Temperature Display for 3D Printer Enclosure

Posted 29 July 2021,

In my ongoing quest to convert my MakerGear M3-ID printer from a nice bench decorative piece into a real functioning printer.

Recently I have been having real problems with using dissolvable filaments with my MakerGear M3-ID dual-extruder 3D printer. I couldn’t get either the PVA (water soluble) or HIPS (Limonene soluble) filaments to stick worth a damn to the BuildTak surface. In the process of troubleshooting the problems, I discovered that the M3-ID has real trouble getting the print bed temperature above 100C – at least in my nicely air-conditioned lab spaces. So, I went on the hunt for a decent enclosure, and found this 3D Upfitters model.

3D Upfitters Enclosure for the MakerGear M3 series (M3 shown, but the same enclosure works for the M3-ID)

The dimensions shown for the enclosure look like they would work for my setup, so I ordered one – we’ll see. In the meantime I started thinking that I might like to control (or at least monitor) the internal temperature, especially the ambient temperature at the location of the Octoprint module and control electronics. My worry is that at high bed temperatures, the ambient temps might get worryingly close to the max temps for the control electronics. 3D Upfitters does offer a temperature readout, but I was pretty sure it wouldn’t accurately represent the ambient temps around the electronics, so I decided I would modify an earlier project to create a custom temperature probe, using the venerable Nokia 5110 LCD display, a DHT11 temperature/humidity sensor, and a Teensy 3.2.

I decided to use the Teensy 3.2 micro-controller rather than an Arduino UNO to avoid the issue with 5/3.3V level conversion and because I’ve used this item several times before in other projects. I’m sure there are cheaper alternatives, but this is what I had available, and they are rock-solid products. Here’s the schematic:

And here are some photos:

And here is the code that interfaces to the sensor and drives the display:

Stay Tuned,

Frank

Caplet Dispenser

Posted 13 July 2021

Being an old fart, I have unfortunately accreted a number of meds that I must take on a daily basis; the current count is four different meds – two in caplet style, and to in pill style. Each night I have to remove the cap from four different bottles, extract just one pill/caplet, and close the cap, all without losing any. After the umpteenth time that I either lost a pill on the floor, or had some very expensive caplets literally go down the drain, I said “there must be a better way…”.

So, I started designing a mechanism that would dispense just one pill/caplet on each cycle. I started out by creating TinkerCad models of the pills and caplets, as shown below:

Meds to be dispensed. Large caplet is about 7 x 19 mm

Then I started working on a dispensing mechanism, and wound up with the following design.

Caplet is loaded in upper image, dispensed in lower one

The mechanism consists of a sliding drawer and a collar with a caplet-sized slot. In the ‘load’ position (upper image above), the caplet falls through the slot into a caplet-sized bay in the drawer. In the ‘dispense’ position, the caplet falls through the open bottom of the drawer bay. Each time the drawer is moved back to the ‘load’ position, another caplet falls into the dispense bay, and is dispensed when the drawer is moved back to the ‘dispense’ position.

The above design worked very well, but there were two significant problems; the drawer wouldn’t stay in the ‘load’ position, so care had to be taken to avoid dispensing several caplets at a time, and it was possible for two caplets to fit vertically into the slot, jamming the mechanism, as shown below:

Two caplets oriented vertically, jamming the drawer mechanism

To keep the drawer in the ‘load’ position, a small rubber band was attached to the collar (the red part above) and around the right end of the drawer, using a hot-glue gun. This keeps the drawer in the ‘load’ position until actively pushed against the rubber band tension to the ‘dispense’ position. The issue with two caplets jamming the drawer was solved by placing a ‘flap’ over the slot in the collar, reducing the hole size such that only one caplet at a time can fit. A caplet goes through the hole vertically, and then slides down into the drawer bay, winding up horizontally in the bay, as shown below:

final design showing reduced-size collar slot allowing only one caplet at a time into bay

The next step in the project was to design and fabricate an adaptor piece to connect the dispensing mechanism to the pill bottle. TinkerCad doesn’t really support morphing from one shape to another, so I had to find a different way. I tried Blender, and while it did work, I had no experience with the product and so stumbled around a lot. Next I tried Open SCAD and discovered the ‘hull()’ feature, which does pretty much exactly what I want. After playing around with this a while, I came up with the following OSCAD script to do what I wanted:

The above code and parameter set produced the following model:

The cylindrical shape at the top just accepts a 56mm diameter bottle cap. The above model was converted to an STL file and then imported into TinkerCad, where it was mated with the dispensing drawer. Then the entire thing was printed in one go using my Prusa MK3S 3D printer, as shown below:

Prusa Slicer 2.3.0 showing ‘sliced’ model with supports, ready to print. This print takes about 3.5 hours.

