Closer To The Final Replacement Of Stick

[Radek]

yeah 50 are a bit much.. this is not a easy mod and many wont do this to their gp2x... atleast not on their own.
anyway i hope you find a way to produce sone a first run 10 or something like that.
i really want to try it.
cant wait to kill my gp2x

​

I will able to make few units anyway so don't worry. Just was looking around as I'd strongly prefer to make PCBs with solder mask (it's more reliable, looks nicer and is easier to solder) as I can not make such ones at home.

As many cunning people have suggested; drill out the area where the dpad will hit the plastic; then
carefully sand it down.

Its labour intensive, slow, and prone to leaving jagged edges. Unless someone makes a custom moulding:

(ahem ....DaveC... )

Take the Dpad '+' design and flatten the underside so its curve matches the 'top' of the GP2X joystick depression.

Drill 4 small holes into the top panel of the GP2X where the 'center' of the pad triangles would be (on a Gamecube/SNES/NES type pad).

The custom Dpad will have 4 small 'pins' that poke thru those holes and allow a microswitch to engage.

The center on the dpad is supported with a rubber column or spring. (or something)

Another solution would be complete change of GP2x's case...
It'd be very doable as the GP2x's mainboard is very simple (and it's good thing) design. Should not be hard to make but unfortunately this is beyond my abilities. But perhaps there would be some market for different "skins" for GP2x?

The only thing I don't know how to figure out is what the Dpad PCB rests on. (if not annother platform held up by metal pins from the old joystick connectors.)

That's actually quite easy part - after removing the stock stick (it can be just cutted off the mainboard and should to preserve its pins at back side) there will be plenty space to mount your own small PCB board to support mechanically new controller.

Microswitches are being used in this case because I think they will give a more positive feedback to the user AND work better (mushy carbon pads dont work well with sharp plastic columns poking into them). Oh, and a single external switch added to the outside of the case. (drill a hole for one of those small momentary push buttons, near the top right perhaps)

I'd prefer microswitches too but membrane ones can work good also.

I think it could really work, now. Keep thinking and working guys / gals. We are getting closer...

​

Just found some other parts needed... not ideal but not expensive either. They will be very helpfull...

Well I've discovered something that could really help with the custom plastic Dpads:

http://www.shapelock.com/index.html

A plastic that can be easily worked with only hot water.

Check it out!

​

Has been anyone using it? How really is it good?
It looks very atractive and could be very handy.

I wonder, if the microswitches inside a mouse might be good for this? I have few old ones...

​

You might experiment but plain microswitches are easy to buy.

EDIT:

Just found an N-Gage keypad for not that bad price. Are these serviceable parts with actual switches or not (they can works by themselves or there is need to make a PCB with contact pads?)

 

[scorpio]

Hi,

For PCBs you might want to check this crowd out. I haven't tried them myself yet, but I put them into my bookmarks last year after seeing some favourable mentions.

http://www.pcb-pool.de/

Also, it looks like the shapelock plastic is just a rebranded "Polymorph" that Maplins (in the UK and Ireland) sell. It's good stuff.

It's quite strong when you have a decent-sized lump of it, but when it's thin ( < 0.5mm) then it does get quite flexible which can be a bit of a drawback. On the plus side of that, it can be stretched and flattened while soft if you need to make cable quides etc.

It's great for prototyping parts or making custom mounts. You can use it to copy parts as well: If you make a warm lump of it and press it onto the part you want to copy (i.e. a d-pad) until it cools, then you can use that as a mould for casting a copy, either from resin (you might need to use silicone spray) or more plastic. This is within shape limitations of course, but there are all sorts of methods to get a full mould.

It cuts quite nicely with something like a Stanley knife, if you need to do further shaping on it.

I originally bought it with the intention of prototyping a project, and I've ended up using it for all sorts of other things.

 

[scorpio]

Hi,

Turning the current stick into a 4-switch would be difficult, you will see when you look at it. The best way if you are going to use the current stick and not replace with a pad would be to rotate the stick 45° to eliminate the diagonal bias, then shim the "hat" closer to the contacts to reduce dead zone. A metal spacer could be placed between the hat and the top metal cover.

​

I had a look at the internals of one of them today and I have a few ideas, but I'll have to try them out and see what works best. One of the things I'm going to try doing is to put a thin layer of solder on the circular edge of the contactor and file it back down to more of an octagonal shape to match the outside, just to see how that behaves. This should also reduce the throw length by a noticeable margin because of the extra metal, and I would guess with ROHS-compliant solder, it should wear pretty well (if it works!)

