Building a gigantic clock along the side of my basement stairs.

This is less a quick print than a huge project I just started.

Basically I want to build something like this:

https://www.diymachines.co.uk/how-to…elf-edge-clock

I like the idea and I have a perfect dead space to use for a wall mount clock shelf, but there are several issues I want to address in my build.

  1. I want the time obtained from the internet. This projects uses a fixed start time, which gets annoying when you have a power outage an need to flash the firmware or wait to a specific time to plug it in.
  2. It uses an Arduino. I want to use an ESP32 chip, as it comes with enough IO pins and wifi to connect to my IoT network.
  3. This is an all American clock. I want a proper 24 hour display, which makes the clock even bigger.
  4. I want blinking dots between the digits. e.g. “19:45” (and it gets bigger again)
  5. Depending how well this project goes, I want to add seconds. “19:45:23” (to make it gigantic) This will be achieved by expanding the board by one meter on the left side and continue the construction their along with a new firmware of course to address the change.

The full clock will be 2800x600mm.

[ATTACH=JSON]{“alt”:“Click image for larger version Name: GiantClock_1_TheTarget.jpg Views: 0 Size: 275.3 KB ID: 6363”,“data-align”:“none”,“data-attachmentid”:“6363”,“data-size”:“full”,“title”:“GiantClock_1_TheTarget.jpg”}[/ATTACH]

This almost empty wall will be the clocks final location. Beside the door bell on the top left and the three images the wall is just white, which just sucks. :smiley:

I will add more entries to this first post while the project is progressing and I hope it will due to the now existing public pressure. :smiley:

During the last week I updated and tweaked my printers. Especially the old ANet A8, as it is, beside the GeitPrinter, the one with the biggest build plate and the GeitPinter is still not 100% finished. However, the old firmware had the 0/0 point of the print area outside of the build plate, which made it impossible to get a bigger print done and positioned right. This was the reason I updated the firmware in the first place.

[ATTACH=JSON]{“alt”:“Click image for larger version Name: GiantClock_2_TheTemplate.jpg Views: 0 Size: 277.9 KB ID: 6364”,“data-align”:“none”,“data-attachmentid”:“6364”,“data-size”:“full”,“title”:“GiantClock_2_TheTemplate.jpg”}[/ATTACH]

As you can see it now fits. Don not get confused. This is not one digit of the clock. This is the template for 1/2 of a 7 segment, which will be acting as a template to layout the drill holes for all clock elements properly. So the more or less full build plate is just the upper circle of the “8”

Many parts are printed in ABS, because my ANet A8 in its enclosure is my ABS printer. ABS is also cheaper and a lot more robust, when it comes to impact, so I plan to do the inner parts as ABS and the outer hull in PLA. Printing will take ages. This small ABS template already took 5 hours. Three inner pieces printed as ABS with 50% infill, so together they take 14 hours. If you only count the segments of the clock it will take a while to print, but I will also need to print the other parts in between I need around forty of them and the same number of shelf walls and surfaces.

This table contains the current process of the printing process: [TABLE=“align: left, border: 1, cellpadding: 1, width: 500”]
[TR]
[TD]Done[/TD]
[TD]Required[/TD]
[TD]Part Name[/TD]
[TD]Filament Type[/TD]
[TD]Printing Time[/TD]
[/TR]
[TR]
[TD]1[/TD]
[TD]1[/TD]
[TD]Drill Template[/TD]
[TD]ABS[/TD]
[TD]5 hours[/TD]
[/TR]
[TR]
[TD]51[/TD]
[TD]66[/TD]
[TD]Core Element[/TD]
[TD]ABS[/TD]
[TD]14 hours/3 pcs[/TD]
[/TR]
[TR]
[TD]1[/TD]
[TD]1[/TD]
[TD]Core Element (electronics)[/TD]
[TD]ABS[/TD]
[TD]4.5 hours[/TD]
[/TR]
[TR]
[TD]67[/TD]
[TD]67[/TD]
[TD]LED Mounting bracket[/TD]
[TD]ABS[/TD]
[TD]5 hours/8 pcs[/TD]
[/TR]
[TR]
[TD]7[/TD]
[TD]16[/TD]
[TD]Sleeve 1 (outer frame, bottom,left,right)[/TD]
[TD]PLA[/TD]
[TD]6.5 hours[/TD]
[/TR]
[TR]
[TD]8[/TD]
[TD]13[/TD]
[TD]Sleeve 2 (2 cbl. / light, top, horizontal)[/TD]
[TD]PLA[/TD]
[TD]6.5 hours[/TD]
[/TR]
[TR]
[TD]12[/TD]
[TD]24[/TD]
[TD]Sleeve 3 (4 cbl, inner vertical)[/TD]
[TD]PLA[/TD]
[TD]6.5 hours[/TD]
[/TR]
[TR]
[TD]6[/TD]
[TD]13[/TD]
[TD]Sleeve 4 (4 cbl. / light, middle horizontal)[/TD]
[TD]PLA[/TD]
[TD]6.5 hours[/TD]
[/TR]
[TR]
[TD]1[/TD]
[TD]1[/TD]
[TD]Sleeve 5 (electronics, top left)[/TD]
[TD]PLA[/TD]
[TD]6.5 hours[/TD]
[/TR]
[TR]
[TD] [/TD]
[TD] [/TD]
[TD] [/TD]
[TD] [/TD]
[TD] [/TD]
[/TR]
[TR]
[TD] [/TD]
[TD] [/TD]
[TD] [/TD]
[TD] [/TD]
[TD] [/TD]
[/TR]
[/TABLE]
[ATTACH=JSON]{“alt”:“Click image for larger version Name: GiantClockStatus.jpg Views: 10 Size: 326.8 KB ID: 6807”,“data-align”:“none”,“data-attachmentid”:“6807”,“data-size”:“full”,“title”:“GiantClockStatus.jpg”}[/ATTACH]

