X5ST Build Log

Created by: Lester Caine, Last modification: 08 Jan 2019 (15:15 UTC)

The original plan was to video blog the building of the printer, but my first attempts at recording were less than successful with poor sound quality. Something of a disappointment since I've been providing high quality video links for hospital training sessions for over 20 years. The main problem is that one needs a second body to do the camera and sound work while presenting, the job I normally undertake on video links, so for now I'll just video demonstrations where appropriate. If you want to have a laugh at my initial videos, they are archived in the image gallery (Video Clips). All the pictures are archived in the main gallery and will be used in the main build log.

Perhaps surprisingly, there seems to be no manufacturer generated build document. TronXY's site seems to be devoid of any useful documentation, and only contains sales blurb. GearBest has produced documents for some of the printers they supply and so have one for the X5S. In addition, other builders have produced both build logs and video logs covering the building of the machine. I've watched a number of the video builds and from those gathered a list of things that needed attention while building in order to improve the final machine's performance. In addition, the X5ST already addresses a few of the earlier limitations but perhaps not with the best alternatives. Trying to work out just where these changes fit in the original build guide ass to the fun as well, which also added to the problem of trying to do a video blog. I have already had the 'Step 1' section apart a couple of times in order to tidy things up a little more. customize-3d.com has proven a useful resource and has an alternate build log which I've been using as a cross reference to the GearBest one.


The starting point as always is opening the boxes. The right hand box is the base printer direct from TronXY while the left hand box has all the extra bits I'd picked up before Christmas. More on those as he build progresses. Unpacking the printer box provided all of the expected parts and also a little head scratching as there seems to be numerous extra small parts that do not seem to fit anywhere. I'd already taken the GearBest guide and assembled a spreadsheet parts list against which I could check off all the parts. The result was less than successful as some parts are now pre-assembled yet the small parts kit seems to include some screws and spacers intended for the original X5S. In addition I an planning to the cap screw stock to replace the larger cross-head screws, something which has already happened in part in the pre-assembled elements.


Trying to take pictures and write notes are all taking a lot longer than actually doing the job in hand, building the printer, but I think that the effort is worth it if only to explain some of the decisions made while building. The current 2 days work has only reached this stage of construction ...


One of the various changes planned while building was the inclusion of corner brackets at each joint. The selected joint is actually less than successful as it is simply not stiff enough to maintain a square corner. It does however work well when all 4 corners are fitted. It would have been better to assemble the lower rectangle of 2020 extrusions first before fitting the upright 2040 extrusions. Ensuring the corners are all square before finally tightening up.


The designed mounting for the rubber feet was simply a bolt washer and t-nut, so my alternative may seem a little overkill, but given the small footprint of clamping between an M4 washer and the V grove of the extrusion, it's all too easy to knock a foot off. By clamping between the M6 washer and the penny washer this is less lightly to happen. I am already considering a bottom panel which will remove the need for the penny washers. It will ensure the base is perfectly square and allow space to relocate the power supply inside the enclosure allowing a back panel to be fitted.



Stage 2 involves assembling the Y Axis carriages on the top rails. The plan had been to replace the rather Mickey Mouse bearing and penny washer idler wheel arrangement with proper timing belt idler wheels and correctly position them to ensure the belts are running in parallel or right angles on all paths. However the new module has a modified idler with a plastic pulley and a pair of bearings. Similar to the bearing arrangement on the tracking wheels. I'm not happy with either as some wheels and pulleys are not running smoothly and the idler pulley is too wide so the belt can move up and down in the gap. Replacing the pulleys with the smaller timing belt idler is proving a little tricky as the inner hub is a little thin and the M5 washers cover the whole bearing, and I'm waiting for some new bearings for the wheels although simply swapping around the good ones from the pulleys should work short term. At some point I know I need to modify the whole carriage anyway as there is no way to adjust the clamping of the wheels on either side of the rail, but that needs the printer working to manufacture the improved elements.


The Mk1 version had both carriage idler's on the one bolt. I'm not sure that this is much better as it requires that the X axis carriage has the belts attached quite far apart. In addition the y axis alignment is still poor, and only requires the left hand idler pair being moved a little further over to keep the belt parallel to the axis. The vertical spacing is also out of alignment, but given the over wide pulley height compensates a little for that.


I'm looking to maintain a more accurate tracking of the belt horizontally and vertically, but even the toothed idlers are a little wider than the belt. The vertical spacing is in need of a slightly longer M5 bolt to allow an extra 2mm offset o perfectly align the idlers, and a small change to the left hand pulley which I will probably replace with an aluminium bracket. This needs to mirror the alignment of the stepper motor on the other side.


The supplied idler will take a 10mm belt while even the toothed idler is capable of handling an 8mm belt. Only using the setup will show if this is still a problem. The original design simply used bare bearings and penny washers which while functional did not allow proper tensioning of the belts, and are on the list of recommended changes


An unforeseen problem with the toothed idler is simply mounting it! I had gone with the M5 hub, since that matched the size of the other elements. The difficulty is that the boss of the hub is just too narrow to clamp against. If I use normal washers it clamps both sides of the bearing rendering it somewhat pointless? I'm not sure just what washer on can use in order to assemble the idler and still have it run freely? Another interesting aspect is the simple fact that one needs to be able to control the tension on each bearing and having two bearings on the one bolt is a problem in itself. The Mk2 option simply used an M6 bolt as a spacer, but replacing it with an M5 nut and adding one to the second idler solves the problem here. The pair of idlers is more of a problem and he one where the use of the right spacer is more critical, but currently it is working reasonably well using the internally serrated washers.


