3D printer teardown 'take two' - The UP Mini 3D printer by PP3DP.

[Fraser]

As some may already be aware, I previously provided a look under the covers of the PP3DP UP! Mini 3D printer. If you have not seen that thread, take a look here:

https://www.eevblog.com/forum/reviews/up!-mini-3d-printer-by-pp3dp-a-look-under-her-bonnet-comments-on-design/msg677679/#msg677679

This time I shall be removing the covers from the printer and delving deeper into its construction.

This more in depth look will hopefully assist anyone else who needs to dismantle their UP! Mini for cleaning or repair.

Enjoy 

[Fraser]

More Pictures

[Fraser]

More Pictures

[Fraser]

More Pictures

[Fraser]

More Pictures

[Fraser]

More Pictures

[Fraser]

More Pictures

[Fraser]

More Pictures

[Fraser]

More Pictures

[Fraser]

More Pictures

[Fraser]

As a little commentary for the above pictures........

This is a second UP! Mini printer that I recently acquired for £107 as "parts or not working"

The seller knew nothing about 3D printers and could not test it or provide much detail on its condition.
From what I could establish it was a customer return but it sounded like the extruder was as new and unused.

The unit arrived yesterday. It is not new but is in very nice condition. It was originally supplied by Amazon and was returned to them for reasons unknown. It then went into an auction of some sort as it has an Auction company label and lot number on it.

From papers that I found in the bottom of the box, under the packaging, it is obvious to me that this was used by a student for a project. I may be being very unfair but I have a suspicion that this printer was bought to complete a students project and then returned on some pretext. Somewhat like people buying suits and digital cameras for weddings and then returning them afterwards.

The good news for me is that I got a very nice condition UP! Mini for a bargain price.

I decided to strip it down for a clean and inspection. I also want to reverse engineer parts of the controller board to better understand the design and potentially change the build plate heater temperature control design. I thought other forum members might also like to see some more in depth views of the units interior.

Having done some investigation I know that this unit was built after December 2014, is fitted with a Version 3 Extruder, and was supplied by Amazon on 17 June 2015. It was returned to Amazon and then auctioned by SimonCharles Auctioniers.

The pictures of the controller board show it to be well made but not perfect by any means. The twisted pair of heat sinks are caused by the heat sink fixing pin hole being incorrectly positioned. The soldering quality looks fine so it should be reliable on that front. Bear in mind that this board lives inside a heated chamber caused by the build plate heater running at 60 Degrees C. Small heat sinks are present on the three servo driver IC's and (extruder?) Power MOSFETS. The power supply switching IC,s are heats inked through the PCB which isn't to my liking. The components all appear to be reasonable quality. I suspect the electrolytic capacitors will be first to succumb to working in a warm environment. I may fit higher temperature rated quality capacitors at some time in the future. I will be looking at the board with a thermal imaging camera to identify hot spots and general operating temperature. I am also looking at forced air cooling of the board after enclosing it in a case to isolate it from the build chamber. This is because I will be raising the build plate temperature to 80 Degrees C.

The controller board is used in other UP! Printer models. It just has different firmware loaded to suit. There are unused connector positions on the board and I have shown them and their designations in the pictures. There are two unused fan supplies that I may make use of to cool the board. A door interlock is not used and other models do not have doors so this may have just been included in case of need. There is an unused connector paralleled to the 3.5mm socket. These connections are for the automatic build platform height calibration system used on other models, but not the Mini.

The two unused fan supplies are designated fan 5 and fan 6 yet I cannot find fan supplies 1 to 4. One of the models must have a lot of fans ! The Mini has only one fitted to the extruder head.
A useful 5v supply connector is also shown on the board. I will investigate these unused connector positions to establish if they may be enabled and used for upgrades.

The 'CPU' is mounted on a daughter board to enable easy installation. This is presumably to facilitate easy upgrade or installation of model specific CPU's. There is nothing underneath the CPU daughter board. The microprocessor has a 16MHz clock crystal but it's identity has been removed by erosion or abrasion. Why bother with anonymising this chip I wonder ?

The V3 extruder heater assembly in now an all in one solution instead of being a machined aluminium block with separate filament guide, heater and thermocouple. The new design is made from a Diecast metal and IMHO appears to be a cheaper approach that is not as good as the V2 extruder heater. Unhelpfully the OEM has also changed the extruder heater mount dimensions to prevent use of the V2 unit. The OEM extruder heater assembly costs the same for both the V2 and V3 spare parts..... An eye watering £100 each  . Fortunately Chinese copies of the V2 are available for £26 complete. If my V3 fails, I will fit a modified V2.

Well that's enough comment for now. Maybe more later.

Fraser

[Fraser]

Today I stripped down the extruder head for a clean.

I can state that the V2 and V3 extruder head assemblies appear the same except for a change in the heated nozzle head and the aluminium heatsink that clamps it in place. It is my opinion that the V3 heated nozzle head is inferior in both design and materials to that of the V2 head.