Very ambitious! I suspect parts storage & organization will be an important aspect of this project, as it looks like it will take months.

Currently I am fighting the printer calibration. The print above failed due to the small “handle” section got bend. I fixed the part to be bigger and the second part printed well.

However, I need to recheck the calibration as it seems the 500mm calibration distance I used where not enough. In the end the part is too small. I also need to do the same on the other printer of the shell or the shelf will not fit over the studs.

I need a way to calibrate the axis movement over the entire distance of the print bed to get the best result.

Using two printers would half the print time and I could print 6 segments and matching studs a day.

I guess you must be printing at a fairly high speed.

55mm/s, so a little slower than the usual defaults.

People print three or four times faster these days, but I never see speed as an usable option. I want the print to finish properly and look good. It makes no sense to increase the print speed of a 10 hour print and after 3 hours of 5 it fails for some reason. It is the same with computers. When they do their job, it is fine. I don´t need to wait in front of my system and wait for the result. You don´t save time when a printer produces trash every few prints. :smiley:

The third and hopefully last template version is now printing and nearly done. After that I start printing the studs. 10 hours one print with 3 of them. The new placement algorithm they implemented into Cura 4.8 together with the community works great. No gigantic spaces which gets filled by useless brim. All parts next to each other. I have seen so much worst stuff over the last few years.

I added a table to keep track of the printed parts and added an ugly picture showing the so far printed elements for the clock.

Editing in this forum still sucks, so I add more stuff here at the bottom. I also will not resize images anymore, as the forum removes the sizing every time. So my prescaled 1024x768 it is. Also why is it adding images at the top and not at the cursor position.

Due to my hardware troubles with the ANet A8 (x belt suddenly had a strange turn) I had some failed prints of the core elements. Since I wasn´t satisfied with the small (to fit on a 210x210 prinbed) template included in the original clock project and I already had the failed brackets I designed spacers:

[ATTACH=JSON]{“title”:“SegmentSpacers_1.jpg”,“data-align”:“none”,“data-size”:“full”,“data-attachmentid”:6832}[/ATTACH]

Together with the cleaned and striped down base plate core mount print fails it looks like this:

[ATTACH=JSON]{“title”:“SegmentSpacers_2.jpg”,“data-align”:“none”,“data-size”:“full”,“data-attachmentid”:6833}[/ATTACH]

As I printed the core elements in ABS, I printed the spacers in ABS as well, so I can weld them together using acetone.

Damn the clock will be huge :smiley:

Just to get some perspective of the size. This is one element. 1 segment wide and 2 segments in height. The final clock will have 13 segments horizontal. :smiley:

[ATTACH=JSON]{“title”:“SegmentSize.jpg”,“data-align”:“none”,“data-size”:“full”,“data-attachmentid”:6834}[/ATTACH]

I was thinking, from your previous descriptions, that it would be massive, and it sure looks like it will be.

After “some” time of printing the basic frame structure is finished and the first 1809 LEDs are mounted.

1809? well, there are 67 segments with 9 RGB LEDs installed right now, so 6793 LEDs.

The shelf mount lights are still missing, but I worked 6 hours today to get the main lights done and I am done, too :smiley:
[ATTACH=JSON]{“title”:“GiantClock_Assembly_2.jpg”,“data-align”:“none”,“data-size”:“full”,“data-attachmentid”:7105}[/ATTACH]

Now some programming, to get the matrix table data right, as currently it only shows garbage, when in non test mode. :smiley:

Wow, you do awesome ambitious projects!

How many kgs of filament will this clock require?

I think about 4 Kg PLA and 2Kg ABS.

Finished the basic programming today and it even shows the time, it automatically gets from the internet now. :smiley:

There is also a web server, which allows to change colors, display and animation modes. I also implemented some test modes.