Stage 3 While the Mk2 print head looks like a nice and complete module ready to simply fit to the X Axis rail and bolt on, however there are a number of not so obvious problems hidden inside. The cooling for the working surface is some considerable distance away, and not even directed at the nozzle. Providing a flow from both sides of the nozzle seems to be essential for the best quality printing? In addition a key element on all designs of heating body is a separation of the hot zone from the feed tube. This is absent in the TronXY design and so I have a couple of alternate feed bodies with a well defined gap between the heater and the feed area, along with a 24V 50W heater to help with the PSU upgrade. I'll install the head as it is currently, but am unlikely to start printing with that setup.


Stage 5 A fairly successful evening outside. I'm definitely at a stage where it would be nice to run up the X and Y motors and see the print head move around. It may not be immediately obvious but rather than the motors moving one or other axis, they work in tandem. When both moving the same way the Y axis moves, while moving opposite directions drives the X axis. This is the reason that it is important that the belts are moving cleanly either parallel or at 90deg to the drive paths. A little cock-up is I supposed to be expected and while 8 of the 10 toothed pulleys have the belts the right way around, the remaining two have the wrong side of the belt to the pulley. I think I can cheat and simply turn two down, which will fine tune the stepper motor to x axis angle, but the stacked idlers ideally need to be split to move the final path away from colliding with the belt going in the opposite direction. One of the mounting plates has developed a crack as well, I should have added washers top side, but an aluminium replacement seems the best option anyway. More concerning however is the very tight movement of both axis. I need to loosen off the drive wheels a little which has been mentioned in other construction blogs and is part of the reason for switching to an adjustable 3 wheel carriage on the Y axis. I think I expect the carriages to move with just a single finger pushing against them?


Stage 6-7 Assembly of the two sides of the Z axis lift is fairly easy, except the build instructions include a washer between the top of the linear rods and the frame, but it has not been possible to fit this. Additionally, the linear bearings have been supplied with soft washers glued to the upper surface of the mount, but this seems redundant as the lower surface is clamped to the cross rail. The backlash nut and spring is similarly fitted below the rail which results in a little more pressure on the spring, however the rail still moves with gentle pressure once in position.


Assembly into the cube is relatively straight forward since the linear rails are simply clamped with a bolt through the top and bottom rail. One just has to ensure that the bottom rail is central in the hole in the motor bracket before tightening up. Running the cross rail up to the top rail before tightening the bearing against the rail ensures the leadscrew is aligned top to bottom. Holes in the cross rail make it easy to tighten the bolts as they are obscured by the rail. As had been anticipated, the cross rail is perhaps not as rigid as would be liked. It does rock a couple of millimetres from end to end. As long as the build is fairly even across the bed, this should not be a problem, but addition of a third Z axis assembly with stepper motor at the back of the cube has been implemented by a few owners to keep the whole bed flat.


Stage 8 Assembling the extruder is a little academic as it was supplied ready assembled. Except that it needs disassembling in order to remove the paper layers off the plastic bracket. At some point a second extruder will also be added to run a second material feed.


Stage 9 After a bit of digging around I've established that while the X5S was supplied with a dual voltage heated bed, the 'Mk3a' is a 12V only design despite the versions actually supplied by the reprap.org project all being dual supply. I don't think the TronXY design has any connection with the non-existent web page it refers to, and annoyingly it would seem to be if little use unless I use both power supplies. Alternatively I simply remove the thermistor and glue the silicon pad to the back instead of getting a second panel. I'm looking into the options at the moment, but while the CRAMPS board has individual supply connections for Steppers, Heaters, Bed and Fans, the TronXY board seems to just have a single supply, but it is marked as 12/24V so I should be OK to run 24V and then use a 24V feed from the bed output to either the 12V heater, or the SSR for the mains pad. I've actually got a nice 30Amp '13.8V' power pack which would allow a bit more power to the 12V option in the short term. I've held off ordering the second ali panel so I can work out what other panels may be useful. I think I will add a solid base in any case, but the rest will probably be plastic.

The glass plate is a "31cm sq Ultrabase Heated Bed Build Surface Glass Plate" which I picked up in the bits order from China. Even with shipping half the price of a UK sourced version and has a good press in the youtube demo's. Just needs a retainer to drop it into on the bed.


Stage 10 Now to see if we can get things moving! Is it worth puling out the ELS? I'd simply run up each motor on the DivisionMaster, but having built a batch before Christmas, they all went and while I've got the DMSuper I can use, getting at the dip switches is a bit of a faff and 4.2Amp/45V is bit too much for a NEMA17 motor.

Stage 10a I will need to cut a couple of aluminium brackets to carry these pulleys. The outside ones are simply tapped M5 into the plastic as they are prevented from going through by the framework below. What has also helped here is switching from the belt provided in the kit to a reinforced rubber one recommended by others. This sits a lot squarer than the plastic one which even when tensioned a little still laid over at an angle. This setup would probably not work with the larger idler pulleys supplied in the kit, but I think the extra expense has resulted in a much improved layout? You will see that the homing switches are in place and the machine homes nicely ... one axis at a time. I also have the default configuration file and after a little work with google I can even even read it. The way that these 3D controllers manage their configuration via gcode seems logical. No more XML files like I have with Mach3. I am still limited to just the functions programmed into the firmware, so can't simply enable the second extruder. I will need a controller I can reconfigure if I'm going to do that.

Back to alignment, the X axis as almost square with the bed, but the crude way the belts are adjusted means that matching the tension perfectly in order to achieve a perfectly square X axis is difficult. Another improvement in the belt paths will be the inclusion of a more controllable tensioning adjuster, such as a screw jack on each leg.