Qidi X Smart 3 revised fan installation

I have been slowly evolving my ideas for fume filtering when using the Qidi X Smart 3 printer. I should say there is nothing majorly wrong with the fan system as shipped but there always seems to be background residual fumes even when printing PLA. The level of fumes does seem to be dependent on the brand of PLA used.

The rear fan as shipped has no HEPA filtering, it simply vents the chamber into the external air. This fan only comes on during the printing process.

I have previously posted some early ideas for improving this using readily available HEPA filters. Following some discussion with Christian, a fellow X Smart user, we worked together to evolve this further.

The first phase was to upgrade my external fan filter duct design to have both HEPA and a carbon granule filter sections. The carbon granules are sandwiched between two 10 gauge stainless steel meshes and an outer cover holds everything in place.

The result of this looked promising but the axial fan as shipped lacked a decent air flow through the filter stack. This was upgraded to a 6028 centrifugal fan mounted on an adapter plate. The Fusion 360 assembly is shown below.

Below is a simplified cross section view. The internal fan adapter plate flange and the external filter stack are bolted together through the back wall of the printer using the original fan mounting holes and M3 screws and nuts. The fan is mounted on the adapter plate using M4 screws and nuts. There are moulded nut cavities to make assembly easier. The cover plate which holds the two meshes and carbon granules in place uses M3 screws into 3D modelled threads in the corner holes.

As previously posted I had fitted a full size Bento filter to my Qidi ifast. Further discussion with Christian revealed he had fitted the Bento Mini to his X Smart and seemed impressed by the internal air scrubbing action. I looked at the design of the Mini and after some thought changed the carbon cavity filtering walls from a printed grid to using the same 10 gauge stainless steel mesh as used above. The Bento Mini printed well and I was impressed by the thought that had been put into the design. There is a version with a hanging bracket designed for the X Smart 3.

I now had a belt and braces solution – the Bento for internal scrubbing and the HEPA/Carbon filter pod on the rear extract fan. Here are some images of all the components in place and the small terminal strip to interconnect the 24V supply.

The fans used on both assemblies are the same dual bearing 6028 models as detailed in the Bento write up. These are rated at 24V @ 80mA. The axial fan originally fitted was rated at 130mA. I removed the original fan and wired the Bento and the extract fan in parallel and connected this to the original supply feed from the control board via a small terminal strip. The two fans are only commanded ON during printing.

That is my hopefully my last solution but as ever it will depend on actual performance to see whether I notice the difference in air quality. Mission creep is always possible.

As mentioned in a previous post, I designed some booster feet to fit over the existing feet on the printer. These increase the air gap below the printer to allow more air flow. If you print them in TPU they gave extra stability to the printer and reduce resonances.

Here is a link to the STL files and write ups for the simple duct, the fan adapter plate, the two stage filter duct, booster foot and also the modified Bento carbon box and lid with mesh divider walls. Note of late I have been printing with a setting of four perimeters which gives stronger modelled threads.

Thanks again to Christian in Germany who has been a great help in bringing these various mods to fruition.

Links to similar or related post are listed below : –

Qidi X Smart 3 tweaks

From the title you will have guessed that I could not resist the temptation to buy an X Smart 3 while they were on special offer at GBP299.

The X Smart 3 is a lovely little machine. It prints excellent high quality models at a very fast speed. My testing suggests around a third of the time as on my Qidi ifast but of course it has nothing like the same chamber build volume.

I have had some issues with the X Smart 3, some of which were finger trouble on my part and some that needed recourse to Qidi’s excellent support team.

Attached to the link below is a ZIP file that has the full write up detailing my modifications to the X Smart 3 to add a LAN socket, modify the processor fan operation and to add extension feet to raise the body of the printer to allow more air flow. The ZIP file contains the PDF document and two STL files, one for the extension feet and one for a printed template to aid positioning the hole for the LAN socket.

Note that the LAN connection will not appear on the printer control panel but is visible by the Qidi Slicer application. The IP address allocated will be automatically defined and you will need an application like Fing to discover what this is. The control board will be discovered as a Raspberry board.

Since the above write up I have progress on a HEPA filter housing for the X Smart 3 rear fan. Below is a write up on this and some changes I also made to my i-fast. Note that since posting this write up I moved the rear fan filter housing inside the chamber so that the fan is pulling air through the filter rather than trying to push it through. This worked out nicely with the rectangular shaped filter housing.

I have also made a replacement top cover with a large HEPA so air can get in and out of the chamber to mitigate having to leave the lid off.

