[00:45]

While I've spent most of today computing my options for the #T^2 TyMist [gd0138], I started to have concern for the methods of controlling the OLED for this project too. I found these relevant projects:

• https://hackaday.io/project/364-mipi-dsi-display-shieldhdmi-adapter
• https://github.com/ylj2000/HDMI_To_MIPI
• https://github.com/peng-zhihui/HDMI-PI

I also don't think I'm going to be able to squeeze and fold over the ready-to-go driver.

The board uses the TC358870XBG Toshiba chip, but it's a dense BGA so manufacture could get a bit difficult.

Remembering that the CM4 supports DSI where the CB1 doesn't, I went to investigate futher:

Using this calculator, the bandwidth needed for 90Hz is between 2.8 and 3Gbps, so it should be a case of wiring things up and writing some drivers... not like I've ever written drivers (and in Linux no less) before. Speaking of Linux, I have been looking around to see how involved it could be to spin up -- but more importantly, maintain -- a custom distro, but this distro's changelog seems managable and this distro was apparently made as a side project, so maybe I'll manage. I've been interested in Powerpoint OSes since about 10 years ago, so I've had a feeling that there would be one reason or another that Future Me would make an actual OS [skin / distro] one day. Anyway, I might want to read up on content like this log to see what goes into driver dev.

I'd have to start looking into alternatives to the CM4 if I don't want to pay inflated prices for one. Last time I checked, I heard that the backlog for stock was going to be sometime mid 2023 soooo not sure about that one. Considering that I'm likely going to have the Timespy as unofficial #Teti [gd0022] and #T^2 TyMist [gd0138] controllers, it might make sense to have a basic edition that uses an SPI/I2C display (and perhaps runs off #Tetrinsic [gd0041] / the M032 series microcontroller). I'm still eyeing that 2.7" Sharp display, even if it's more expensive than I'd like:

[20:15]

Attempt7.X

I think the latest was 7.3, but the adjustments have been slight since Attempt7 so I'm just going to bundle them all in this paragraph. 

The changes was that I changed the strap location to be in the middle of these Tetrinsics. I also tried the thumb at a 12 degree angle and didn't feel like I needed to change anything. Additionally, I replaced the somewhat arbritrary 8.X degree main draft with a more concrete 3-point plane.

A few days ago, I had also discoved this strap and bought 2 of them.

What I was doing was looking for a silicone strap and found one that looked like the white one I have now and one that had more holes and a dual colour design:

Then I remembered what I said, which was that the strap is the main thing in contact with my arm. Thus, the ideal solution would be a strap that was breathable. 

However, in the corner of my eye, I also saw this silicone strap with a clasp design:

I thought, further motivated because of how large Timespy is, it'll be a nice feature if I didn't have to carefully button through the correct hole and then neatly tuck away the extra strap length every time I put it on. This also reminded me that more strap materials exist than just silicone.

First, I did some more research on watch straps and saw:

and I think "Manne I got to choose wisely because the strap can also make or break the entire look!".

I widen my search results and soon find this strap:

I loved the look of this, would be breathable and looks really flexible, so I start searching specifically for mesh watch bands.

BAM! I see this and immediatly exclaim "The future is WOW thanks to SCIENCE!!". Infinitely adjustable! Completely breathable! Cheap and looks amazing! 

I start watching YouTube videos to see if this is legit (seemed so) and then read both positive and negative reviews on Amazon. Everyone that commented on the magnet strength said that it was strong enough for the job (which was my main concern, can confirm) and the only review that gave me concern was one talking about the poor paint durability of the black version (lightly grazed a pair of scissors over one of the lug holders whilst trying to optimise my way of installing them and the paint discoloured, so I unfortunately can also confirm). After hours of deciding, I decided to go with the black version. 

I'm going to play it safe and use dark grey or black as the accent colour IRL (the light green parts in the renders) because I don't want to strart tracking down a light green PETG (or any FDM filament with the #SecSavr Suspense [gd0105] in view).

This is what Attempt7.3(?) looks like:

The strap is thin enough to easily slip slightly underneath my jumper sleeve.

Attempt8

The only thing that was annoying me and the only thing I changed was the position of the strap. Basically, Attempt7.3 constantly was constantly squashing some bone bump where my forearm bone ends and wrist starts. In Attempt8, I reworked some faces and aligned the strap to Tetrinsic Finger2:

Printed it, put it on and my reaction was, essentially:

It's the moment when everything finally aligns ergonomically. It doesn't feel too tight and loose at the same time or slides around too much or slides over the wrist.

It's the first time I'm just mirroring the STL in Cura and reprinting, with 0 changes required!

I printed Attempt5 and, whilst that sharp edge adjacent to Finger5 got in the way, it was determined that a solution... may exist. 

🎊🎊🎊

Attempt6

I noticed that a more acute angle may improve screen visibility, so I tried an 84 degree angle:

As you can also see, stylistic choices like a section of logo bracket and filleted chamfers have made an appearance to obtain a solution that didn't look incredibly disjointed or overly sharp.

