PB+ Assembly Tips

From Printrbottalk Wiki
Jump to: navigation, search

Here are a few tips on putting the printrbot+ together. They should be, more or less, in the order of assembly. They represent my own experience with the kit of parts I received in early June. Due to the inconsistency of the parts, variation in the materials and processes used to make the parts and poor quality control during the kitting process (what you get may be quite different from what I got!) Your Mileage May Vary. I have tried to express this as representative of my own experiences during the process of getting my printrbot+ to work the way I wanted it to. I have now added attribution to some of the edits done by CL1 where they do not agree with my own experience. Use your own judgement while considering this.

Use of Washers

Brook is pretty nonchalant about putting washers under the screw heads. However, the use of washers will spread the load of the clamped screw over a much larger area on the wood or ABS and should reduce the deformation, allow tighter clamping and possibly diminish some of the eventual need to re-tighten due to creep of the material under long term load. Another way to minimize nut-creep is to use Locktite or elastic stop-nuts. I prefer the stop-nuts (where they will fit) because they are a bit easier to deal with for maintenance.

CL1 info: Washers are only needed where they are used to bridge a rectangular cutout. Using them at other joint locations is not recommended! The addition of a washer will actually increase the likelihood of loosening under vibration, and are contrary to the LC and Plus designers intent. YMMV. Another way to minimize nut-creep is to use Locktite or elastic stop-nuts, commonly called Nyloc. There are many places where these nuts will not fit. Locktite Blue or Purple Threadlock works fine for those and is available at most Hardware and DIY Box stores.


Be sure to tighten the cable tie for the z-endstop from the outside. Placing it on the inside may be neater but it will interfere with the z motor.

CL1 info: This switch is critical to the easy operation and repeatability of your PrintrBot. The wire-tye may not provide enough security from unintended movement of the switch, which means you will end up adjusting the switch actuating screw or Z Leadscrews before many or all prints. In the Assembly Notes for LC and Plus, you will find some ideas for how to improve the mounting. Using screws (not supplied with the Kit) or at least some glue can be a worthwhile step. I suggest doing it now to avoid unnecessary problems later. The PrintrBot will work fine with just the Wire-tye, but it is far more likely that the switch will slip over time and be variable with use.


Relocate z-axis Limit Switch

Adjusting the z-axis limit switch is a real PITA! I was having to do it way too often so I did decide to move the whole process to the left front side of the base. The switch actuator is now aligned with the left leadscrew (along the x-axis at least) which reduces one adjustment interaction. The adjustment screw is ground flat or slightly spherical on the end so that rotation has reduced effect on the contact point and then locked with a jam nut to secure the adjustment. Also, I attached the microswitch with screws in place of the cable tie to improve the stability for this critical sensor. The holes in the microswitch are small, requiring a 2-56 or 2.25 mm screw to fit.


Linear Bearing Alignment

It will help maintain the bearing alignment if the rails (smooth rods) are inserted before the cable ties (zip-ties, TyRaps) are being tightened.

CL1 info: Do NOT cut the ends of the cable ties flush as you will need to re-tighten them after a few hours of use! They may be cut shorter, but leave enough tail to grab with pliers for this re-tightening, which is like the first check-up of a new automobile.


y-axis drive Assembly

Note that the pulley mounts to the y motor shaft with the clamping screw away from the motor, unlike the extruder gear and the x motor shaft.


Weird Noise in the y-axis

I was hearing a weird noise when the y-axis moved rapidly. I was finally able to track it down to the belt rubbing on the bearing retainer. Scooting the belt ends as low as possible in the platform clips was not enough to eliminate the rubbing so I added an additional set of 5/16" washers below the bearing to move the whole bearing assembly up a bit. Using additional washers here requires care in assembly as the printed retainer does not extend far enough to center the washers. Also the pulley clamp M3 head must be above the bearings to avoid interference and you may need longer M3s through the bearings into the y-axis motor.

CL1 info: In the picture below it looks like the Y Motor shaft was rubbing on the PrintBed. This is the first report I've seen of the problem described in the previous paragraph. The plastic GUID batwing shape piece is made for single washers under each leg and therefore added washers will not have central support. If you have the problem as described, consider adding spacers between the CLIPs and PrintBed instead!

The above mentioned rubbing seems to have come from the warping of the printbed discussed in more detail below.


Upgrade x and y axis drive to Gates GT2 belts and pulleys

I was seeing slight overshoot in sharp corners on my prints and have had trouble getting the y-axis belt to run straight and not rub against things.

With this in mind (and also in the interest of obsessive tweaking) I decided to change the drive components from XL style belts and printed pulleys to GT2 components. There is a detailed writeup on the RepRap Wiki here: [1] This covers the alternatives and explains the benefits of using a system designed for the application.

The printrbot can benefit from something that was designed to provide precise positioning and not just transmit power. Think about the difference between powering your alternator or power steering pump and driving a camshaft in an internal combustion engine.


