Skip to content

Archive for April, 2016



I made a clock for the granddaughter a couple of years back and it’s been working fine until a couple weeks ago. The display was fairly small TFT 1.8″, consisted of a DS3231 clock module, power converter, and Arduino.

Each day the display would come up with a witty quote that was read from a SD card. Optionally there was a monthly calendar that would toggle on/off if a button was pressed.

IMG 4981

Keira mentioned that although she used to like a “night light” in her room at night, now she didn’t want any light at all. I tried, unsuccessfully, to modify the LED backlighting on the TFT to get it to turn on and off. She just laid the clock face down at night to stop the light as a stop gap.

When I got the clock back from her, the DS3231 backup battery was dead, and the chip itself was generating random times and dates. Which were probably correct somewhere in the world, but not in our locale.

Hence, time to build version 2. The first order was to find out what Keira liked and disliked about the clock. The first like was the calendar portion. The strong dislike was the TFT backlighting.

Time to assemble the parts. An Arduino MEGA, breakout board for the TFT display and a high resolution 480 x 320 3.2″ display. To make darn sure the time was going to be accurate, I added a Skylab GPS module.

IMG 1331

The reason for the 3.2″ high res TFT was that Keira is a cat lover. Lynx in particular. My original idea was to have a background picture, and draw the time/date over top of the display. What I soon found is there is no functional library that does this sort of blending of text on graphics. Granted I could write one, but I decided to keep it simple.

I soldered wires directly on to the GPS module, sandwiched the boards together and presto, a relatively compact clock.

IMG 1333

Rather than add a bunch of push button switches, I used a couple of 3mm round head screws with the ADC touch library. I soldered two wires on the Mega’s A0 and A1 pins to use the library. If you noticed the display I am using also sports a touch screen but I didn’t like the idea of finger prints on the display so I didn’t use it.

The display is a Cold Tears 3.2″ TFT 480 x 320 display. Very nice colour and, when I looked at the back of it, there is a control for using PWM (pulse width modulation) to control the brightness of the display. What is not clear is how you make it do this…

In reality you need to unsolder one jumper on the TFT PCB and then on the adapter board PCB you need to add a short jumper wire because they never put a trace/jumper on the board to do it. Once done, pin 44 on the Mega with a PWM signal will control the brightness of the TFT.

Next up it was off to the internet to find some cat pictures that would scale nicely for the display. I had previously written some software that would open a graphic, scale it for a TFT display and then save it in RAW format so the UTFT library would read the raw file and display it. The graphic size that scaled best was 600 x 400 so that’s what I used as the exact size in the search.

Scaled with the date/time showing:

IMG 1334

I purposely looked for images that didn’t have something on the bottom so the date/time wouldn’t overlay the important part of the picture.

With the backlighting on the TFT, it wasn’t until I built my own Tardis clock for our bedroom that I fully understood just how bright even a dimmed LED can be. Let’s just say I completely agree with Keira on the lighting.

For a case to hold the clock, well, it helps to have a 3D printer. I designed a case in Tinkercad that would hold the clock, and then a separate base to stop the clock from tipping over. You can see the two “touch” screws at the top right corner of the case. Those are used to kill the backlight on the TFT and to toggle between the cat/time/date and monthly calendar views.

IMG 1351

There are 100 images on the SD card mounted on the back of the TFT display and each day at 2am, the clock will pick another photo at random. Since the clock is GPS smart, it also nows daylight savings time from regular time.

While Keira can turn the TFT backlight on when she goes to sleep, I have it coded so that at 7:30am in the morning, it will turn itself back on for the day. Or you can turn it on manually before that time.

For anyone who might have noticed that the dim image in the last photo looks like a golf club, well, I AM an avoid golfer, I do things a little different. But I did keep with the “cat” theme….look at the brand name and logo on the club head…

IMG 1352


3D Printer Nozzles

With just under 1300 hours on Bullwinkle (my first Kossel) and about 20 hours on Rocky (my second Kossel) you’d probably think that I’ve learned all there is to know about…well…3D printing.

In actual fact, the longer I am in this hobby; if one can call using a heated inverted tripod to drool PLA around a heated bed, the less I realize that I actually know. Oh sure there’s some obvious stuff I have learned, but sometimes the things that I’ve never paid any attention to are rather important.