After the print finishes, the support material between the drawer collar and the drawer itself has to be removed manually with an Exacto knife. This is a bit of a PITA, but worth it to have the entire thing printed as a single piece. The photo below shows the finished product.

30 August 2021 Update:

I just recently acquired a Flashforge Creator Pro 2 dual-independent extruder printer to replace the MakerGear M3-ID I sold on eBay for about a third of what I paid for it. Even after a year and a half of diligent work, I could NOT get the M3-ID to print with dissolvable filament worth a damn. I tried everything, including a brand-new roll of PVA filament and adding a BuilTak removable build plate system (see this post). I even ordered the 3D upFitters enclosure for the M3-ID, but came to my senses before I assembled and installed it. After some more web research, I came across the FlashForge Creator Pro 2 IDEX (FFCP2) system, available for about 1/4 the price of the M3-ID, and this printer came with glowing YouTube recommendations from many reputable 3D printer enthusiasts. After receiving my FFCP2 and setting it up, I was able to print the above pill/caplet dispenser design using PLA for the structure and PVA dissolvable filament for the support material. The result was a very high-quality build and the support material dissolved out after just a few hours in warm water – YAY!

Stay tuned,

Frank

ILI9341-Based Digital Clock Project

A while ago, the clock/time display on our microwave started having problems; it’s a 7-segment vacuum-flourescent display, and a couple of the segments a no longer lighting up. Instead of “End”, we now see “Erd”, and the clock is getting harder to read. The microwave itself is still running fine, but because this particular display is our primary time display in the house (aside from our phones), not having a good clock display is annoying.

I investigated getting the appropriate repair part for the display, but that module costs more than the entire microwave did originally, so that didn’t seem like a practical idea. And while I can get a cheap stick-on clock for just a few bucks, what’s the fun in that?

So, I’m in the middle of a design project to build a digital clock to replace the one on the microwave. At first I built up a clock using a Teensy 3.2 and one of my spare Nokia 5110 LCD displays, but that turned out not to be very practical, as the display is basically unreadable from more than just a few feet away, with or without backlighting.

So, I started looking around for better displays, and ran across the ILI9341 TFT display with a touch-sensitive screen. Even better, this was available with Teensy-optimized drivers, as shown here – what a deal! I started thinking that with a touch-sensitive screen, I might be able to implement some on-screen buttons for setting the time, which would be way cool!

So, I ordered two of these displays from PJRC, and started working on the design. Here’s the initial schematic

Here’s my initial breadboard setup:

Teensy 3.2 in foreground, Adafruit DS3231 RTC in background, ILI9341 TFT touch display

I originally used the Arial proportional font provided by the library, but I discovered that it produced bad artifacts after a few hours, as shown below: The only way to avoid these artifacts is to refresh the entire screen on every pass through the 1-second timing loop, which causes a very annoying ‘blink’. Eventually I figured this out, and changed to a non-proportional font, as shown below;

Non-proportional font looks worse, but doesn’t suffer from artifacts

Once I got everything working (well, except for the touch screen stuff that I plan to add later), I fabricated a nice box using Open SCAD and TinkerCad, and made a more permanent version using a half-sized ‘perma-proto’ board, as shown below

Here’s the more-or-less final schematic

7 March 2021 Update:

After a long and somewhat agonizing journey, I finally got the touch-screen stuff working, so I can now adjust the time on my digital clock using on-screen touch-sensitive controls. Here’s the updated schematic:

The only real difference between this one and the previous schematic is the T_IRQ line is no longer connected; the Teensy-optimized XPT2046_Touchscreen.h/cpp library doesn’t use it.

Here’s a short video showing how the touch-screen enabled time/date adjustment feature works:

Although I hope to clean up this code considerably in the future, I include it here in it’s entirety in it’s current state

10 March 2021 Update:

After getting all of the above to work, I decided to re-tackle the proportional fonts issue. In my first attempt, I had used the ‘ILI9341_t3’ version of the ILI9341 Teensy library, and there is a newer ‘ILI9341_t3n‘ version out now. So, I modified one of the example programs (unfortunately still written for the old library) to be compatible and got proportional fonts (well, just the Arial one) working on the display.