The x-axis upgrade was trivial. The GT2 pulleys come with 5 mm brass bushings for the shaft and setscrews (uses a 0.050" hex key) and the smallest, readily available, size is 36 teeth. This is great for the x-axis because it is almost the same diameter as the 608ZZ idler bearing on the other end of the belt. This removes the small bit of cosine error that comes from the belt angle with respect to the axis of motion changing as the carriage moves. Change the pulley, string the belt and tighten the cable ties and you are done.


The y-axis presents some additional challenge. As the printrbot is designed there would be zero clearance between the 36 tooth GT2 pulley and the 608ZZ bearings that are used for belt guides mounted on the y-motor 3 mm mounting bolts.

I modified the belt guide support to offset the bearing supports by moving the 3 mm mounting holes off center by 1.5 mm. This doesn't leave much wall left but since the support is captured inside the bearing ID it seems to be more than adequate.


Additionally the bearings need to move up from the surface of the printrbot base by an additional 5/16" washer's thickness in order to clear the pulley flange and have the pulley clear the base. I ground off one side of the washers so they would have plenty of clearance with the pulley flange. You could also remove some of the flange but that seems less desirable.

I moved the holes and lengthened the bearing shafts of the bearing support so that the additional washers would also be centered by the support.


The pulley then needs to be placed on the y-axis motor shaft so that the bearings fit between the pulley flanges and the bottom flange clears the base surface.


An additional benefit of this upgrade is that the more precise pulleys assure that the belts run true and are less likely to rub on the bearing support or base and the pulleys won't work loose on the shafts or delaminate like the printed pulleys seem to do repeatedly. And they sound better too!

I bought the parts from Stock Drive Products, their part numbers are A 6Z51M036DF0605 for the pulleys and A 6R51M762060 for a 1524 mm belt, enough for both axis'. TechPaladin is a good alternative as their minimum shipping charge may be less. The RepRap Wiki shown above has links and more possible vendors. These are 6 mm wide belts with a 2 mm tooth pitch.

Don't forget to recalibrate your x and y axis as the drive ratio has changed, I am using 44.5 steps/mm now.

Heatbed Temperature Measurement and Insulation

To get accurate temperature readings for the heatbed it is helpful to have good thermal conductivity between the heatbed and the thermistor, and low conductivity to the structure beneath the heatbed. To achieve this I choose to place an insulator below the heatbed. I used two pieces of cork sheet from the hardware store. The top layer is 1/16" and has an area cut out to clear the thermistor and its leads. A piece of 1/8" cork forms the lower layer. I had previously tried 1/2" cork and cardboard (with foil and Kapton tape) but they were either too thick, too messy or too uneven in thickness to fit the platform without gaps. Small strips of paper can be used a shims between the cork and the platform to refine the bed leveling. Another benefit of this cork material is that it is flat enough to eliminate the need for springs to support the heatbed.


I placed a small piece of Kapton just between the heatbed and the exposed wires of the thermistor (NOT UNDER THE GLASS BEAD OF THE THERMISTOR) and then placed a drop of epoxy between the thermistor glass bead and the heatbed's lower surface. Epoxy is not a great conductor of heat but it is WAY better than air! Common CPU Heatsink epoxy or grease are alternative choices. The thermistor is then covered with another layer of Kapton for good measure.



The insulation should slightly improve the response of the heatbed to temperature change commands but wood is also a pretty good insulator. The real benefit comes from isolating the thermistor from things other than the heatbed so that it will more accurately measure the heatbed temperature.

Mounting the heatbed may require longer M3 bolts, depending on the thickness of your insulating layer and any clips you might use to retain a piece of glass. This also will slightly reduce the z-axis build envelope.

I found it very difficult to get the headbed to lie flat, even without the insulator. Part of the problem is that the screw holes in the platform did not match up with the heatbed screw holes. This applies a bending moment to the heatbed making it unflat! Enlarging the holes and match drilling them to assure they are aligned with the holes in the heatbed should reduce this tendency.

CL1 info: Many owners simply omit the screws entirely and use small binder clips to hold the Glass and HeatBed to the PrintBed.

This, however, requires some care in placement to avoid interference with the y-axis linear bearings and their cable ties and may result in some additional loss of build area.


One clue to this problem is seen in this side view where the screw is not perpendicular to the platform.


I ended up using a piece of Pyrex (borosilicate glass) from McMaster-Carr (8476K181) which I would recommend. The only downside I can see is a slower warm-up time, the glass is very flat. With the heatbed set to 100 degrees C and no fan running this is the temperature profile I am seeing. Maybe it isn't any better than with a simpler solution but I am getting good results. Due to the DC drive, there is a temperature gradient across the heating element. Change the DC polarity to the HeatBed and the gradient will reverse.


PrintBed Flatness and Stability

I found lots of instability in the setting of the z-axis limit switch. There seemed to be some hysteresis/sticktion etc. and I was constantly readjusting the setting. Also, the heatbed was very difficult to level. It became increasingly apparent that the printbed was not flat or stable, a losing proposition for any kind of system stability. I decided to stiffen it and straighten it with some bracing on the underside.


I attached two pieces of aluminum channel in the areas that did not interfere with the rails or y-axis drive belt. The channel came from the hardware store and may be marked as 3/4" plywood edge trim. Flathead screws and elastic stop nuts retain the channels and they have the additional function of allowing shims to be added between the channel and the plywood to fine tune the straightness.