Update – Fan Noise

The only remaining frustration with the printer is the power supply fan noise. This is present all the time the printer is running whether under full print load or just on standby. Unlike the ifast the X Smart 3 only has one power supply. It is therefore impossible to get the power supply to switch off when it needs to be on to run things. Contrast the ifast where it goes completely quiet. There are suggestions that there maybe better substitute power supplies that have proportional speed fan control driven by the supply current demand. I debated this but decided there would still be noise even when backed off.

My solution is a bit steam age. I bought in an Alexa compatible AC switch. I now tell Alexa to turn the printer on and off. Much easier than trying to get round the back of the printer to the power switch.

On the subject of Alexa devices, I also bought in a Google/Alexa air quality monitor so I can check the VOC level around the printer. This seems to be quite sensitive and suggests the HEPA filter modifications are working.

Links to similar or related post are listed below : –

Automated 3D printed collet storage using Fusion 360 parameters

How sad is this ?

I recently bought an Imperial set of ER25 collets to compliment my metric versions. I also have a set of metric ER20 and ER11. They have all been delivered in little yellow boxes which is good from a shipping point of view but a faff when needing to use them. In the past I have made storage blocks using odd pieces of wood with tapered holes cut using a cone cut drill bit. The cone cut has a 6 degree taper and ER collets all have 8 degrees so it sort of works but not quite ideal for my perfectionist mind.

As ever, the 3D printer sits in the corner begging to be used and Fusion 360 challenges me to do something a bit more professional. The scene is set for an overengineered collet storage tray.

Now I could just model and print something to hold the twelve new Imperial size delivery but what if I add to them at a later date? What if I decide to get some ER20 Imperial versions also ? ……….. Could I create a Fusion based automated model that will cope with any sized matrix of holes and any size of collet?

The first exploratory step was to create an Excel spreadsheet (another of my fascinations) as a means of identifying the steps that might be involved in automating the design and to tabulate the different parameters of the various ER series collets. Here is a screen shot.

A little bit more detail will be needed for you to understand what this all means.

The top table just details the published specification dimensions of the ER range of collets. I needed the height of the tapered section and this is calculated. Note that not all collets have the parallel section at the top of the taper. If you extrapolate the taper it then coincides with the dimension D1 and this is one I have used in the calculations.

Each collet holder hole in the block will be an 8 degree tapered hole that the collet sits in such that it does not sit all the way through the block (I opted for a 1mm bottom gap) and such that a percentage of its height protrudes from the top of the hole (I opted for 30%). The middle right hand sketch shows the various parameters and the red outline represents the collet.

The lower table shows the set up for the model parameters with spacing and borders defined. This section allows entry of the all important ‘how many holes do I want’ and the aspect ratio of the matrix. The final two red lines shows how big the resulting storage block will be.

All with it so far ? This is so embarrassingly sad … should I even be thinking of posting it…

These various parameters are now mirrored and entered in Fusion 360’s Fx parameter set up. I have tried to be logical in the naming convention and the blue stars indicate a function that is defined and entered by the user and will also appear in parallel in the Fx Favourites listing. This allows a focussed entry of just the variable parameters without all the other background calculation clutter.

I wanted to round the resulting block size to a nearest whole number. My preferred function in Excel is “mRound” where you can define the rounding to an increment value. There is no such equivalent in Fusion. In the end I used the “Ceil” in Fusion and “Ceiling” in Excel but Excel’s “Ceiling” requires a ’rounding to value’ hence the entry line (R) in the Excel listing which is set to ‘1’ to make it match the Fusion calculations.

By setting the length and width hole count to ‘1’ a single holder can be printed and checked to dimension before doing a full matrix as needed. I found that on my first single hole print the collet sat slightly higher in the hole than designed but this was traced to a print artefact on the taper wall. This was was easily rubbed down. On similar tack, I have allowed a fudge factor for print shrinkage should this be needed. Any errors in this respect will cause the collet to sit higher on the block. This will be most apparent as the collet size reduces. The value entered (say 0.1mm) adds a linear amount to the top and bottom hole sizes and therefore the taper.

Once you are happy with the single hole print you can define the number and aspect ratio of how you want the holes to be printed on the finished block. A 6 x 2 print will match the normal 12 piece ER25 collet set. On the Qidi ifast printer this took around 4 hours at 0.25mm layer height. Note that I used the ‘Shell’ command on the block lower surface of the print to reduce material use. This means you need to place on the printer bed ‘bottom side up’.

Well I did say this was going to be a sad nerdy post but as ever I learned a bit more about Fusion, Excel, refreshed my school geometry and made the workshop even more tidy and organised. What’s not to like ? …..