After several hours of running this example with no apparent artifacts or problems, I decided to update my complete clock program to use the Arial font. After making the modifications and running it overnight, the time & date displays were still rock-solid the next day, as shown in the following photo – YAY!!

So now all that is left to do is to upload the new Arial-based code (with the time background color switched back to ‘black’) to my ‘working’ clock module, sit back and enjoy the proportional font display.

14 July 2021 Update:

I noticed that my clock had some ‘issues’ with time/date adjustments, so I ‘put it back in the shop’ for some additional TLC. While I was at it, I noticed that the system schematic didn’t include the DS3231 (hard to have a clock without a RTC!), so I updated it as well. Here’s the updated schematic.

The updates made were to make the time/date adjustments more robust. The updated code is included in its entirety below. First the main program:

And then the ‘CustomBox’ class file (no .cpp file – everything is in the .h file)

Stay tuned,

Frank

MakerGear MK3-ID BuildTak FlexPlate System Installation

I’ve had my MakerGear MK3-ID for a year or so now, along with a Prusa MK3S. The Prusa MK3S with it’s removable flexible build plate is my ‘go to’ printer for almost everything that can be done with a single extruder. The flexible build plate is wonderful – it makes printing so much easier. Fortunately there is now a comparable option for the MakerGear M3-ID – BuildTak’s 8×10 FlexPlate System. I was able to find the correct size system for my M3-IE at Partsbuilt 3D and it arrived in just a few days (unfortunately with Ohio state tax applied as both the company and I are in the same state). This system makes a HUGE improvement in the ease of printing with the M3-ID, and consequently make it a much more appealing alternative to printing everything on my Prusa. Here’s a photo of the FlexPlate base on my M3-ID.

However, there were a few ‘gotcha’s’ during the installation, necessitating some modifications to the M3-ID.

  • The stock M3-ID comes with a 4mm glass build plate, mounted on top of the heating element, and held down with screw-down rotating clamps at each corner . There’s no need to keep the glass build plate with the BuildTak FlexPlate system, but the resulting 4mm gap means that the stock clamps will no longer hold the build plate down firmly (or at all, for that matter).
  • The FlexPlate system consists of a mounting plate with a number of embedded magnets that is affixed to the flat heating plate with an adhesive film, and a flexible steel build plate onto which the actual build surface (either a BuildTak or PEI sheet) is adhered. The steel build plate attaches magnetically to the mounting plate, and can be easily removed and flexed to pop the print off, a la the Prusa MKS system. However, it turns out that while the mounting plate with embedded magnets mounts quite nicely on the heated base, the 8×10″ steel build plate is just a smidge too large in both dimensions for easy mounting/dismounting. It tends to overhang in one or both dimensions, meaning that one corner rests atop its corner hold-down block, and then the plate isn’t quite flat. This problem could easily be fixed if the steel build plate was just a few millimeters smaller in both dimensions.
  • The Prusa MK3S has a simple and elegant way of achieving excellent physical registration of the flexible build plate when it is placed onto the printer. There are two small protruding posts (3 or 4mm screws, actually) at the rear of the print area, and the flexible plate is notched so that when it is placed against the screws, it lines up perfectly every time. There is no such physical registration feature on the M3-ID, so getting the plate down correctly so it isn’t overlapping one of the corner pieces is a bit of a PITA

So, I needed to make a couple of modifications to the M3-ID corner hold-down system. The idea was to make the build plate fit without binding, and to achieve better physical registration of the build plate when it is re-installed onto the M3-ID after popping off a print.

The first thing I did was simply place a couple of 3mm hex nuts under the hold-down clamp foot at each corner, filling the gap left by the 4mm glass build plate, as shown below.