Some warping remains in the x direction but this is of less or no consequence due to the way the heatbed is supported. At the same time I moved the heatbed slightly rearward and match drilled some new holes so it would not be under any additional strain from the support screws.


Extruder Assembly Caution

When assembling the extruder be careful not to drill out the screw holes for the pinch roller pivot and the hotend retaining screws all the way through. Just drill out the lead-in side so that the screw threads will engage the mount.


Extruder Gear Alignment

To achieve smooth extruder operation it is important for the extruder herringbone gears to be coplanar. In other words, they need to line up so that they will mesh properly.


In my case this required that the extruder stepper motor be spaced back from the mount by the thickness of a #6 washer. In order to get adequate engagement from the motor mounting screws I needed to use slightly longer M3s.


Also, the motor gear had to be moved as close to the motor body as possible. This required flatting the shaft to within a few mm of the motor body.

Shortening the extruder motor gear clamping screw by about a mm provided the needed clearance with the extruder mount. For reference, the supplied motor threaded mounting holes are approximately 5 mm deep.


x-endstop Adjustment

It will be easier to adjust the x-endstop from the open side. Since the screw may miss the endstop switch actuator arm you can put a double nut on the end of the adjustment screw and then add locking nut to keep the screw secure.


z-axis Bearings and Binding

Getting the z-axis to move smoothly was a major challenge for me. I think I had multiple problems, all conspiring to cause the movement to be halting at best. In summary I had:

1 Crappy (low quality) bearings
2 Overly tightened x-carriage retainer plate screws
3 The wrong size bearings for my laser cut bridge
4 Poorly lubricated bearings

The first clue is illustrated by the deformation of the laser cut bridge wood shown below.


Close inspection showed that the z-axis rails were binding on the wood as indicated below.


Also, over-tightening the screws will slightly deform the bridge and aggravate the lack of clearance. Then I found that the 12mm linear bearings I received (21mm OD) sit too deeply in the laser cut recesses in the bridge. I used some 0.01" brass shim stock (hardware store or hobby shop) to move the bearings out in their slots enough to continue.


I would recommend placing the cable ties with the clamp piece oriented toward the center of the bridge. There is more space there so that the cable tie clamps don't interfere with the x-carriage retainer plates when they are tightened. This photo shows them incorrectly mounted. I went through a lot of cable ties working on this!


Finally I found that slathering the bearings with wheel bearing grease was helpful in relieving the rough motion. I think now that I will order some 22mm OD higher quality bearings.

The carriage should return to the z-axis home position smoothly even without the the extruder mounted. I found this to be an excellent test case.

New bearings to the rescue

I am reasonably sure that I was sent the incorrect sized bearings for my z-axis. They measure 21mm OD and 30mm long. I inquired of PBHQ via email about this but have not received any response. I just went ahead and ordered some 12mm X 22mm X 32mm ball bushings from VXB and they were not only the correct size but they are MUCH smoother as well. I suspect that many of the complaints about "crappy bearings" may be incorrectly sized bearings or mismatched parts that span design revisions. Still, I needed one set of 0.01" shims on my bridge in order to be able to tighten the x-carriage retainer plates screws snugly without causing the rails to bind on the bridge. The z-axis is now smooth, does not bind and moves down under gravity power without the weight of the extruder or x-axis motor present.

z-axis Drive Upgrade to Acme Threaded Rods

I decided to forestall any degradation of the zinc plated all-thread leadscrews by substituting acme threaded rods. I bought 1/4-16 X 12" lengths from McMaster-Carr (93410A904) and mating nuts (94815A007) along with some cheap imported aluminum couplers from Ebay (search for "6.35mm x 5mm Shaft Coupling"). Beware of poor concentricity in these couplers, runout can lead to layer thickness variation with a period of one revolution of the z-axis, this looks like ripples in the side walls of a print.


I know that this is steel on steel but I have not yet found any 1/4-16 brass or bronze acme nuts (that were reasonably priced). Good lubrication should handle the light loading. I made two carriers that are printed around the nuts and fit into the bridge assembly. These were necessary because the nuts were too large to fit directly.


Installation took about 30 seconds! Don't forget to correct your z-axis calibration by a factor of 16/18 for the change in thread pitch.

Preliminary Axis Calibration

I found it easy to set the preliminary calibration by measuring each axis' travel directly. Enter the calibration factors in the Gcode direct entry field (lower right corner of Pronterface). The command is M92, in my case M92 Z2023.3308 for the z-axis (I have 16 tpi leadscrews, see above). Then set up a scale on a reference point and move the carriage 50 or 100 mm using the Pronterface direct motion entry buttons. Measure the actual travel, adjust the calibration factor and retest. Repeat until you are satisfied. This is testing the actual travel and provides no compensation for extrusion width or other variables.


Other Upgrades For PB+

See Printrbot+ Upgrades Parts List for list of upgrades and replacements for various parts, suggested by the community. It's not a given that any of these will improve your printer's performance, but some people have tried them and found them useful for specific issues.