in the hexagon shaped hole in the bottom of the X-Axis Idler.
6. Reattach the Z-Axis microswitch wiring
7. Repeat steps 1-6 for the opposite side.

Final Assembly:
Once you have everything back together, check that the microswitches touch and are engaged when the adapter is fully inserted into the hexagonal hole. Please note that this model purposely has a tapered hexagonal protrusion to gradually consume all of the slop that Robo built into their use of threaded couplers with the M8 threaded rods. The taper is to remove as much of the backlash as possible. Backlash is a contributing factor to Z ribbing and the Robo upgrade does nothing to address it. This could be the reason there is a varying level of success even with the upgrade installed. If the fit is too snug, just a few minutes with a file will take down a bit of the hex dimension. If your printer is calibrated for dimensional accuracy over looks of the output, you should have no issues. There are various 3D model source formats included with the Thingiverse item for modification based on your particular setup.

You will need to perform a paper level of the bed, this will ensure that the switches are engaged and are at the correct height on both sides of the X axis. And check that they disengage freely. Print these adapters with a filament that has a lower coefficient of friction than PLA or ABS. I would suggest Nylon or PETG as the preferred filament type.

Before your first print:
Now there is one last thing that must be done to assure that you are set up correctly. You must modify and upload a new version of Marlin. This guide is not intended to explain how to upgrade your version of Marlin, it is assumed you know or can search for instructions. Look for the following line in Configuration.h

Code:

#define DEFAULT_AXIS_STEPS_PER_UNIT {80,80,2560,723.38}

and change the third parameter (2560) to so that it reads as follows

Code:

#define DEFAULT_AXIS_STEPS_PER_UNIT {80,80,1600,723.38}

This will ensure that the Z height of your models print correctly. The number listed above is for the most commonly available TR8 lead screw. If your lead screw has a different specification then use the calculator located here to provide the correct number of steps above: Prusa Lead Screw Step Calculator.

One last thing about moving from threaded rods to lead screws of any number of starts. Since we increasing the rate of travel per revolution, finding exactly the correct spot for each lead screw with regard to manually leveling the bed becomes more sensitive. I little more concentration is necessary to get the level just right.

Note: Once I have everything installed on my particular machine, I will update this thread and provide additional pictures.

 

Parts being delivered today, I will update with photos.

The cost is surprisingly small for this upgrade, why spend $100 when you can spend $25-40. Here are some links. Don't forget that you also need a few screws and the cost of printing two of the plastic parts, plus your time:

As cheap as possible links, total cost $13.08:

1. Leadscrew w/Nut:http://www.banggood.com/3D-Printer-...r-Motor-p-1112698.html?p=T72314257377201308LP (2 @ $4.94 each)
2. Solid motor couplings: http://www.banggood.com/5mm-8mm-Sha...Spanner-p-1070710.html?p=T72314257377201308LP (2 @ $1.60 each)

Amazon, total cost $38.44 (please note that Amazon ships these from China, so don't buy from these links):

1. Leadscrew w/Nut: Nothing matching currently available from Amazon
2. Solid motor coupling: http://www.amazon.com/Shaft-Couplin...69365&sr=1-7&keywords=motor+shaft+coupler+5mm ( 1 set @ $4.80)

US Based Ultibots.com, total cost $44.95 + shipping :

1. Leadscrew w/Nut: http://www.ultibots.com/z-axis-leadscrew-nut-pair/ (1 set @ 44.95 plus shipping)
2. Solid motor coupling: N/A

Personally I went with the first links above. It was an experiment and I wanted it to be as cheap as possible. I rarely buy parts like this from Amazon, unless I need it in two days (Amazon Prime) as most of this stuff ships from China anyway. Almost all US resellers on Amazon double the Chinese pricing. Ultibots.com and other sites only become viable when they have something no one else does, or they run a special (discount).

These parts will essentially do the same thing as the "Official" Leadscrew Upgrade Kit, with a couple of refinements. I built the model to help eliminate backlash on the Z axis. The Robo relies entirely on the weight of the carriage and gravity to keep the backlash to a minimum and with the slop they built in to the X Axis Idler there is little chance of eliminating any backlash. It is another contributing factor to Z ribbing, and why success is variable with the official upgrade.

None of the links I provided are affiliated with me in any way. I do not make any money from this. If you want to support Robo3D then purchase their kit, otherwise I present this in the hopes that everyone who wants to, can have the same benefits at a fraction of the cost.

One more note on why Lead Screws are much better than threaded rods. Threaded rods are designed to be used under tension, not compression as in this application. The threaded coupling is manufactured for ease of threading and as such is not a precise fit. In fact it fits rather loosely on the threads. Consider that for auto-leveling the motor moves in steps that are less than 0.12° per microstep. Even with a minimum of slop that entire microstep and many more can be swallowed up in the backlash built into the system. The auto-leveling may be moving the rods, but essentially not moving the bed. Lead screws are precision components that are designed for linear motion and as such eliminate most if not all of the backlash between the nut/rod interface. This is a very good reason why all pre R1+PLUS owners should consider replacing their Z Axis rods (kickstarter/beta/R1) with lead screws.