2ea 3mm hex nuts under the hold-down clamp foot, replacing the thickness of the removed glass build plate

This worked – sorta – but it wasn’t very elegant and didn’t address the other issues, i.e. the flex plate interference, and the problem of physical registration. So, I started designing some alternate hold-down clamp and corner pieces to do a better job. The first effort was a modifed hold-down clamp, as shown below:

Modified hold-down clamp requires only 1ea 3mm hex nut spacer

The thing I love about 3D printing is that you can iterate through several designs rapidly and learning from each version. I usually go through half a dozen versions before getting where I want to go, but it is so cheap and fast that doing it this way is MUCH faster than trying to do a ‘moon-shot’ on-off solution (which generally doesn’t work anyway, due to some unknown unknown).

So, back to the drawing board (TinkerCad) and some more designs. I started with designs to replace the stock corner blocks, and after six revs, had the following final design

MakerGear M3-ID replacement corner blocks – V6A & B are left/right mirror images

and the final design for the hold-down clamp:

MakerGear M3-ID replacement hold-down clamps – left/right mirror images

Both the corner blocks and hold-down clamps were printed with left/right mirror images to match the mirror-image corner requirement.

After getting all the final versions printed up, I installed everything on the M3-ID, as shown below:

Here’s a short video showing me placing the flexplate onto the magnetic base, using the rear corner blocks for physical registration.

Me placing the BuildTak build plate onto the magnetic base

I’ve already made a number of prints with this system (all the corner blocks – about a dozen by the time I was through) and I have to say it’s like night and day compared to the hassle with my stock M3-ID. If you have a MakerGear M3-ID, this is a must-have!

I created a Thingiverse ‘thing’ with the final designs for the corner block and hold-down clamps.

Stay tuned,

Frank

Dewalt DWS713 Miter Saw Dust Port Vacuum Hose Adaptor

My wife got me this really cool Dewalt DWS713 miter saw for Christmas this year, and I have been having fun pimping it out a bit. In a previous post I described how I added a cut shadow line LED light and control box, and this post describes how I added a 3D-printed coupler from the saw’s sawdust exhaust port to a small vacuum I keep in the shop for small cleanup jobs.

My first attempt at a coupler was a straight piece that connected on one side to the exhaust port, and on the other side to the rigid end of the vacuum hose, as shown below:

Unfortunately, I had neglected to consider what was going to happen when I actually tried to use the saw. As soon as I released the downlock and raised the saw, the coupler snapped when the rigid part of the vacuum hose ran up against items on my bench behind the saw – oops!

So, I started over again. I found the TinkerCad ‘bent pipe’ shape generator, and made a few versions that incorporated 45-60 deg bends into the rigid portion of the vacuum hose, but all of these suffered from the same problem with physical interference and cracking when the saw was raised. After some more head-scratching, I discovered I could remove the rigid end of the vacuum hose, leaving just the flexible part. As shown below, this particular vacuum has a detachable end piece that fits inside the flexible hose

Flexible hose connector. Note circular ribs.

So, I decided that I could replace the rigid piece from the end of the vacuum hose with a short coupling piece to connect the sawdust exhaust port directly to the flexible hose. As I normally do with complex projects, I started by printing a test piece – just the portion that couples to the flexible vacuum hose, as shown below:

Short piece to test the flexible hose coupling geometry

Once I had the flexible hose coupling geometry nailed, I did the same thing with the other end, and then connected the two coupling ends with a ‘curved pipe’ shape from TinkerCad, resulting (after a number of revisions) in the piece shown below:

Version 10 – this one worked!

As can be seen in the above screenshot, the final working version was version 10. One of the more wonderful things about 3D printing is the ability to make and discard multiple revisions – all it costs is a little time and a bit of very cheap filament. No need to hyperventilate over mistakes – just throw it away and try again!

The next few photos show the finished coupler installed on the miter saw.

I had uploaded the previous straight-line coupler to Thingiverse here, and I edited it to provide the new design as well

Stay tuned!

Frank

Lab Power Supply

Posted 30 June 2020,

Almost exactly one year ago I ran across some posts regarding a very nice lab power supply regulator and display called the DPS500x power supply front-end.  My existing linear style supply was getting a little long in the tooth, and ran out of gas pretty quickly above about 12V and 1A.  So, I decided it was time to upgrade, and wound up with a very nice, lightweight well-performing unit.  Unfortunately I forgot to document the project and when I recently wanted to point someone else to this nifty product, I didn’t have anything comprehensive to point to!

So, this post is a belated documentation post for the project.