The V2 heater head uses a stainless steel combined filament feed tube and head mount. This is attached the the aluminium heater block using two stainless steel machine screws and feeds into the Brass extruder nozzle.

The V3 heater head appears to be made from Mazak (Zinc + aluminium + copper) aka Zamak. It is a one piece casting with the filament feed tube and head mount integrated into its shape. The heater and temperature sensors are the same as those used in the V2 head and are secured into the moulding using ceramic paste (as used to secure projector lamps into their reflector assemblies).

The V3 heater head is a far more fragile unit when compared to the V2. Mazak is a very brittle material and the heated head is secured to the extruder assembly via a very thin Mazak tube. The heated nozzle is screwed onto a Mazak thread. If too much torque is applied, the thread section will shear away.  The V3 is definitely cheaper to produce as it is a one piece design and negates the need for screws or assembly beyond fitting the heater and temperature sensor. IMHO though, its a cr*p design.

I shall be modifying the extruder head heatsink to take the V2 heated head assembly. You will see in the pictures that the V3 heated head mounting point has had one dimension reduced by removing the curved 'cheeks' that are present on the V2. Thankfully that just requires me to remove material from two sides of the heatsink mount. The other dimensions remain unchanged. I have to wonder if this small change was introduced to prevent the use of the far cheaper, but equally good, Chinese copies of the UP! printer heated head. £100 Vs £26. I know which I will use !

The extruder head assembly is mounted onto the printers chassis using three strong magnets. At first I thought this a bad idea but it has proven reliable. It is designed to reduce or prevent damage if the build platform is driven up into the extruder nozzle. The extruder assembly just slides up and out of harms way 

The extruder assembly consists of the heated nozzle head that is fed with the 1.75mm ABS or PLA filament using a stepper motor driving a cog in close proximity to a ball bearing race. The filament is trapped between the ball bearing race and the rotating cog. A neat design that is simple yet works well. The plastic chassis of the Extruder assembly is made by a 3D printer in ABS. The OEM provides the STL files to enable the owner to print replacements or update the design as and when the OEM makes design changes. Each issue of the printers software contains a new set of STL files for all the ABS parts of the printer. Nice idea. There have been several versions of the Extruder assembly chassis with minor tweaks and improvements incorporated.

As parts wear out you can print new ones which is important as the Extruder assembly does suffer wear in the filament path as it passes through the ABS. The alternative approach would have been to make the Extruder assembly from aluminium, but that would have increased the production cost considerably. I suspect users could use the STL files to have parts produced from different materials or with a finer finish.

The Extruder assembly is also fitted with a cooling fan that is dual purpose. It cools the interface between the heated head and the stepper motor to which it is mounted (via the heatsink) and a down-facing vent can be opened to blow air onto the printed layers to help them cool. This is used for PLA printing rather than ABS.

In order to help the user in assessing the print, the assembly also contains three powerful LED's to illuminate the build platform and extrusion process. The lighting is toggled on or off using the single front panel "Power on/off/stage lighting" button. Neat idea.

Well enough writing about the Extruder head.... on with the pictures........

[Fraser]

The V3 and V2 Extruders detail.

The V3 is more rounded as it is a moulded Mazak item.

The V2 is made from separate components and a machined aluminium block.

[Fraser]

Detail of the heated head mounting point in the heatsink. The width of the rectangular hole needs to be increased by approximately 1mm to accept the V2 heated head feed tube. The heated head feed tube is clamped between the heatsink recess and a spacer that may be seen nearest the stepper motor. Note I have temporarily reassembled the spacer incorrectly with the notch in the wrong place. I have now corrected that 

[Fraser]

That's all folks

Fraser

[SeanB]

Interesting there Fraser.

[Fraser]

Another UP! Mini owner has kindly annotated an image of the controller board. He misidentified the 3.5mm port as for diagnostics (JTAG) when it is in fact the platform height sensor input (not used on the Mini)

Credit to 'Brainzilla' on the pp3dp forum

http://www.pp3dp.com/forum/viewtopic.php?f=28&t=22192&start=60#p42739

Fraser

[Fraser]

A small update on the UP! 3D printers.

I am going to modify both of my printers to provide the motherboard with some thermal isolation from the heated build chamber. I am doing this because the present open design means that the CPU and other heat generating components are running very close to their thermal maximum due to the ambient offset of around 40 degrees C. The CPU has a maximum temperature rating of 85C and has been known to play up at 70C. There are reports of CPU failure as well. As the OEM wants almost $400 for the CPU board alone, I intend to protect it as best I can from thermal stress.

I have designed a metal thermal screen that encloses the motherboard and allows me to employ a small fan to force air through the enclosure and cool the motherboard. The UP! Mini only has passive convection cooling via a pair of rear panel louvre sets that are not in any way efficient. I have already modelled the screen in card so I know its size, shape and required hole  locations. I can recommend making mock ups in card to anyone doing similar. Its a good idea and can save a lot of time due to errors on a final product.