I printed that and my takeaways were:

• Finger5 does not fail the solution.
  • wohoo
• The 84 degree screen angle fails the solution.
  • sigh
  • The whole thing just feels crampt and the bottom of the screen started colliding with things like my legs or work surface.
• The ergonomics of Finger5 (and the other ones, to a lesser extent) is dependant on the arm's angle.
  • So I was onto something with that screen angle idea, at least. Unfortunately, moving that angle to the Tetrinsics is unlikely to work, since the ends closest to the screen would increase in Z height (relative to top of arm).
• The draft angle when used on a work surface could be increased to give Finger5 more space, and there needs to be more worksurface support at the end corner.

Attempt7

Attempt6 showed promise, so I:

• Increased the main draft angle from 8 to 9 degrees (which increases the longest length to 95.7mm)/
• Returned to a 90 degree screen angle.
• Moved the "bracket bend" so that it was 2mm away from the edge of the milky-green section (2mm is the offset from the white L to the milky-green section).
• Move the strap back to centered-to-screen.

Attempt 7.1

I printed 7 overnight and it turns out Fusion360 threw one of the strap holes somewhere, so I couldn't mount the strap.

The increased draft angle worked, but I thought that the angled face that the thumb pressed on was too high, so I reduced it such that both sides looked more symmetrical.

I accidentally tried Attempt5 on the wrong arm, and it felt more comfortable than Attempt4 and its centered-to-sceren strap location, so I'm now going to try centered-to-back:

I've also increased the height of where the front strap starts.

I'm also going to try keeping this face parallel to the belt so that the top face doesn't look all misaligned. If you look at the Attempt6 prerender, you'd see that it angles away.

[20 minutes later, whilst setting up the render]

Just did an offset on the face to fix this fillet that has been annoying me. Restarting the print now.

Comparing Concept1 Attempt[latest] to Concept2 Attempt7.1 

Shakes head in disgust and disappointment.

All the sacrifices I made, all because I said / thought:

• 60x60mm height and width MAX.
• Finger5 would just get in the way, complicate the design and thus should be avoided.

Both designs have a 9mm thickness behind the screen, similar thicknesses under the Tetrinsics (20 vs 20.4mm) and similar max lengths (91.7 vs 93.9mm), yet Attempt7.1 looks sleeker, thanks to its more modern looking design. The fact that A7.1 is 80mm at its shortest length also helps. Because of the strap, even the 70x70mm Attempt7.1 doesn't fully cover the top of my wrist, so Concept1 likely would've had issues with thumb reachability anyway.

Unlike Concept1, C2 A7.1 can be taken off and used on a flat surface. It's also got a larger, higher PPI and refresh rate screen with thinner bezels now. 

Admittedly, this entire run was sponsored by #Tetent TestCut. Thank you Me In The Past 😭!

Attempt7.1 renders

Whilst using Attempt4 for a bit, I noticed that Finger5 might be able to slide across the same surface as the other fingers. Theoretically, the finger responsible for [space] and other things would only have to move +/- 2mm for most typing, and printed concepts of the #Tetent TestCut have shown that the thumb could comfortably move well enough to share the same layout as Fingers 2-4. The same can't really be said for Finger5 at the moment, and the reason why I want to try and get a solution that includes Finger5.

This is the next solution I've come up with. I've changed the main draft angle so that it's like the Testcut. The design is 80mm at the shortest section and 94mm at the longest, with the L thickness being 9mm. I'm waiting for it to print right now, and I hope this one actually works because then I'd be able to use them as a split, slightly smaller TestCut:

That specific cut around Finger5 is so that said finger doesn't hit the sharp edge of the white section.

In other news, it's quite nice to go all the way back to the first Timespy project log and read that I partially started the project because of the 90Hz screen I found, reject it and then accept it months later because of a design aesthetic change.

This is the shortest version yet. I was hoping that Finger5 would be able to reach around the side. I printed a mirrored version to see how it'll feel on my right arm. I also wore Attempt3 whilst out shopping and it was fine other than the length (addressed in Attempt4) and that even my coat sleeve shrinks at the end for some reason. I've never really noticed it until now.

Unfortunately, Finger5 still fails the solution. The smaller length is nice though, but I think I've moved the strap a tad too far from the wrist so I'll nudge it back in Attempt5.

Attempt2 Print

I'm going to save my Yousu Galaxy Blue for end-use prints since it's one of the few fancy colours I still have in stock and I don't really want to buy more filament because I want to eventually switch over to TeamResin with the #SecSavr Suspense [gd0105].

It was all excited when the bottom strap loop actually worked, but I didn't have enough clearances for the top strap so it never clicked in, unfortunately. I've now made sure that there's a 2mm offset clearance from the 1.5mm pinhole (so 2.75mm clearance total). 