While this is an almost facetious waste of time and effort, the principles used in the Fx programming has many other applications.

The planning spreadsheet has been added to my spreadsheet compendium which along with the Fusion file (for ER25) is attached on the following ZIP file link.

Links to similar or related post are listed below : –

3D Printed Four Facet Sharpener

A Hybrid Version of John Moran’s Elegant Design

While investigating four facet drill sharpening, I came across John Moran’s website which detailed his design for a four facet drill sharpener.   Four facet sharpening is popular for providing more accurate starting of a hole, less drilling pressure, more accurate hole size and better retention of drill point sharpness. John’s design is intended for sharpening 2mm to 12mm drill bits.

I have a Quorn grinder but it is not something that I can quickly pull out from under the bench to sharpen the odd drill.   John’s design looked easy to set up and use and certainly easier to carry. I have been working with him to produce this hybrid version. I recommend that you read John’s write up before embarking on either his original design or my 3D printed hybrid version. He also presents a YouTube video on how to use the sharpener.


Here is a picture of John’s finished design. Very nicely finished and presented.

Here is a picture (not quite so pretty …) of my 3D printed hybrid version. A new baseplate might be in order to smarten it up.

The concept of combining 3D printed components with essential metal parts works very well and the resulting sharpener appears to be stable and repeatable in its results. There is a huge advantage in combining technologies in this manner. All the ‘boring’ or ‘dreading to make / leave until tomorrow’ parts are easily produced on a 3D printer and the remaining metal parts are straightforward to produce with conventional resources and skills.

Rather than go into details as a blog post, below is a link to a ZIP file that contains my full write up, 2D PDF drawings of the metal parts, STEP files of the 3D printed files and the raw Fusion files for those wanting to tweak.

If you can’t work with STEP files then let me know and I will run STL versions. STEP files are more accurate than STLs and also have the advantage of being more easily editable than STLs.

I think you will be impressed with the end result of John’s design.

If this write up or my general waffle is of interest then please subscribe to receive notices of new posts. If you build the sharpener and it works let me know. If it doesn’t work and it’s my fault for not providing enough information than email me so I can correct things.

As an alternative to a 3D printed version you can watch Mark Presling on YouTube who has recently construct the same sharpener in metal.


I have already had some feedback on the 3D printed design so some of you are quick off the mark.

This feedback related to the friction of the collet holder shank in the block and also the print roughness against the depth setting screw.   Both of these will relate to the print quality from your printer.

As an experiment I ran a new version of the block with a larger bore and then lined it with an aluminium sleeve that was a smooth fit on the shank.   I also turned a flange on the end of this sleeve so that the adjuster nut had a smooth metal surface to ride against.

The addition of the sleeve looks promising in solving these two issues.  I admit it is going against the concept of a 3D printed set of parts but the sleeve is easy to turn up on the lathe and the new block in PLA is still much easier to produce that trying to make the block totally in metal.

You can download the edited write up, new collet chuch holder block STEP and the sleeve drawing in this add on ZIP file.

Links to similar or related post are listed below : –

A 3D Printed Passive Demagnetiser using Neodymium magnets

I saw a recent post on the HomeMadeTools.net site from Tony Foale about using Neodymium magnets in a rotating holder to act as a demagnetiser. 

Tony made the holder from aluminium but from my point of view with Fusion and 3D printing facilities the idea lent itself to be run as a 3D print.  The finished device can be held in a lathe or drill press chuck using the central boss.  The boss does not have to be rigidly clamped as the magnets will grab and locate the assembly up to the chuck.   The magnets are very strong so be very careful as you bring the object to be demagnetised up to the lower surface or it will also get ‘grabbed’. 

Attached is a ZIP file containing the Fusion 360 file.   This is designed to take 10mm diameter magnets which will be a tight push into the mounting holes.   I used 10mm long magnets but smaller ones will probably work just as well. The magnet mounting holes have a 1mm thickness bottom surface.   The magnets need to be inserted with the alternating orientation as shown.  Tony made a circular ‘keeper’ for the bottom surface from a steel disc.  This reduces the attraction to swarf.   The ZIP file also contains a simple lid to fit over the top side to protect the holes from accumulating magnetic swarf.

Screen shot of the demagnetiser fusion file
Pictorial screenshot view from Fusion 360 of the demagnetiser 3D model

ZIP file download is on the link below and contains the body and lid as a Fusion file and as separate STEP files  – demagnetiser disc body and lid

A closing thought (which I haven’t tried …) but if all the magnets were in the disc the same way round would this act as a magnetiser to save having to stroke the item with a magnet ?

Similar or related subjects : –

Verified by ExactMetrics