From looking at the way the contacts are arranged, and from looking at a (pretty shaky) picture of the GP2X circuit board I found on one of the forums, it looks like it might not be possible to just rotate the unit. :-(

Also, correct me if I'm wrong, but from the picture it looks like they (GPH) have actually mounted it on the PCB at a slight angle?

 

[bacteria]

This is what I love about these boards - the mountain of knowledge, and people wanting to help!

 

[bacteria]

The link to Maplin's PolyMorph is:
http://www.maplin.co.uk/Search.aspx?criter...=15m3&source=15

 

[Radek]

The link to Maplin's PolyMorph is:
http://www.maplin.co.uk/Search.aspx?criter...=15m3&source=15

​

Excellent!

I also found source of least expensive N-GAGE (classic and QD) pads finally! What are you thinking about first N-Gage pad? They are very cheap now but do look decent.

Could someone provide me with pictures of backsides of these pads?
(and other info like exact sizes)

Soon I should have more parts to play with. I will able to think about drilling gp2x's case then.

 

[DaveC]

Hi at

Turning the current stick into a 4-switch would be difficult, you will see when you look at it.  The best way if you are going to use the current stick and not replace with a pad would be to rotate the stick 45° to eliminate the diagonal bias, then shim the "hat" closer to the contacts to reduce dead zone.  A metal spacer could be placed between the hat and the top metal cover.

​

I had a look at the internals of one of them today and I have a few ideas, but I'll have to try them out and see what works best. One of the things I'm going to try doing is to put a thin layer of solder on the circular edge of the contactor and file it back down to more of an octagonal shape to match the outside, just to see how that behaves. This should also reduce the throw length by a noticeable margin because of the extra metal, and I would guess with ROHS-compliant solder, it should wear pretty well (if it works!)

From looking at the way the contacts are arranged, and from looking at a (pretty shaky) picture of the GP2X circuit board I found on one of the forums, it looks like it might not be possible to just rotate the unit. :-(

Also, correct me if I'm wrong, but from the picture it looks like they (GPH) have actually mounted it on the PCB at a slight angle?

​

I was originally thinking of soldering some very thin pieces of sheet metal to the ordinal contacts. That would solve two problems. It would increase ordinal bias, plus the thickness of the metal would reduce the dead zone as well.

 

[scorpio]

Hi,

I was originally thinking of soldering some very thin pieces of sheet metal to the ordinal contacts. That would solve two problems. It would increase ordinal bias, plus the thickness of the metal would reduce the dead zone as well.

​

I know where you're coming from all right. I was trying to figure out the exact mechanical operation of the joystick, and it's more complicated than I first thought, as it basically involves the stick driving a lever, with the lever having two sliding fulcrums. The bottom fulcrum drives the contactor down onto the appropriate contact at the base. The spring around the outside of the contactor acts as the opposing force to ensure that only part of the contactor touches down. (You know all this already, I'm sure, but hopefully this will benefit anyone else reading who might be wondering...)

I drew myself a diagram to try and clarify it (this is just something knocked up in DPaint, so it's somewhat impressionistic.)

I'm looking at it from the point of it being a single-axis switch, with a 360 degree range of rotation about the y-axis. (For simplicity's sake, I'm also ignoring the switch in the centre of the base, which also acts as a restraint to stop the stick pushing the contactor all the way down.)

From looking at it pretty closely, I have a feeling there is going to be very little room in there for shims etc., which is why I've given up the idea of a slotted film to overlay on the contacts, as any (non-specialist) film thin enough to allow smooth transitions between directions is not going to last long enough to be realistic in implementation.

I realised that if we go too far in reducing the diagonal bias, we may end up with a lot of dead areas between each of the 8 directions, which could be just as frustrating. I have some ideas involving 8 sprung ball bearings mounted horizontally through the sides, but this is just fantasy at the moment, and probably completely unworkable.

I'm hoping that using solder to build up the contact points and shape them will allow me to fine tune it and find the best shape. Logically, an octagonal shape would be best, but we know how these things usually work out, so...


I'd forgotten that over here, we have Friday off this week. Perhaps I'll have some time to experiment after all!

 

[Radek]

Another small update:

I just shipped for these pads:

Got them dirt cheap and they are serviceable parts for classic Nokia N-Gage. Not best perhaps but at their price it was hard to resist. Get 5 of them and paid the same amount of money just for postal service.

Here I have something better:

This dpad is from N-Gage QD and people were raving about it here so I will try to get it (it's actually on auction and I'm hoping to get it at reasonable price).