After a weekend of soldering and adding around 20m of wires, the basic installation is now done and my back is thankful for that. The constant bending over the table to solder stuff in was tough for my back. Last step was to add the 26 shelf lights, which light the small boxes and their future content.

I added shelf light control to the web interface, so I can now configure those, too. As you can see the majority of covers are still missing. This is because of my TronXY X5 being lazy during the last week. He is now back to 24/7 duty and should deliver around 5 covers a day.

[ATTACH=JSON]{“data-align”:“none”,“data-size”:“medium”,“data-attachmentid”:7156}[/ATTACH] [ATTACH=JSON]{“data-align”:“none”,“data-size”:“medium”,“data-attachmentid”:7157}[/ATTACH]

The wooden plate lacks some side coverage, but unfortunately the home depots are closed till January at least. I also need to sort out some minor element alignment issues mainly on the right side. I have no clue how that happens, but I need to fix this somehow with the wiring already installed.

The ESP32 developer board is dangling on the outside of the electronics housing, too. I still want to add some sound module and a light sensor, so the time is readable at day and not flashing people at night. For now I use the time to dim the light, which (of course) isn´t optimal in all situations.

At last I want to finish the basic programming. It is working quite well so far and I have some fancy modes, where the dots are bouncing and the numbers are scrolling up or down, but I want more effects to select from. :smiley: Maybe a random mode for the shelfs, so every hour a random set get lit.

Fantastic! Thanks for the updates.

I’m impressed!

I’m impressed too. It is interesting to put electronics into the 3d printer world. That’s going to be an expensive clock by the time you are finished.

It is quite okisch and still cheaper than getting some real and boring furniture to fill up the empty wall. Filament is quiet cheap. I payed around 12 Euro per ABS and 14 per PLA spool. No p&p as the shop is around the corner. Processor and LEDs are peanuts and the board itself was 35 Euro. The wires are all from my private dumpster in the basement, where I keep all obsolete cables for a later usage. So they got a second live.

Most work was the soldering. It took me a few days to solder all the LEDs and hot glue the contacts afterwards, as I want to avoid as much maintenance as possible, because the clock will be mounted more or less out of reach. There is no way I can put a ladder next to the wall to fix some wires later on.

Most wires came from an old 10m VGA cable and the rest from a broken CAT5 network cable.

Fun part: When taking apart the CAT5 cable, I found out, why it was so bad. The “made in china” text beside the CAT5 on the cable should have been an indication, but when opening there was basically just wires and plastic insulation. There was absolutely no shielding. Not even the foil you find in all cables. Also no mesh around the wires. Nothing. Just single copper wires twisted around each other. I now have an idea why this cable was so bad behaving in my network. :slight_smile:

I rate it as an interactive multimedia art installation. So it is priceless anyway. :smiley:

The electronics are not even complicated. All elements are connected by the same three wires in series, which end up on the power supply and the processor board. Beside these components I just added one resistor to avoid the processor burning out, when power from the LEDs gets cut (broken cable or soldering).

Programming is a little more complicated as you need to know at least basic C and C++ programming, unless you want to use some ready to use code, which I didn´t. :smiley:

Geit, I have a small correction for you. I was involved with some of the very 1st twisted pair Ethernet development, as a tester of SynOptics Corporation’s LattisNet product, in the mid 80s. LattisNet was the very 1st Ethernet cabling system that didn’t rely on COAX cable.

The design specifically forbade the use of shielding in the cable. Shielded cables performed much worse than unshielded cables. Twisted pair Ethernet is designed to accept external interference and filter it out at the endpoints. The idea is that, as long as the cable is unshielded, the outside interference will be induced equally in the individual wires. This makes it relatively simple to create a filter that ignores all the signal that’s common to the wires, and only keep signal that’s different. It’s the prime reason why the cables have to be twisted in pairs, each pair with its own specific number of twists per inch.

Since then, development of the system has led to cables, such as CAT6, that do incorporate shielding. However, CAT5e still does not have shielding, and it’s not supposed to have any. I’m not sure how much the technology has changed from the early days but, given that CAT6 requires a cross-shaped divider to keep individual twisted pairs apart from each other, I can only assume that filtering of interference is still important.

It’s possible your cable’s poor performance isn’t because it lacks shielding. It might be that that manufacturer didn’t organize the wire pairs properly or didn’t use the correct number of twists per inch. Perhaps the copper itself is not the proper alloy. Is it possible the cable is longer than 100 meters?

Ok, the last elements are about to be printed.

I created a thingiverse thingy as I (of course) over enginiered the entire thing adding even more parts to it, but I think the new black components make a huge differences and they solve some minor issues I had with the original design.

Well done my friend. It’s a huge project & you’ve brought it home very well.

Ambitious project. Well done.