I purchased some 0.9mm aluminium sheet, a selection of small 12V 40mm fans, a 38mm hole cutter and a nice new 18" capacity sheet metal bender. I now have the parts and tools to make the final aluminium thermal brake. I will fit the modification to both of my printers and release details of it here in due course.

I dislike the Version 3 Extruder hot end that UP! now use. For this reason I purchased some Version 2 hot ends to use. A far better design all round. I wanted to recover the original UP! heater and PT100 sensor so this evening I took a hack saw to the Version 3 hot end and removed a section from the heater and sensor tubes. The heater and sensor are held in place with thermal cement and trying to drift them out of the tubes risked damaging them. Opening the tubes proved reasonably easy even with a full size hack saw (I only had that to hand). I should have used a junior hacksaw for greater control but I got away with it.

After extracting the heater and sensor elements I cleaned them up with a Castle block and they look like new again, ready for reuse. Cleaning revealed teh heaters rating. It is marked "1311 80W". It is a 24V 80W heater, as expected. This is higher power than the heater used in a RepRap which uses a 40W unit.

I attach pictures of the extracted heater and PT100 sensor. Both survived the operation and are ready to be fitted into one of my Version 2 Aluminium shells. I will secure them in place with common thermal cement that is used in fire places etc.

Fraser

[Fraser]

Another modification that I intend to install on the two UP! Mini printers is designed to address the well known issue of the build platform temperature being too low for large ABS object printing, at only 60 Degrees C.

Many users have found that large ABS objects tend to suffer warping of the lower layers due to insufficient heating of the build platform. The larger brother to the Up! Mini uses a build platform temperature of around 110 Degrees C !

Users of the UP! Mini soon established that the printer used a very simple temperature control for the build platform. An etched FR4 PCB forms the element of the heater and this is supplied with approximately 20V D.C. Without any temperature control, the platform rises to a temperature of around 85 Degrees C. This temperature has been found to solve the warping issue with large projects.

The thermostat employed by the OEM for this printer appears to be a cost reduction measure as it is simply a TO220 housed bi-metallic fixed temperature device placed in series with the heater supply. You can hear it clicking on and off just like the thermostats found in old clothes irons ! There is bad news on two fronts with the use of this particular component. Firstly, the OEM used a 60C thermostat for reasons that are not fully understood. And secondly the component is rated at only 1A ! These thermostats are known to fail open after a period of time, likely due to overloading. Some users have elected to short out the thermostat and run the build platform 'open loop' at whatever temperature it settles at. Others chose a slightly better option and fitted a higher temperature thermostat. The component is not rare or expensive.

I have bought three potential replacements for the thermostat fitted in my machines....... 70C, 80C and 100C. The 80C was a likely candidate as I know it works having read of others using it with success. The 100C device was bought in case I decided upon a different temperature control method and just needed a safety cut-out in case something went badly wrong with the control loop.

I am still undecided on which path I am going to take but I have just purchased some very neat little PID's that run on 12V and can handle switching 10A via the internal relay. They only cost around £7.00 from a UK warehouse !  These are looking like a good option as they will actively control the build platform temperature at whatever level suits the task at hand. PLA printing does not need a heated platform and 80C would be counter productive. I will also gain a visual check on the platform temperature so as not to start printing until the platform reaches the correct level. The PID's sensor is a common 10K NTC thermistor. One is provided but I may use something smaller and with more flexible leads as it attaches to the moving platform.

It will be a relatively easy install as 12V is available from unused Fan connectors on the motherboard and breaking into the platform heater feed coming from the motherboard is pretty simple too. The only challenge will be a decent installation of the thermistor.

Fraser

Pictures attached.

[SeanB]

Nice work there Fraser, you might also want to use some old PCB board under the aluminium plate as thermal insulation so that you reduce the radiated heat into the CPU area even more. Just a plate will heat up from radiated heat, and will radiate it into the lower chamber as well, though it will keep the airflow from the fan confined. With the heater, that cement is likely overkill, but will work. i generally use some Copaslip so that I can remove failed heaters again, but I work with 400W cartridge units, where the operating surface temperature is similar, but they have a screw that retains them, along with the stiff solid core PTFE wire, not flexing like that one.

The heaters rarely fail, just they have insulation resistance failures, which trips the RCD they are on. Even though the quick start the machine uses involves running the 220V 400W heaters at 400VAC for 4 minutes as a fast ramp up method. When I change to the new PT100 sensor at the end of the year ( part of the scheduled service at the moment) I will have to get some new relays as well ( hard to get a 5 contact plug in relay, and the original 40 year old Brown Boveri is starting to show it's age a little, so will be a quartet of 3 pole relays where I will do a little interlocking as well for safety), the machine is running on some heaters that are at least 20 years old. The failure there is the price of the Fenwel Thermoswitch, which is double the cost of the upgrade, and only available on order with a 2 month lead time. Lucky if it does die ( they fail as either open or a very large hysteresis range) I have a single usable one.