As the topic suggests, I’m going to write about nozzles. I have ordered more than just a few off eBay because I was told that they wear out. What was failed to be mentioned was at what point they actually wear out. With as many hours on them as I have, I would have thought they would be wore out completely now. Yeah, not so much.

Thus I have not managed to wear one out yet. I have managed to plug them though. Using a small butane torch to blast the filament inside to ashes works fine to recover the nozzle. Although I have to admit for the price of the butane torch, you can buy a few replacement nozzles.

For my Kossels, I have used MK8 styled nozzles. I changed one on Bullwinkle at about the 700 hour mark, so 500 hours ago and it’s still printing just fine.

A month ago I built Rocky. Overall a better build than the first printer largely because I know what to mind for when building one now. None the less, with the same firmware settings, I was having some issues getting the filament to flow properly. When I got the extrusion right, the extruder would skip (it clicks when it skips from the bowden tube back pressure).

Didn’t make a lick of sense to me because both printers are identical builds for the hot ends, bowden and extruder. Firmware steps were also the same and when measured, ask for 100mm of filament, I got 99.91mm on average.

Cranking up the tension on the extruder resulted in the Bowden tube pulling the top pneumatic fitting right out of the extruder housing. So yeah, it takes a whack of force to do that, but I dood it…

After re-printing the extruder, did I mention it’s handy to have two of these so on can fix the other? No, not what my wife thinks either…:-) Any way, with the new extruder I started to look at why the hot end wasn’t working right.

I already know that the temperature thermistors on hot end can vary, sometimes a considerable amount (+- 8C). But that did’t seem to be the problem when I measured it with my Fluke temp probe.

Nope it was something else…

So I thought, what the heck, I’ll change the nozzle. And low and behold it suddenly worked perfect. If you’re thinking that the old nozzle might not have been the right size, I have a .4 drill bit that I run through all my nozzles before pressing them into service. Hence, it was the correct size.

None the less, the first layer pattern looked completely different. As in GOOD different compared to the old nozzle. So I checked the type of nozzle I put in, compared it to my stock of nozzles and realized, by shear luck I’d snagged a great nozzle.

Here’s my line up of nozzles that I have in my parts trays:

IMG 1347

Kind of hard to believe those are all MK8 nozzles. But they are. All .4mm. The three nozzles on the left side of that picture are 7mm ones. The two on the right are 6mm ones and the type that came in my Kossel kit.

It’s quite easy to see that for as far as mass goes, the nozzle on the extreme left in the picture has the most mass. Which means it’s going to hold its heat quite well and it’s going to transfer that heat to the filament as well.

Looking down at the tops:

Nozzle Top

Again, easy to see where all the mass is. In fact, so much mass that they grind off the top to keep it at the 12mm overall height.

So back to my nozzle that works so well, it’s the one on the left. Not only does it have large mass, but there’s another little surprise when you look at the inlet side of the nozzles.

IMG 1348

That’s standard 1.75mm filament. The four nozzles on the right vary in the inlet size from 1.84 to 1.91mm. So 1.75mm filament fits in with a bit of wiggle room.

However, the nozzle that works the best for extruding the proper amount is again, that extreme left one. The inlet size on it is 2.25mm. The filament has LOTS of room.

IMG 1349

While it looks like an optical illusion, those right nozzle inlets are darn near the same size as I mentioned, but that left one, huge. Like a cave.

What I think is happening is that with the larger mass, and bigger “reservoir” the heat is transferred better to the filament and it flows out very nicely. Whereas the smaller ones don’t. In my experience any way.

I’ve noticed that some of the eBay sellers have started to put inlet sizes on the nozzles they sell. Today I spotted one that had 2.5mm inlet. Would it better as good or better? No idea.

If you’ve been experiencing extruder problems, i.e. clicking, under extruding, have a look at that nozzle. Might be worth ordered a couple of the .4mm nozzles but with more mass and a bigger inlet. If you want to print faster, this might also go a long way to allow you to do that. All one can do is experiment.

One note of caution, if you do change the nozzle, remember you might have to re-level the hot end. If you can find the same 12mm length nozzle though it’s not going to be much of a change and the benefits could out weigh the time it takes to re-level.