My Deltex || A Large-ish Format Delta Printer || Build log
The goal of this post is to keep a more or less practical overview of all the threads I have started that are related to my large-ish format delta build, that I have dubbed my Deltex. I will keep this 1st post updated and edit is as new threads are added.
My first idea was to make this overview just for me in some private spreadsheet as a way to keep track of the threads I made. But then I though well, if this info could be useful in some way shape or form to anyone, I might as well put it up here. So here we are I guess ¯_(ツ)_/¯
The story of this build starts a couple of years ago in the 2016-2017 era with me getting a Rostock Max V3 kit from Seemecnc. This was my first delta and first 3D printer for that matter. I enjoyed building the kit and playing with the printer, but before long I started tinkering with the printer.
March 2017 - FSR probing
I installed a JohnSL FSR kit as an upgrade to the factory installed accelerometer probe. This was for two reasons, one I believed (and still do in most case) that is was an upgrade. And two, I was having concrete plans to upgrade the stock rambo board to our long loved Duet Wifi. While interfacing between the accelerometer and the Duet was done before by some people, it was not plug and play by any means. So I did not think of using the Duet and accelerometer as an option I was going to do. The 3rd picture in this thread shows the newly mounted bed.
March 2017 - Duet Wifi
With the FSR's probing installed, the road was paved to install a Duet Wifi. I need not convince you guys here of why I did that I guess . There are really no pictures of this process, nor is there a thread. I mainly followed Michael Hackney's excellent guide.
May 2017 - E3D v6 and Bondtech BMG
Convinced that I would solve stringing issues, and give me the option to print in TPU in the future, I looked for ways to install a V6 with a Bondtech BMG mounted as a direct drive. I again looked at Michael Hackney and slightly modified his underswung mount documented on his site. I used this setup for quite a while and was fairly happy with it, not really any public picture available though.
April 2018 - Linear Rails
An upgrade talked quite a bit about over on their forums at that time was swapping out the carriages for linear rails like (600mm) mgn12's. I went looking for a solution introduced during the Bondtech upgrade. With the underswung mount, there was quite a length between the nozzle and the ball joints' plane. This introduced effector tilt in the system, especially at higher movement speeds/acceleration/jerk settings. The idea was that the added stiffness compared to the plastic carriages would result in less tilt to be amplified by the long effector. A close up picture of the mount I designed for this available here.
July 2018 - AC heated bed
To combat slow the slow heating time the 12V pcb heater boasted I decided to upgrade to an AC heated bed, controlled by an SSR. A thread about choosing a silicone heater and SSR can be found here. In the end got a 950W heater made, available here. A 2nd thread about the electricals can be found here.
November 2018 - 0.9° Steppers
Continuing the upgrade frenzy I got myself these 0.9° steppers and installed them. The conversion was fairly plug and play.
November 2018 -> Jan 2019 - 24V PSU
Just about simultaneous with the stepper upgrade, I got myself a nice passively cooled 24V PSU. The exact model is a Meanwell HPRG-200-24, that I bought from Amazon. This upgrade not only gave my more headroom for higher speeds without the back-EMF becoming a problem with the new steppers, it also gave me some much need quietness. The stock 12V PSU made a filthy amount of fan noise once it kicked into high gear after a while. Info about the conversion is documented in this thread.
July -> August 2018 - Full Custom Effector and Carriages
Continuing on the effector tilt problem of April 2018, I concluded that the only way to really improve the stiffness of the system was to solve the root problem, the underswung effector. But moving the hot end cold end assembly up was not straightforward at all. I tried and battled to fit the Bondtech+pancake stepper between the arms (space 50mm from each other) but failed to do so. There just was not enough room to avoid the stepper from hitting the arms while printing in some positions.
To solve the space issue, I believe there were two possible ways to go about it. Either increase diameter of the bolt circle the ball joints are located on or use a wider arm spacing to make more room. In practice I think both are useful, so that is in fact what I did. I seem to recall increasing the rod spacing to 75mm, but I do not know how much larger I made the effector. I do know that it was not much more then what was strictly necessary to fit the extruder, hotend and fans. The main drawback of this solution was a loss in printable area, but what can you do.
October 2018 - External 5V PSU
At this time a realized remote monitoring of my printer in action was something I wanted. So I got myself a Raspberry Pi Zero W and an accompanying camera. Powering the Pi is a simple as sourcing 5V, sadly I did not really have an extra wall socket available to plug something in. Quickly I figured I could tap from the 240V that enters the printer and work with that. So I got myself a Meanwell RS-25-5 (25W, 5VDC). With a 5V rail available in my printer the road was also opened to connect the 5V ext input on the Duet and feed it from that. This allows me to switch off the main PSU and still keep the board powered, which is useful for pre-uploading and simulation print jobs.