There are three ‘big’ pieces (big in terms of importance, but not in actual size or weight) used in this project; the controller head itself, an appropriate housing, and a AC/DC power supply that will fit inside the housing

The DPS5005 50V 5A constant voltage/constant current digital controller head

This is a really neat little gadget that takes a DC power supply as input and steps it down to the desired output voltage using a highly efficient buck voltage converter technique, and applying some constant voltage/constant current magic to the output, all in a package that is maybe 40 x 60 x 30 mm.  I got mine from Banggood.com, but they are available everywhere.

DPS5005 from Banggood

A suitable housing

In this case, an aluminum two-piece housing and hardware kit custom designed to house the DPS500x series of power supply controller heads.  The upper and lower halves slide together via a tongue-and-groove arrangement, and is VERY well done.  In addition, the housing has three sets of internal rails that make it easy to align/mount internal assemblies – NICE!

As the second photo above suggests, one suggested layout is to have an external 25-50V power supply connected to this piece, with basically nothing inside.  However, I did some research and convinced myself that I could fit a small open-frame AC/DC power supply inside the housing and wind up with a complete unit, just with lower wattage.

An AC/DC power supply

I wanted one that could fit inside the housing between the front and back plates to provide the ‘raw’ DC voltage to the controller head.  In the past I have used a number of Mean Well supplies and found them to be small, reliable, and cheap, so I started looking for a unit that could deliver 24V at 2A or greater (the most I thought I would need for a general-purpose bench-top power supply) while still fitting into the housing.  After a bit of research, I found that the Mean Well EPS-65-24 24V 2.7A open-frame power supply would do the trick nicely, and was available from Mouser for $13.80

Side view of power supply with model number shown

With all the ‘big’ parts identified and ordered, there some smaller issues that needed to be addressed:

Custom 3D printed back panel:

Because I was planning to use an internal AC/DC power supply rather than an external one, I needed a panel-mounted AC plug instead of DC Banana plugs, and I wanted an AC ON/OFF switch as well.  So using the basic dimensions and layout of the existing back panel, I designed a new one in TinkerCad to meet my needs.  I found some designs for 40mm fan grills on thingiverse and used them to create a cutout directly into the back panel, and took the cutout dimensions for the power switch and AC plug from the manufacturer’s specs for the parts.  When I was finished, I had a nice 3D printed panel as shown below:

TinkerCad back panel design

Custom back panel installed on power supply housing

Power Supply Mounting Rails:

The housing sports three sets of internal rails that can be used for parts mountings, so I designed and printed some ‘runners’ that would attache to the bottom of the power supply and mate with the internal rail geometry, as shown below:

3D-printed runners to mate with housing internal rail structures

Tilt Stand:

The finished power supply worked great, but it was so small and flat that it was difficult to get my fingers on the controls, so I designed and fabricated a custom tilt stand, as shown below

Front view showing tilt stand. Ignore the missing screw 😉

Miscellaneous:

Mean Well AC/DC power supply AC input connector:

Mean Well AC/DC power supply DC output connector:

HP style AC cord panel-mount receptacle:

AC Power switch: KCD3 T85 16A 250VAC, 20A 125VAC.

Here’s a link to the 3D print (STL) files for the tilt stand and the custom rear panel.  If you don’t have access to a 3D printer, there will surely be someone in your local area who can print them for you.

 

Frank

 

Flashforge Creator Pro 3D Printer Motherboard replacement

Posted 07 October 2019

I have owned a Microcenter clone of the very popular Flashforge Creator Pro for several years now and just the other day it gave up the ghost and died; the internal LED lighting and the front panel LCD display started flickering, and then went dark altogether. This wasn’t an entirely unexpected event, as for the last week or so I had been smelling burnt insulation every time I used the printer.

So I turned the printer on its side and removed the bottom panel to have a look around. Everything looked normal until I examined the main power connector to the motherboard; this connector looked a bit brown and charred as shown below:

Motherboard power input connector. Note the discoloration and bubbling

If I wiggled this connector with the power to the printer enabled, I could get the display and internal LEDs to light up briefly and then go out again, so it seemed pretty reasonable that this was the problem.  After removing the motherboard from its mounting posts, I was able to get a better look at the connector, as shown below:

 

Motherboard power connector as viewed from the side

As can be seen in the photo, the plastic power connector has melted and bubbled out to the side, and the mating halves of the connector are fused together.  In order to disconnect the power cable I had to physically pry the two halves apart, as shown in the following photos:

Motherboard half of the burnt power connector

Cable half of the power connector

I didn’t know if the connector fried and caused the motherboard to die, or the motherboard died and caused the connector to die (or maybe a little of both?).    Anyway, I decided to try an replace the motherboard with a new one purchased on eBay.