I was also evaulating different chips for my logic adaptor but... funnily these old cmos 4000 series chips what I'm using are really the best from almost every standpoint.
(work asynchronous, have very little power drain counted at µW and are small enough even in DIP version).

And thanks about the polymorph info - it might be very handy.

I have to say... I'm starting to be more confident about this project.

 

[Trip]

I was going to mention Maplin's PolyMorph, but was beaten to it .
I did stumle across this stuff the other day though if it helps.. PlastiDip
Not sure if it will hold up well enough for coating the actual pads, but it might come in handy for something else.

Nice work so far by the way

 

[Iorgy77]

Are you going to fit an extra button to the gp2x for the joy button?

Very impressive progress so far.

 

[PSyMastR]

Are you going to fit an extra button to the gp2x for the joy button?

Very impressive progress so far.

​

The ngage original (dunno about the qd) had a click button.

 

[DaveC]

It would be nice to be able to find discrete D-pad assemblies somewhere. Cannibalizing various devices is not a reliable or many times economical long term solution. I wonder if there are replacement parts available for something to be able to order in quantity?

 

[grahf]

Those N-gage QD pads look like what the doctor ordered. All one piece, and "built in" carbon pads. While its true what Dave is saying, its not good to canibalize.. But the N-gages were made with swappable faceplates in mind so parts are very easy to get. Im sure there are more keypads aroudn then Gp2xs sold so far

 

[Epicenter]

The day we can easily replace the stick on our '2xes with a replacement D-Pad, (or better yet for those of us with less than the best soldering skills in the business, mail it in to be replaced).. will be the day the '2x ascends from great portable game system to tiny god. Keep up the excellent work.

 

[DaveC]

Hi at

I was originally thinking of soldering some very thin pieces of sheet metal to the ordinal contacts.  That would solve two problems. It would increase ordinal bias, plus the thickness of the metal would reduce the dead zone as well.

​

From looking at it pretty closely, I have a feeling there is going to be very little room in there for shims etc., which is why I've given up the idea of a slotted film to overlay on the contacts, as any (non-specialist) film thin enough to allow smooth transitions between directions is not going to last long enough to be realistic in implementation.

I realised that if we go too far in reducing the diagonal bias, we may end up with a lot of dead areas between each of the 8 directions, which could be just as frustrating. I have some ideas involving 8 sprung ball bearings mounted horizontally through the sides, but this is just fantasy at the moment, and probably completely unworkable.

​

Actually there IS room for shims to be soldered on the smaller ordinal contacts. They just have to be thin enough. I don't know if you misunderstood but I am not talking about insulating film, I am talking about conducting metal. If a thin sheet was soldered to the ordinal contacts it would raise the targets higher than the diagonals. Because the hat is circular when tipped it would come into contact with the raised ordinals easier then the diagonals. This would also reduce the dead zone because you wouldn't have to move the stick as far to make contact because the contacts are higher. There would be no dead areas between the 8 areas either as the raised ordinals would just cause a ordinal and diagonal to be pressed at the same time when pressed diagonal which would give SOME kind of input.

I was even thinking of having these contacts laser cut into the right shapes so that the shims could be mass produced and mass mods of the stick could be made. A bunch of the alps sticks could be bought, modded and then sold for a direct replacement. I was going to experiment but haven't found the time and I am waiting to see what turns up here too. Replacing the stick with Radek's chip would be great. I do think it may be a bit more complex for alot of users though as it involves alot of modding to the case itself. If a direct replacement with a modded standard stick was available it may be a good alternative for those that aren't up to major surgery on the GP2X case.

 

[Radek]

Are you going to fit an extra button to the gp2x for the joy button?

Very impressive progress so far.

​

Eventually. That would be easy to do but stick's button isn't used too often anyway.

It would be nice to be able to find discrete D-pad assemblies somewhere. Cannibalizing various devices is not a reliable or many times economical long term solution. I wonder if there are replacement parts available for something to be able to order in quantity?

​

I didn't get these from actual n-gages! It'd be terribly costly to do so. They are serviceable parts and they can be bought alone.

The day we can easily replace the stick on our '2xes with a replacement D-Pad, (or better yet for those of us with less than the best soldering skills in the business, mail it in to be replaced).. will be the day the '2x ascends from great portable game system to tiny god. Keep up the excellent work.

​

Don't worry about soldering as it isn't necessary. Open your GP2x (it's easy to do) and look at the backside of GP2x's mainboard. You will see that stock stick's pins are standing high enough to connect wires to them without resorting to soldering. I wasn't soldering in GP2x anything yet and I don't think it will be needed.