February 2019 - Controlling PSU state
Along the path of upgrades a external 5V PSU was added and connected to the Duet. This allows me to switch off the main PSU and still keep the board powered. The next logical step was to make the PSU switching available from inside DWC. The functionality was already there as DWC supports this for ATX style supplies and the Meanwell 24V supply has a remote control terminal as well. But do not fool yourself, no plug and play. The logic levels of both are not the same. This is solved by building an inverter based on a 74HC00 nand gate and some resistors. The components are place on a tiny piece of proto-board and seal with hot snot. Simple but very effective and useful.
January 2019 - The Frankenstein is nearing it's end
I have long started to realize that the only important original parts in this build are the white melanin plates and the aluminium extrusions and at this point it has become a Frankenstein machine. Furthermore, there is not really something that is easily upgraded.
A short Youtube video of one of its final prints in April 2019: https://youtu.be/ljWF8_4XOw4
A new bad boy in town
It has been decided! A new delta, designed from the ground up by me, is coming to town.
For a while now I was toying with the idea of designing my own delta from the ground up, but it is at the start of 2019 that I really went to town in CAD and ordering materials left and right. I used Autodesk Inventor for all the designing had have about 95% of the components incorporated in the models. The main concept behind the Deltex is that I do not want to limited in terms of upgradability and possibilities again. If you decide to spend the money, spend it once and do it right, right? Now, I am not swimming in cash, so there is a limit that philosophy
May 2019 - First Light
The basic details of the printer are as follows:
- 1,5m long towers out of 30150 aluminium extrusions
- 1m long MGN12H style linear rails
- A combination of 8 en 5 mm aluminium plates
Quite a few components are salvaged from the old Rostock, may he rest in pieces.
- Duet Wifi of course
- 0.9° steppers
- 24V PSU Meanwell HPRG-200-24
- 5V PSU Meanwell RS-25-5
- Heated bed
- JohnSL FSR's
- E3D v6 + Bondtech BMG + effector
A short video of one the first functional prints I did on the new machine can be seen on Youtube here.
January 2019 -> IP? - Watercooling || Smart Effector || 4th Axis
This part of the build might be the most interesting, at least it is something that sets this build apart from most delta's. The process has been going on for more than a year, spread across different threads so writing a chronological paragraph might not be exactly straight forward. Bear with me though
Back in the Rostock days I moved from an accelerometer based probing sensor to FSR's placed under the bed (check the March 2017 paragraph). The well known issue with FSR under the bed is differences in trigger height based on XY location. This was already an issue with the original smaller bed but would become truly problematic with the full size bed (see July 2019 paragraph). So I went looking for alternatives, as per thread the first, I decided on switching to a Duet Smart Effector.
At this moment a couple of different ideas/issues came to mind more or less simultaneously: mounting the smart effector, watercooling and a 4th axis. After some digging around the best choice for me to integrate a water cooled E3D heatsink was an off the shelf drop in replacement heatsink that I got from a German store. This item is sadly only available with a groovemount, which is not compatible with a Smart Effector. Therefore, an alternative mount was engineered and fabricated. This process is documented in the 2nd and 3rd thread. A tiny thread about adding a virtual heater to monitor the water temperature inside the loop can be found here as well (thread 5).
Next issue up for solving: shifting as much weight as possible from the effector to somewhere else. That somewhere else is a brand new addition to the printer, a 4th axis mounted at the side. The original idea for a 4th axis with just the extruder mounted on it dates back from end May 2019 (5th thread). At that time it was limited to just a concept idea. Even the implementation in RRF was still quite new then.
The 6th thread focuses on how to physically ad the extra axis to your printer and what rules you need to abide to, as RRF makes some assumptions on where you position your axis. The tl;dr here is that the 4th axis endstop needs to be homed at the same time as the other axis. This thread also contains some pictures of various stages in the design process. If you are looking to get more info on how I designed the 4th axis itself and the integration process for the water cooling, the third thread is were you need to be .
January 2020 - Supervolcano on a Smart Effector
I guess the print volume warrants an E3D Supervolcano with a 1mm nozzle, just for those plus size prints. The first thread deals with the compatibility with the Smart Effector regarding electricals. The second thread deal with the mounting on the Smart Effector. Technically only a regular v6 is mentioned there, but mounting a Supervolcano is the same as a v6, so it all applies. The 3rd thread is just some fun stuff about the long simulation times that accompany the large prints.
July 2019 - Bed upgrade
When switching from the old Rostock to this printer, I initially reused the bed. As can be seen in the video in one of the first posts about the Deltex, the bed was quite small compared to the available space on the machine frame. To accommodate the new larger size bed I upgraded to a new 2kW silicon heater, which might be considered overkill but actually gives me similar heating times to the hotend. To switch the higher powered heater, I got myself a replacement SSR for the el cheapo one I had previously. This is talked about in the first thread.