Having Fingers 2-4 on the same level is perfectly fine, surprisingly. Finger5 is still what's failing the solution.

Attempt3 Model

My next idea was this. I put my scale against a wall and measured how much force Finger5 could push outwards, and it can do 200g without much effort! Unfortunately, this solution increases the overal length to an eye watering 108mm. The white edge also increased 8 -> 9mm.

Attempt3 Print

Finger5 still fails. On the plus side, the strap succeeds!

So I wore it overnight and here are my thoughts and observations:

• The strap needs to be more centred.
• The design only needs a bit of persuasion to get through the end part of a sleeve. After that, it can slide though the main length without impedement.
• The cuboid design looks fine on my arm. Larger than a watch? Absolutely, but since the sleeve neatly slides under it, it's seems to blend in. (White gloves would take up about the same visual area.)
• That steep angle needs to go.
• It doesn't seem that my skin had trouble breathing. I'll stick to 20mm bands though, since the band is the main thing touching my wrist.

For starters, I forgot to change the height of the screen from 67 to 68mm, so the bezels look ever so slightly thicker. I've reduced the thickness to 20.5mm, meaning that all Tetrinsics are at the same height. They've all been given a slight angle, other than Finger3. Finger2 has been moved as far towards-wrist-ways as possible.

For the Finger5 cutout, I think I've obtained a good looking solution that has that filleted chamfer aesthetic.

I've then added the holes for the 20mm strap. I've modelled in a 20.5mm spacing, 1.5mm hole diameter and 2mm hole depth. The strap position is such that the edge closer to the camera is halfway across the front screen. This is so that sleeves can neatly slide under the other half.

Finally, the length has been reduced from 103mm to 95mm, which is about the same as Concept1. The design would probably be fine without the angled bezel now, but I'm going to keep it since it's still a Tetent at the end of the day and I may decide to unstrap and use it on a desk as a portable trackpoint+keyboard. In this mode, the thumb is used to keep the Timespy in place. 

Hmmm that sounds like a good idea actually... I've got another CB1 that isn't planned for anything... and that sweet 90Hz OLED is only £30+VAT... and I might look better wearing 2 instead of 1 because of symmetry... and I can have a Tetent input method always on me...

I like that the screen looks large, expansive and a square-ish aspect ratio, which could be attributed to the thin bezels. It's also a good thing that I opted for this shorter design than the 121mm long one; this design doesn't end too far down my forearm and I think it should be possible to use a single watch strap. I'm actually quite a fan of the entire look from all angles, but I might make that 40 degree slope next to the wrist the same angle as the one closest to the camera (see below).

The issue is the locations of the Tetrinsics. My fingers just have much larger angles than expected. I'm going to have to do some rework. It also seems that 23mm is still too high. I'm going to try for 20 or 21mm again and see what I can get.

I'm also going to add a place for the straps. The i3 AMOLED has convinient little levers that make removing them super easy, barely an inconvinience.

I'll have a look on aliexpress to see if there are any bands like this. This one says "20mm" on it, so I hope that it implies that wider straps exist.

searches on aliexpress

Well that didn't take long at all, and there's even a variety of colours to choose from!

As I was modelling the new concept, I came across these:

They both look quite similar to the basic Timespy design, just thicker. These devices are over 10 years old. It seems that Timespy is the modern day equivalent of these wearable PCs. Thinner bezels, slightly larger, 90Hz screen (3.5 vs 3.8), a (theoretically) fast keyboard instead of a numpad...

This change has come about due to the Tetent TestCut concept.

I started with a 5" 720P touchscreen I found from the same AliExpress seller as the current 3.5" one. 

Dimensions went from 96 -> 121mm and 62.x -> 67.6mm.

For some reason, the bezels seem too slim, which is a nice problem to have. I also had the below idea to make Timespy look more slender than it actually is:

I had some worries that 121mm was just going to be too long, and I also liked the idea of a 90Hz OLED.

The screen only increased the height/width from 67.6 -> 68.6mm, but was quite a bit shorter and now the battery and expected Tetent layout were the limiting factors on length.

I decided to try angling this side face, which was an idea from TestCut.

I thought it looked a lot fresher instead of that waste-o-space bezel.

Next, I added the Tetrinsics, this time giving Finger 2, 4 and 5 a bit of an angle instead of parallel with the top face. I also added a coloured section to the bezel, which I think finally brings the design into "fresh look" territory. The look was the reason I got my current white smartwatch in the first place.

I thought that the rounded edges of the Tetrinsic trenches weren't complementing the look, so I removed them:

I also remembered that Finger4 was hitting the edge of my previous concept, so I needed to rework this section also:

And that got me to this solution:

However...

I wasn't expecting it to look like I put an architectural model on my wrist.

It looks kinda ok from the front though.

I think the seperated bottom section is a nice place to add a splash of colour, so I've coloured it the same milky green as the front.