Hardware review: few observations


#1

Hi! I think creating hardware is a wonderfully difficult process to get it right (and I have no experience besides building some hobby stuff). So first, I really like and admire the work behind the current state of the device and this is just to open some discussions about future possible improvements.

Here are my observations about the hardware:

1. D-PAD

It feels weird. If behind there are 4 individual buttons I would rather prefer 4 individual directional buttons.

2. Scratches

There are small scratches around the sides

3. Micro-SD

The slot is too wide and the micro sd does not gets flush when inserted

4. Micro-USB

The slot is too wide (why the original design included the secondary micro USB?) and too near the pokitto’s leg (so, it is impossible to plug bulky ended cables)

5. Screen clamp

The mechanism for holding the screen looks flimsy and botched as a quick solution, in my two cases the little clip/clamp does not sit flush to the screen

6. Power switch

The switch cover can be a bit wider so the switch cover does not move diagonally. Also, I am not sure if OFF should go on top.

7. Back buttons

The rubber buttons do not fit properly. They are not flush to the case and do not look symmetric (centered inside the hole). And why there is no icons?

8. Embossed rectangle on the back

What is that? :stuck_out_tongue:


LCD holder replacement
#2

Thanks for the analysis, now some answers

I respectfully disagree. This allows better diagonal (UP/DOWN+L/R simultaneously) movement. Theres a reason Nintendo uses this arrangement. The benefit will come our when games using this begin to appear. I also have a Gamebuino and 4 directional buttons IMO is not good.

Intentional artifact caused by a so-called “natural” mold (no sideways actions). Nintendo or Sony would use sideways moving action in the molds. I couldn’t afford such things as they make the mold many times more expensive. Sell 100k Pokittos for me and I will remove those “scratches”.

They’re not scratches. They’re how plastic is made :wink:

I didn’t want it to go flush. Space around MicroSD cradle left intentionally to allow us to use 1 of 3 different component suppliers

Because… this is not the last Pokitto you will see. Call it Pokitto Pro or Pokitto Plus… its there for a reason. As for being too close to the leg, when we still had the silicone buttons, the silicone pac came to within tenths of mm of the USB connector component inside the case.

Use a smaller USB plug or take a knife and cut the cable end :smile:

Again, if natural mold is used, the only way to introduce “clips” would be to have sideways moving action in the mold. This was the only way to avoid high cost OR having a “hole” in the face of the Pokitto. Couldn’t be avoided. If you want, I can give a detailed explanation. But basically: mold cost.

Agreed. Too late to change that now.

They are not flush to the case because it would be very difficult to press them if they were. I actually tested this particular thing very carefully and adjusted the cutting tool accordingly.

As for the buttons: did you notice they are from Gameboy Color? Not making my own silicone mold I saved 700-1000USD. Theres some variation in the buttons I get from China.

Was intending to, ran out of time. It introduces yet another step in manufacturing, and setting it up (either silk screening or tampo printing) would have set me back another 1-2 months. Will be added in the future. EDIT: This means there will be more info than just the icons. The extension header pinout will be printed for reference also.

Hardware debugging (look underneath). Its for crazy hackers to cut out and stick a SWD connector in.

EDIT: in case someone thinks I am being defensive. No, I do not mind this kind of conversation at all. I do this same at work every day. But many of the decisions were decisions, not mistakes, and I want to make that clear. I am very interested in all feedback so please keep it coming!


Pokitto parts?
#3

The only issues I have are not liking the type of buttons used, but I know the reasoning behind it and I’m more than happy that the reasons are good.
It would be nice if the micro sd slot wasn’t as tall, I have already dropped a card inside a couple of times. But then, most people wont be clicking the card in and out all of the time.

The d-pad bugs me a little. I have never liked it if you can press all 4 directions at once. Most dpads have a bump underneath to prevent people pressing up+down or left+right at the same time.

What would happen if someone added a second usb connector? Does something happen already?


#4

No. Its still a dead end. But had to be taken into account now, bc changing plastic is so much harder.

Its there for the future “pro” version, the timeline for launch is still (?). Pro will have, among other things, a HW debugger built in. The reason why its not in this first version is simply cost: it would add 12-25€ to the price. People would have been scared away. And most people wouldn’t know what to do with a HW debugger anyhow.


#5

Me included :slight_smile:


#6

now I need Pokitto Pro


#7

We’re here to change that. In a couple of years, there’ll hopefully be lots of people who will say “I got into embedded programming because of Pokitto” - which is the aim of the whole project.


#8

adding to the list is the cutout for the audio jack.
also would have been nice to have a way to lock the sd card like for example wen powering on the device the slider would partially cover the sd card slot so it could not accidentally be removed, would still be possible to do if the switch had more travel and reversed the on off

the wedge for the display is minimal would have like do see a screw on the top or like a V back plate for it since the top of the display can flex and potentially break the screen holder wen in a bag or something


#9

Thanks for the answers!