One last note: This differs from @Marquis Johnson's solution in two key areas. First his model follows closely the dimensions of the threaded hexagonal coupler originally installed on the Robo. As a result there is no compensation for the backlash introduced by the purposely designed slack in the X Axis Idler. Second and less important, the Z Axis Stabilizers (where the switches mount) are incorporated into the part, Why have two plastic parts bolted together when all you need is one. You can use my model with the Sainsmart Integrated Linear Steppers. I thank him for all the work he did on that video and without it I would not have been inspired to do this.

 

Yea, if you are buying from China just go direct to BangGood.

Good job posting this.

 

All the parts gathered in one place ready to install.



After inserting the lead screw nut and moving the Z axis microswitches into place.



In place with TR8*8 lead screw inserted, verifying that the Z axis microswitches engage properly.


Setup complete, making sure X Axis rails are parallel with the bed before putting the cover back on. All that is left to do is a bit of lithium grease and the lead screws and it's done. One thing to note, in the original post I mentioned 300 mm lead screws. If you use 300 mm the result is build height remains exactly the same as the default Robo. I used 350 mm lead screws, which fit without modification, and the result is I actually gained about 30 mm of build height. As a consequence of using 350 mm the actual cost to me was just under $27.00.

Here are a few pictures, I already test printed an object and all the ribbing is gone, will post images tomorrow, sometime between the Parade and the Panthers game. You can see that the adapters, which were printed before the upgrade, show a patterned Z Axis ribbing on all sides.

 

Thanks @Mike Kelly but need to replace the failed power supply first. Quite the coincidence that the power supply fails right after doing my first test print with eSUN PolyCarbonate filament and only my second print after upgrading the lead screws.

 

Yea, that PolyCarb totally nuked the power supply

 

Of course it did,

Sent from my SAMSUNG-SM-N910A using Tapatalk

 

Here is the result of the Z Resonance print test. Please bear in mind that I normally do not print in PLA and this was using a default profile, there area few little blobs on the print from incorrectly set retraction settings. and I found out when I was done that I had inadvertently set the feedrate to 130%. Still pretty good as far as I am concerned. I am going to consider this a successful mod.

 

Ran another test on Z probing to see if the adapters have any affect on the results. Here is the data from the middle of the bed, the most flexible part of the platform.

Code:

Send: M48 V3
Recv: M48 Z-Probe Repeatability test
Recv: Positioning the probe...
Recv: 1 of 10   z: 0.175000 mean: 0.175000   sigma: 0.000000
Recv: 2 of 10   z: 0.177500 mean: 0.176250   sigma: 0.001250
Recv: 3 of 10   z: 0.162500 mean: 0.171667   sigma: 0.006562
Recv: 4 of 10   z: 0.175000 mean: 0.172500   sigma: 0.005863
Recv: 5 of 10   z: 0.157500 mean: 0.169500   sigma: 0.007969
Recv: 6 of 10   z: 0.152500 mean: 0.166667   sigma: 0.009647
Recv: 7 of 10   z: 0.167500 mean: 0.166786   sigma: 0.008936
Recv: 8 of 10   z: 0.170000 mean: 0.167188   sigma: 0.008426
Recv: 9 of 10   z: 0.165000 mean: 0.166944   sigma: 0.007974
Recv: 10 of 10   z: 0.175000 mean: 0.167750   sigma: 0.007941
Recv: Mean: 0.167750
Recv: Standard Deviation: 0.007941

 

My fault, forgot to put it in a shared album.

 

Better?

The new lead screws in combination with the new Y Axis bearings have completely eliminated any anomalies in the layer placement. Next step is to "tighten" up the carriage using the same solid bearings. After that all that is left will be firmware based adjusting various acceleration settings as well as Jerk settings.

 

There is still just a bit of fluctuation but not like it was by default. I believe what's left is due to X/Y resonance and not Z resonance. Z resonance seems to be more about shifting layers, rather than what we refer to as ribbing.
http://makezine.com/2014/11/07/how-to-evaluate-the-2015-make-3dp-test-probes/

 

Okay here is a better picture of the end results, this is the Make Magazine XY Resonance Test model, with the 1.0 mm walls. There are still some acceleration artifacts that can be seen as very slight variations in in density, especially where the solid side has a groove start about half way up. You have to really be picking at gnat's balls to see it under normal conditions.

 

The switches that go into the printed parts are for the auto-leveling/homing only. If you are doing this upgrade to a Robo without auto-leveling you will need to purchase two microswitches.