When the new motherboard arrived, I started preparing for the exchange by carefully labelling all the cables, so I could make sure I got them back to the right places after the exchange. The labelling step is critical, as many of the motherboard connectors and the corresponding cables are indistinguishable from each other; without the labels there would be no way to tell which cable goes to which connector. Then I moved all the cables except for the power cable from the old motherboard to the new one, as shown in the following photos:

After moving all the cables, I still had a problem; the cable end of the power connector was so badly damaged that I couldn’t use it on the new motherboard, and without a power connection, there was no way to test the new board.  I solved this problem by temporarily disconnecting my after-market extruder cooling fan from the ‘EXTRA’ connector on the motherboard, and using that cable connector for the power connection to the motherboard.  After making this change, the printer came up normally when I applied power – YAY!!

So, I still had the problem of not having a connector for my printhead cooling fan cable.  After some more web research, I found this link by Aaron Gilliam (on Thingiverse of all places) describing the part numbers for the entire Flashforge Creator Pro in detail – thanks Aaron!

Flashforge Creator Pro Motherboard connector

Flashforge Creator Pro cable connector

The connector I was looking for was ‘2 pin DIGIKEY # ED2779-ND’ So, off I went to Digikey where I ordered several of the cable connectors, and also several of the mating motherboard connectors.  My plan was to first get the printhead cooling fan back on line, and then maybe try and replace the motherboard connector on the old motherboard to see if that was the only problem; if so, then I would have a complete spare motherboard available – cool!

Stay tuned!

Frank

 

 

 

 

Speaker Amplifier Project, Part VI – Second Production Run

Posted 29 September 2019,

I got an email from Dr. Betty Lise Anderson of the Electrical Engineering Department (I think it’s actually Electrical and Computer Engineering now) at The Ohio State University, asking me if I still had the documentation for the speaker amplifiers I created a couple of years ago for her STEM outreach program.  . Dr. Anderson said these units were very well-liked by her STEM outreach students; so well liked in fact that they apparently walked away on their own!  She asked me if I would be willing to fabricate another couple of amps, and said she would happily pay for all the parts.

Since I never throw anything away, I did indeed have the documentation and even some remaining parts from the original project.  I still had a half-dozen or so of the custom audio level indicator PCBs and at least one Adafruit 20W Class D amplifier left over.  I figured I’d need a couple of wall-wart 12V power supplies and one more amplifier – everything else was already available in my parts bins.  I figured the hundred bucks or so required to get all the parts was not worth worrying about, and besides I could probably write it off as advertising expense for EM Workbench LLC.

The enclosure:

When I made the first set, I 3D printed an enclosure that was a modified version of the nice rounded-corner box design published by Adafruit for just the amplifier.  However, when I tried this trick again, I wound up not liking the result.  Instead, I decided I should be able to create my own rounded corner box.  I searched around on Thingiverse and found a few parameterized rounded box designs, but they all seemed sort of half-baked.  So I broke out my copy of Open SCAD and started figuring out how to do it myself.  I ran across a video that demonstrated the rounded-corner technique using a ‘minkowski’ function, and then I was off and running.  After just a few hours (OK, more than a few, but definitely less than infinity) I had coded a nice, compact Open SCAD module to generate an arbitrary shaped rounded-corner box with an optional companion nesting lid.  The code is available on Thingiverse here.  Using the Open SCAD module, I generated an enclosure and companion lid and exported the result as an STL file, which I then sucked into Tinkercad to add the required cutouts and such for the amplifier project.

Amplifier enclosure as generated in Open SCAD

Amplifier enclosure after importing the STL file into Tinkercad

Amplifier enclosure after modification for the Adafruit amplifier and level indicator PCB

After getting the enclosure design all spiffed up, I started printing it on my trusty PowerSpec 3D Pro 3D printer, only to have it die on me – so much for ‘trusty’!  This was not an entirely unexpected event, as I had been noticing a ‘burnt insulation’ smell coming from it over the last few weeks, and suspected that it might be on its last legs.  So, this batch of amplifier enclosures would have to be single-color (the last one was dual-color red or the enclosure and gray for the text) – at least until my new MakerGear M3-ID 3D printer shows up :-)).  Here’s the result.