The major obstacle is drilling of the front gp2x's panels to fit a dpad here.

Even the stock stick's can be removed without any soldering too. It can be just cut off from mainboard. It's very important to save the already soldered stick's pins in the place.

I was even thinking of having these contacts laser cut into the right shapes so that the shims could be mass produced and mass mods of the stick could be made. A bunch of the alps sticks could be bought, modded and then sold for a direct replacement. I was going to experiment but haven't found the time and I am waiting to see what turns up here too. Replacing the stick with Radek's chip would be great. I do think it may be a bit more complex for alot of users though as it involves alot of modding to the case itself. If a direct replacement with a modded standard stick was available it may be a good alternative for those that aren't up to major surgery on the GP2X case.

​

If you could fit four microswitches inside the stock controller's bo and make them react to the movement of the stick it wouldn't need any resoldering and would save a gp2x's front panel from any drilling. And you would get full benefit of 4-switches accuracy.

Imho even small modification to the gp2x's stick will need opening it so it'd be better to think about more drastic measures and completely change what is inside it than just try to tweak it.

I will open that stick and see how it looks inside myself.

edi: Well, well...

How did you open the stick on the GP2x's mainboard? I can not without cutting off few things:

Are these connection needed? If so then we might forget about tweaking build in stick. (try to figure that later today)

BTW - how much free space is inside GP2x?

As you see there is lots of space - GP2x is very modding friendly.

 

[scorpio]

I've just had a crazy idea, and I'm going to ask my friend's permission for me to open his GP2X so I can verify it. From the pictures just posted it looks like there's loads of room in there.

If you were to put a disc over the shaft of the joystick, so that it rests on the flat top of the joystick enclosure, and be held flat when case is put back together, and have the disc a few mm wider than the stick unit, you could mount 4 sub-miniature microswitches on the end of each axis, which would be activated by the edge of the disc as it gets pushed around by the joystick.

I'll sketch something up when I get home tonight and see if it's plausible.

If it looks workable, then it would only require some snipping of the joystick connections from the PCB, and Radek's circuit to recreate the necessary signals.

Could be really quick, and not too drastic...

** EDIT: No, there wouldn't be enough space at the left edge of the unit to fit a switch in, so it looks like it'll have to be a complete replacement.

PS. I think those connections you've labelled are the other end of the joystick circuit. The voltage comes up those connections, and then is transferred back down through the contactor to the contacts in the base. I haven't wired up my ones yet, so I don't know the exact specifics of the wiring needed.

 

[Radek]

Dissected it...

And putted it back again:

Didn't work then but I'm not surprissed. The connection what were providing voltage were cut off. Will experiment with it soon. But at least I'm certain the bottom part of stock stick can be used as socket for mounting the replacement. Eventually its pads the could be used to transfer signals dramatically reducing needed wiring. A very smart design would have a base pcb hosting logic chips and a dpad itself and there would be only two wires to the batteries.

edit: more research on the "stick"

First I fixed it after being dissected (thought it was quick hack):

Providing the voltages from appropriate pins:

Works now but needs to be pushed much harder to register. Probably due to loss of some of kind of lubricant what was inside it before dissecting.
(another great thing about it - controller with lubricant...)

But this isn't most interesting thing about it... Look here:

Looks like it is an insulator but why?!

Actually I had to provide voltage for both sides seperately because of it. Bizarre?

Enough for me of it... I will now focus solely on making replacement. This joystick is weird with its lubricant, not equal pads, diagonal bias and who knows what else!

 

[scorpio]

Looks like it is an insulator but why?!

Actually I had to provide voltage for both sides seperately because of it. Bizarre?

Enough for me of it... I will now focus solely on making replacement. This joystick is weird with its lubricant, not equal pads, diagonal bias and who knows what else!

​

I had forgotten about that myself, I don't understand why they did that, but there must be a reason to complicate the design like that.

Go here, ALPS web catalogue and enter RKJXL into the search box and you'll be able to find a PDF which will show you the exact details of the connections.

I would have linked directly, but it uses session-based URLs, which will have timed out by them time you're reading this.

My last experiment on the joystick didn't turn out a success, so I'm going to give it one more try with the second one I got, but since my thoughts the other evening, I'm leaning towards a joystick replacement now, rather than a d-pad.

I think you're on the right track with your ideas, but rather than duplicate the effort, I think I'll concentrate on a microswitched alternative, and between us we should be able to come up with some useful options.

BTW, what camera are you using to take your pictures? They're very clear, especially compared to some of the rubbish macro shots my cheap camera takes.