I think the problem is the tactile feedback. I understand how you can press two directions at once (and the Nintendo reason can be invalidated with Sony doing the opposite) but the problem is as @spinal says there is no center pivot, so the second click feels weaker (and uncertain). I guess you can get used to the d-pad. Need to play games in it.

I understand that. The problem is that the microsd can fall inside the device :smiley:

A taller piece can press against the pcb and keep the display from flexing when you touch it.

No I didn’t, that is f**king clever :sunglasses:

Nice too! but maybe the rectangle should be on the inside.

It is really cool to hear the technical details, and you do not sound defensive. I got in love with the magic of hardware while studying computer engineering (after one course), and my knowledge is quite limited so it is really interesting to hear the details.


#10

@eried
One of the reasons for preferring the Nintendo-style approach of a d-pad to the Sony-style approach of having 4 individual buttons is that it makes the programmer’s job easier.

If a programmer can reasonably assume that up & down cannot occur at the same time and that left & right cannot occur at the same time, it means they don’t have to worry about what their programs do when that happens.

Equally, some programmers might be annoyed because they wanted to add a special feature for those conditions.
It’s all a big balancing act.

(Personally from a programming point of view I prefer controllers with proper joysticks where the output is an analogue axis instead of an on/off, but that would be impractical on the Pokitto.)


@jonne

Annoying but understandable.
To be frank the screen securing mechanism is probably my biggest gripe of what I’ve seen so far.

I noticed but I thought you had just intentionally tried to make them like Gameboy buttons, I didn’t realise you had actually used the same mold.

Also I agree with the others, it didn’t come off as being defensive because you explained the reasons properly and rationally (and 80% of the reasons were ‘money’).
Hopefully the Pokitto will eventually attract enough people so we can afford the nice things.


I wonder how cheap it would be to make stickers for the icons?
Might serve as a temporary solution for the first batches.


#11

I also would like to have Pokitto Pro, supposing it fits to the same case. I have already somewhat suffered the limitations of USB-serial tracing on HW. If the device crashes on SW error, you get no information about the reason.

We could improve the crash situation handling, like this:

  • Print the error directly on screen, like the volume icon is drawn.
  • Or we could print the error to the USB-serial, or to a pre-created file in SD card
  • In HW error cases, use the infamous “blinking led codes” (Amiga :wink: http://www.lemonamiga.com/forum/viewtopic.php?t=13978)

#12

Depends what the error is.

I know some chips detect certain errors and have a specialised interrupt in the interrupt table to handle errors so it would be useful if the M0+ has that. (Especially for things like stack overflow if it can detect that.)

Some things like infinite loops would be trickier to detect without the programmer manually inserting debug statements.
Some kind of assert function would be good.


#13

I am not sure about that, but hey, who is a controller designer here :stuck_out_tongue: it is all based on personal opinions. The Sony-style one is actually like the Nintendo internally (bump in the middle do not allow opposite directions to be pushed at once). not four individual plastic buttons.

All my comments are based on the “travel” and clicky feeling the buttons have. I do like the clicky sound, but when you press LEFT for example and then you want to engage LEFT-UP diagonal, the UP clicky sound is muffled so it does not feel reliable for “hardcore” oldschool gaming.

I do not like the lever (third class) the piece does. If you push the screen it feels bouncy and may break the case. I need to design a new one to see if I can really make a better solution (supporting the screen against the PCB)

Considering the price of the case and buttons, likely is going to be cheaper to get a new back part of the case with the labels :stuck_out_tongue:


#14

If you can come up with an alternative that’s economically viable I’ll be right behind it.
Even having two of the slidey things seems like a better solution.

Personally I’d still prefer the d-pad, but that’s probably just because the GBA is my favourite handheld console :P.

Hrm, I will see in person when I get chance to collect and use mine.
There’s probably some DIY solutions available.

I dunno, I think printing stickers is probably cheaper than changing a plasting mold (from a mass-producing business perspective).
I could be wrong though, I don’t know much about mass production and the cost of it.


#15

The label wont be a change in the mold, but some step in post production for printing the labels. The pinout has too many details to be embossed “physically”


#16

preferably offset printed decal instead of a flimsy sticker


#17

put me in the ‘disagree’ camp regarding the Dpad - keep single unit please!
PS sony have never done 4 seperate buttons for a d-pad, they just hide the middle so it looks like it (to get around Nintendo’s patent of the ‘+’ pad at the time) . it still rocks around as a single plastic unit.

FWIW I like the current setup. the dpad might be a little tall, and the buttons slightly loose, but they have a nice tactile click response.


#18

My sticker suggestion was more of a solution for older units that won’t already have it engraved/printed on them.

@eried did mention that:

But I didn’t know it was because Nintendo patented the plus-style d-pad, that’s nice to know.


#19

Just tried my wii, wiiu and xbox 360 controllers

in all of them you can press the whole dpad down from the middle


#20

Is the hw debugger compatible with a stabdard 20-pin, 0.1" target ribbon cable ? Does not look so just looking to the image posted.