Amplifier and Activity Indicator:

In reviewing the documentation from the original project, I saw that the activity indicator schematic wasn’t entirely accurate, so I brought it up to date – mostly cosmetic/lettering, but…

View showing power indicator LED installation before installing power input terminal connector

View showing 2.2K current limiting resistor for power indicator LED

View showing connections between activity monitor PCB and amplifier board

The finished product:

Two complete amplifiers with companion power supplies

A large part of the motivation for this post was to thoroughly document all aspects of fabricating the second run of OSU/STEM speaker amplifiers, so that when I get that next call from Dr. Betty Lise Anderson… 😉

Frank

Custom B-Ball Face Mask Project

Posted 14 May 2019

In March of this year, I suffered yet another broken nose while playing basketball.  Off to the emergency room where, following the normal interminable wait, I was told “yep – you have broken your nose – here’s a referral to an ENT guy – have a nice day!”  The next day I went to the ENT guy, who said “yep – you have a broken nose, and there’s nothing I can do for you; you need an ‘open reduction’ (aka ‘nose job’), and here’s the name of the plastic surgeon I recommend”.  At this next appointment Dr. Bapna (the plastic surgeon) said “yep – you have a broken nose, and you’re going to need an open reduction.  It’s not going to be a whole lot of fun, but I should be able to get you squared away (literally)” (or words to that effect, anyway).

So, in early April I endured a ‘functional rhinoplasty’ (aka nose job), and indeed it wasn’t much fun.  Fortunately I had learned from an earlier rotator cuff operation that I could rent a powered recliner on a short-term basis, and this at least made the convalescence a little less terrible.

In the subsequent post-op appointments with Dr. Bapna, he made it quite clear that while the operation was an unqualified success, another broken nose while playing basketball might not be repairable.  He strongly recommended that I either give up basketball (and what is a 70-year old man doing playing b-ball anyway?) or wear a protective face mask.  Since I wasn’t really interested in giving up round-ball, I started investigating face mask options.

Some research showed that a number of clear face masks are available on Amazon and other retail outlets, and there were a few firms advertising custom face masks.  When I mentioned this to Dr. Babna, he told me that a local prosthetic business (Capital Prosthetic and Orthotic Center, Inc) also does custom face masks (who knew?).  Apparently the process involves making a plaster impression of the face, and then using the impression as the mold for a custom polycarbonate mask.  While I was researching the possibilities, it occurred to me that I might be able to use the knowledge of 3D modelling I had gained from an earlier project to create a duplicate of a chess piece to create a 3D model of my face, and then print a full-size plastic face replica to use as the basis for a polycarbonate mask.  This would eliminate the need to make a plaster impression, and might open up a new technique for custom face mask fabrication.

So, I talked my lovely wife into helping me make a 3D representation of my head, using the same Canon SX260 HX digital camera I used for my chess piece replication project.  It took us a couple of iterations to get enough good shots, but soon I had sucked 185 photos into Meshroom and it was busily cranking away to create the 3D model.

Except when it crashed.  I had experienced this problem during the chess piece project, and had solved it by finding and removing the problem photos, usually a shot that was badly out of focus. So, I found and removed the photo pointed to by the crash log, and restarted Meshroom’s processing.

And it crashed again, and kept crashing even as I removed more and more photos.  In addition, there wasn’t anything apparently wrong with the photos that caused the crash.

After a LOT of research on the Meshroom GitHub site, I finally ran across a post where one responder noted that Meshroom-2019.1.0-win64, the version I was using had ‘issues’ with photos that weren’t exactly perfect, and recommended downgrading to the 2018.1.0 version.

So, I downgraded to 2018.1.0, and voila – Meshroom processed all 185 photos without complaint and produced a startlingly accurate 3D model of my head, shown below

Screenshot of Meshroom 2018.1.0. From left to right; input photos, selected photo for comparison, textured 3D model

Leveraging on my experiences with the chess piece project, I immediately sucked the 57+ MByte texturedMesh.obj output from Meshroom into Microsoft 3D builder, and set about removing all the background artifacts, resulting in the revised model shown in the screenshot below:

Model in Microsoft 3D Builder, after removal of background artifacts

If you are doing the sorts of 3D modelling projects involving lots of photos and 50+ MByte object files, I highly recommend Microsoft 3D builder; it seems to be one of those little-known unappreciated gems in the Microsoft ecosystem; using 3D Builder was like expecting a tricycle and actually getting a 12,000HP supercar; 3D Builder not only accommodated my 57+ MByte  .OBJ file, it didn’t even seem to be breathing hard;  more like “Yawn – is that all you’ve got?”

