Model MF - GB CLOSED

[Techno Trousers]

Everything's coming together so nicely!

[E TwentyNine]

Sheet metal caps arrived... They look good. Not too rigid, not too soft. Just right.

I'm wondering if I can find something like pictured below that will plug the space and not muck with the board. Maybe something low profile metal over rubber/plastic.

[lot_lizard]

I'm wondering if I can find something like pictured below that will plug the space and not muck with the board. Maybe something low profile metal over rubber/plastic.

These will typically always be metal, and will certainly scratch the powder coating a bit when you drive them in, but as long as you weren't taking them in and out often (unlikely)... it would certainly be a sound approach. The only bit you would be missing would be to cover the holes for the XT style barrel lock, and portions of the AT. I doubt too much debris gets through those tiny holes though.


The only other trick would be to make sure they were 12mm. 7/16" and 1/2" will not work, and 15/32" would be hard to find I would think. Normally these little jewels are in inches unless you have them shipped from overseas.

If you really wanted something to fit it gloriously, I would just mock up a mesh and then have it printed in some plastic that is mildly elastic. Then you could even paint it if you preferred to match the plate.

[DMA]

Part of the trick with the controller grounding will be making sure I can find screws with the correct depth (so they don't come up through the back plate and contact the PCB).

Coutnersunk screws looking outwards? Like those you have for PCB-to-backplate grounds, only going thru back plate from under the PCB?
|>=||≡|≡=

Where > is bolt head, = is bolt shaft, | is PCB, || is metal plate, ≡ is the nut.
Proooobably will be hard to assemble though (slip on the connector, then catch those bolts with bolt holes..). But the nut between controller and backplate will help with separation (it's there chiefly so that screw doesn't slip when you tighten the nut on the controller side). Experiment is needed. It may happen to be a really stupid idea. f.e. because PCB tolerances + drilling tolerances + connector soldering tolerances will prevent assembly on some units (because I don't see any oval holes anywhere so I assume there's no room for adjustment).

Don't like the idea of notches under those solder joints. You'll easily have 1-2mm positioning uncertainty there. Also the aestetics of having notches on the back plate is questionable

PS: nuts probably won't eliminate need for isolation layer. Better safe than shorting VBUS and ground. Those fuses in motherboard are theoretically auto-resettable, but..

[lot_lizard]

Don't like the idea of notches under those solder joints. You'll easily have 1-2mm positioning uncertainty there. Also the aestetics of having notches on the backplate is questionable

I agree... I don't like the cutout either (hideous and slightly less solid), plus it would lock the header pin positioning into future controller layouts.


So to describe what we are going to test in the final prototype better, I put the following (crude but to scale) diagram together help see it a bit better. The layers are:

• PCB (top green)
• PCB insulating/protective layer (light grey)
• Steel backplate (darker grey)
• Header pin lift of the controller off the backplate (white)
• Controller insulating layer, which would cover the header pin protrusions as well (black)
• Controller (green)

In grounding the controller, m3x4mm pan head screws would have ~1mm bite into the backplate (plenty without passing through) if we have a ~1.5mm riser (lime green). The same screws would be used to ground the PCB to the backplate. The PCB ground screws will pass through the backplate ~1.8mm, but that doesn't bother me at all aesthetically. We want them to pass through some to achieve as much thread contact as possible.

To be honest, this securing/grounding of the controller to the backplate is unnecessary altogether really since we are grounding the PCB, but I am going to have the final prototype made with the needed backplate taps just to test it. If it is amazing, we will go with it... if not, the controller will only be secured by the triomate slip-on connectors (which is quite sturdy). The "riser" (lime green) would be a loose part similar to a washer, and would not be permanently attached to the backplate.

groundScrews.png (4.6 KiB) Viewed 4728 times

Appreciate the thoughts... your opinions echo mine.

[chzel]

M3 has a thread pitch of .5mm and considering that the first/second threads are usually not fully formed, I'd be surprised if you managed to grab more than half a thread there. And even before that, how are you going to cut the threads on the plate?

Happy New Year!!!

Moment of clarity: I just realised that it doesn't have to be a blind hole...