To upgrade to the full size 400mm diameter bed, I decided a something like a heatspreader was a necessity (even on the smaller bed heat gradients were an issue). Around the forums, it seems to be agreed on that cast aluminium tooling plate is the material of choice for this application, so that is what I got. Due to the increased heat capacity thanks to the heatspreader, I figured waiting for the bed to cool each time was not an option and a removable bed was needed. The details of a discussion about this can be found in the second thread. The tl;dr is that I went with a 2nd, thinner alu place that a PEI sheet is attached to and that did not work well. The problem with this setup is that the 2nd aluminium plate I got was quite soft and did not bend back after flexing to remove parts. The 2nd issue was that there was not easy way to keep it in place.
The solution was a removable bed based on magnetism. The choice of steel is discussed in the third thread, but in the end was largely decided by the fact that I could not source spring steel large enough, so I got some other steel (I fail to recall what grade I ended up with at this time). The fourth thread deal with the magnets not being strong enough the really assure that the bed stays in place. Some small locating pin where installed at the side of the bed and the work just fine.
~~ Continued in the next post ~~
September 2019 - Part Cooling Fan
After struggling with some cheap fans that could not build up the pressure I needed for my printer. Long story short, I stumbled on to Mark Rehorst's blog talking about the mother of all print cooling fans. The details about that can be found in his blog. As the design was already available I just needed to hook it up to the Duet, which turned out not so trivial. The BLDC driver board has no proper documentation and apparently requires a positive side switched control signal, while the Duet outputs with low side switching. A hardware fix was hacked together in the linked thread.
January 2020 - Epoxy Granite Machine Base
The full reasoning behind the epoxy granite base can be found in the linked thread. The tl;dr is that epoxy granite is excellent at absorbing vibrations. This combined with its relatively high density, makes for good machine frame material. The goal was to reduce the vibrations in the machine, and especially prevent them from being coupled to the floor and then being audible all around the house.
May 2020 - Final Upgrade: Hotswapable Super Volcano
I have started acquiring parts for the Deltex version 2, more about that in the next post. I guess this means this V1 is coming to its end and it is time for one final upgrade.
Most of the time I print with a water cooled E3D v6 with 0.4 mm nozzle. Though, since the print volume is quite large, some times some more brute force is needed. As detailed in the January 2020 post I got myself an E3D Super Volcano and it eats plastic for breakfast, supper and dinner. Since I got myself into water cooling however, swapping the hotends around was not really practical.
Long story short, I got myself a second smart effector, water cooled heatsink and all else needed. Another vital component is quick disconnects for the water lines. I got myself two sets of them from Alphacool and they work a treat so far. There are some drops that spill when disconnected but that is expected and not an issue at all.
In a short video I posted on Youtube you can see the printer in action. This video also serves as a timestamp for the final look of the V1.
March 2020 - A Deltex V2 Arises
While the Deltex V1 served me very well, there were a couple of things that were kind of bothering me.
The first and most important of which, optimal use of rail length. Even at Z0 and at the bed edge, the linear rail sliders were still about 10-15cm from reaching the end of the rail, which is wasted height of course. Also, my existing rails were sourced oversees, while far for from the cheapest and decent quality, there were some spots that the block ran less smooth on. Combined with the fact that the rails are just a tad over 12mm wide and the tslot channels in my extrusions are just a tad under 12mm wide, alignment was always a bit wonky.
A second issue is belt tension. The bearings in the motors I had were not capable of supporting the tangential loads on the stepper shaft caused by the belt tensions I wanted.
On a similar topic, the last big thing I wanted to change was the motors. Similar to the linear rails, they were sourced oversees and were working fine, but quality and specs were a questionmark.
April/May 2020 - Proper delta endstops
As a part of the big V2 update, I decided I wanted to upgrade the endstops to something better then a run of the mill microswitch. In short, I ended up getting a set of Metrol STE060PB's. I have not performed any measurements but they feel and look absolute quality.
April-July 2020 - Alternative Belt And Motor Mounting
The bearings in the nema 17 motors I had were not capable of supporting the tangential loads on the stepper shaft caused by the belt tensions I wanted. After some nice discussions I ended up designing a mount that holds 2nd shaft. That shaft is supported by two needle bearing belts on either side of the belt pulley and connects to the stepper with a 8 to 8mm rigid shaft coupling. This way the tangential loads are supported by the two relatively large bearings.