After removing all the background artifacts, I exported the model from 3D Builder as a .3MF file that I was delighted to see is compatible with Prusa’s Slic3r PE, as shown below

The .3MF file from 3D Builder imported into Slic3r PE

I fired up my Prusa MK3 printer and printed out the model, and got the following off the printer

Very small scale version of my 3D head model. 0.5mm mechanical pencil provided for scale

Then I scaled the model up a bit and reprinted it, getting the following model:

 

Once I was convinced the model was reasonably accurate, I set out to print a full-sized model.  To get the proper scale multiplier, I measured the distance between the outer rims of my eye sockets and compared this to the measurement on my mid-scale model. This gave me a scale factor of almost exactly 3.5, so I used this to print the full-scale model. The full scale model just barely fit on my Prusa MK3/S print bed, and took an astounding 24 hours to print!  Also, this is the only model I’ve ever printed that actually cost a non-trivial amount of money –

Full scale print setup. Note the print time of almost 24 hours, and the used filament – over 100 m/$8 in cost – wow!!

Partially finished model, showing the internal structure (5% fill)

Finished print

With the finished 3D model, it should be possible to create the desired custom face mask directly, without having to take a plaster cast impression of my face.  However,  to verify that the full scale model was in fact a faithful representation of my face/nose structure, I decided to make a plaster cast of the printed model, and then compare the plaster cast to my actual face.  This is sort of the backwards process used by a prosthetics house to create a custom face mask; they make a plaster cast using the patient’s face, and then use the plaster cast as the model for the final product.

Plaster cast impression using the 3D printed model instead of my face

Plaster cast separated from the 3D model

Side view of plaster cast on my face, showing that the 3D model is an accurate representation.

Summary:

All in all, this project was a blast; I was able to create an accurate 3D model of my face, which should be usable for the purpose of creating a custom face mask for me so I can go back to abusing my body playing basketball.  However, I have to say that if I added up all the time and effort required to take all the photos, deal with Meshroom’s idiosyncrasies, actually print the full-scale model (24 hours, $8), and still have to take the plaster cast impression to verify the model, I might have been better off to just get the plaster cast impression made by a professional.  OTOH, I learned a lot and had loads of fun, so…

Stay tuned!

Frank

 

 

 

 

Alzheimer’s Light Strobe Therapy Project

Posted 24 March, 2019

A friend told me about a recent medical study at MIT where lab mice (genetically engineered to form amyloid plaques in their brains to emulate a syndrome commonly associated with Alzheimer’s) were subjected a 40Hz strobe light several hours per day.  After repeated exposures, the mice showed significantly reduced plaque density in their brains, leading the researchers to speculate that ‘light strobe therapy’ might be an effective treatment for Alzheimer’s in humans.

The friend’s spouse has been diagnosed with Alzheimer’s, so naturally he was keen to try this with his spouse, and asked me if I knew anything about strobe lights and strobe timing, etc.  I told him I could probably come up with something fairly quickly, and so I started a project to design and fabricate a light strobe therapy box.

The project involves a 3D printed housing and 9V battery clip, along with a white LED and a Sparkfun Pro Micro 5V/16MHz microcontroller, as shown in the following schematic.

Strobe Therapy schematic

I had a reflector hanging around from another project, so I used it just as much for the aesthetics as for functionality, and I designed and printed up a 2-part cylindrical housing. I also downloaded and printed a 9V battery clip to hold the battery, as shown in the following photos

Finished Strobe Therapy Unit

Internal parts arrangement

Closeup showing Sparkfun Pro Micro microcontroller

The program to generate the 40Hz strobe pulses is simplicity itself.  I used the Arduino ‘elapsedMillis’ library for more accurate frequency tuning, but ‘delay()’ would probably be close enough as well.

 

I’m not sure if this will do any good, but I was happy to help someone whose loved one is suffering from this cruel disease.

Frank