April-July 2020 - Stepper Upgrades Once Again
The necessity of going to nema 23 steppers has been questioned and discussed about quite a bit. I don't think there is a universally agreed answer, yet I opted to go the nema 23 route. After comparing quite a few high power models (within the Duet's specs ofc), I ended up getting a set of real nice PKP266MD28B's. They look and feel nice quality, offer impressive torque and are quiet by themselves.
April-July 2020 - Main rail upgrade
Even at Z0 and at the bed edge, the linear rail sliders were still about 10-15cm from reaching the end of the rail, which is wasted height of course. Also, my existing rails were sourced oversees, while far for from the cheapest and decent quality, there were some spots that the block ran less smooth on. Combined with the fact that the rails are just a tad over 12mm wide and the tslot channels in my extrusions are just a tad under 12mm wide, alignment was always a bit wonky.
I decided to get proper brand name rails that are at least wide enough to cover the 12mm of the frame gap comfortably. The set of THK-SRS-12WM's I got are 40mm wide and 1030mm long. For sure overkill in terms of load capabilities, but boy are they nice. They also feature caged balls inside the sliders, making them very quiet and easy to install. The balls don't go all spring sproing when you try to slide a block on the edge of a rail after cleaning for example, which is surprisingly nice. The specific set I got are coated black (no idea if it is anodized or what process they used), thus preventing surface rust.
###July - December 2020 - The noise
With the upgrade to the Oriental' nema 23 steppersa new problem arrose. Resonances in the frame caused absolute terrible noises that could be heard at hindering levels around the whole house. Over the course of several thread and over 200 posts a have experimented with several successfull and less successfull solutions to combat these vibrations. In this section I will try to summerize these post into something more digestible
This whole process started with the vertical extrusions resonating. They were vibrating enough the clearly feel it when I put my hand on them. Over in the first thread, a user suggested bolting something heavy to the towers and quickly concrete became the product of choice. I poured some concrete (with blue die ) slabs and bolted them to the towers. The added mass (~30kg each) did what it was supposed too and silenced the towers.
Sadly the noise issue was not resolved at all, just reduced from loud enough that the neighbours could complain to loud enough that people in the same house complain. At this point I looked into building an enclosure to reduce the noise levels outside of the machine. Here I considered to big options, one was to build a box and put the printer inside of it and the other was to integrate the enclosure and the printer into one enclosed system. I opted for the latter appoach and build 3 panels out of MDF, lined with sound dampening material, that would bolt to the towers. Altough they look quite nice, provide mouting options for a variety of things including watercooling and 4th axis rails, they do not serve their original purpose of sound isolation. I think the big mistake here is that the panels effectively provide more surfaces that can vibrate with the machine. And while the sound dampening material no doubt does its job, the vibrations couple straight into the wood and practicaly bypass all dampening.
Another thing I investigated is the motor mount. One issue that I found out was that the back of the motor not just vibrated in air but vibrated against the plate it is mounted on, causing even more noise. After ajusting the motor mount to give some clearance between the stepper and the frame that issue was resolved and quite noticeable.
A second change to the motor mount is the addition of flexible couplings on the axle. I did not really notice a change in noise levels but I needed something that allows for some minor misalignment between the stepper's axle and the axle that pullies sit on.
I also added the well know Astrosyn style face dampers between the motors and their brackets. This did improve the noise noticably.
Finally, I completed the bracket design to my desires and had them machined outof stainless (stainless is overkill but due to a deal it was the cheapest metal available at the time). They look absolutely awesome and clean in addition to a minor reduction is noise.
This is where this section on noise will be closed for a while. The noise levels are quite acceptable but not perfect. A number of options are still on the drawing board, such as different steppers and a closet style enclosure.
Update: completed sections
July 2019 - Bed upgrade
January 2020 - Epoxy Granite Machine Base
Update: finally got this post more or less updated to current time, enjoy the read
I am also pleased to announce a Deltex V2 which is just about confirmed to be coming. It will not be a 3rd printer as it will reuse the majority of the existing frame and other components. Some changes that are currently planned:
- Linear rail upgrade from Robotdigg MGN12's to original THK SRS series rails.
- Optimization of rail placement should increase build height ~10-20 cm.
- A sound and temperature insulation enclosure.
- The existing fourth axis with v-slot rail with be swapped for the old MGN12 rails. Furthermore, the 4th axis mounting points will be incorporated into the enclosure.
- Water cooling components might be moved to be stationary on the outside of the enclosure walls.
- New way of mounting the steppers/belt idlers to relieve stress from the stepper bearing at higher belt tensions.
- Existing printed corner brackets in the frame will be replaced by metal brackets.
- Possibly the introduction of even more metal parts as replacement for current plastics.
Update: Finished writing the final piece of the Deltex V1 build log. Currently designs are being drawn and parts are on order for the V2, exiting time for sure.
Updated again: added the section about the noise issues I have been having and some picture