[lot_lizard]

M3 has a thread pitch of .5mm and considering that the first/second threads are usually not fully formed, I'd be surprised if you managed to grab more than half a thread there. And even before that, how are you going to cut the threads on the plate?

The threads will be cut into the plate using a traditional taper tap. For the m3x4mm screw, I had trouble finding them, but these screws have no "pitch diameter" (think flat tail). Normally a screw is made of three diameters (major, minor, and pitch). The pitch diameter only comes into play on the final thread as you mentioned where the tail tapers off. In this case, it wouldn't taper (the thread just dies flat). It does make it a bit tricky to thread properly when screwing in though. You would always want to counter-clockwise rotate softly until you hear/feel the "click", and then clockwise rotate.


Foreshadowing: I think we nix the controller ground screws in the end the more I have the working prototype in front of me, but want to know for certain. Having standoff's welded as the "riser" (lime green from before) is just a wasted expense that we really don't have a margin for.

And... Happy New Year to you too. Always reminds me of one of my favorite Trading Places moments:

[DMA]

In grounding the controller, m3x4mm pan head screws would have ~1mm bite into the backplate (plenty without passing through)

And then some extra-strong user will just strip the thread in that blind hole.
And it would be game over, because you can't replace the bolt which isn't there.

But, since you mention that you won't be doing those controller grounding screws - this all is theoretical.

[lot_lizard]

Exactly... In fact, it's officially off the table. Pulling on this in tests right now, if you manage to wreck this triomate only mount, you need to find a new hobby (like working for the airlines handling luggage ) Along with pan heads securing the PCB ground on the 2 unneeded tensions, I'll make all 4 grounded, with the 2 tensions using short standoffs. The standoff will act as the pan head in the previous diagram.

[lot_lizard]

Please only vote if you participated in the GB...

I created a poll for the MF GB to determine the screw drive head type you want to have. I have always assumed Phillips because it is more common, but the hex is more attractive in my opinion. Please vote in the poll if you would in the coming day or two before I make our order to Fastenal. Again, these are stainless steel, countersunk, and used to fasten the top and bottom assembly plates together.

Note we have 2 screw sizes (m2.5 and m3). So we potentially need two unique allen wrenches supplied if hex is chosen (I would have to verify).

Phillips example


Hex example

[ramnes]

Why do you say that the hex is more attractive?

[lot_lizard]

Why do you say that the hex is more attractive?

I'll correct the previous post with an "imo" (just personally opinion). Both will look nice, and I would be very pleased with either. I think hex looks better, but will not be as easy to deal with for the general person without either keeping up with supplied wrenches or having some on hand. My vote would be for the Phillips given the ease of working with, but I realized I never even got the opinions of others. Note that hex in small allen wrench sizes has a tendency to strip the wrench itself if any real torque is given.

[chzel]

Aesthetically allen (hex) is nicer imo, but a countersunk M2.5 will need a 1.5mm allen key which is a bit too fiddly and sensitive to torque (M2.5 needs 1.5mm key and m3 needs 2mm). I vote Philips at least for the m2.5.

[Wodan]

Any chance we can get Torx?

I loooove TORX ... it's the love child of PZ/PH and HEX ...

Other than that, I agree that Hex has the better looks ... but I'm super torn between the two.

[lot_lizard]

Any chance we can get Torx?

I loooove TORX ... it's the love child of PZ/PH and HEX ...

Other than that, I agree that Hex has the better looks ... but I'm super torn between the two.

I can check... I agree it is a nicer bit. The only issue would be sourcing some wrenches to include. Some of us will have them on hand, but I doubt many (especially the m2.5). The m2.5 is a requirement unfortunately because of our edge tolerances. I do like the idea of making these easy to work on in the future without having to keep up with specialty tools (for the average person).

[fohat]

I have been pleased that this project has kept a "keep it simple" approach to most of the assembly process, since the physical difficulty of working with original IBM hardware intimidates many people.

Everyone has a Philips screwdriver and it is universally available. If someone wants to buy a set of specialized replacements that they feel is more attractive, they are always free to do that. $5 on ebay will get you a spoonful of just about any screws that you want.

[chzel]

Torx has indeed a very nice drive geometry and the M2.5mm would be T8, so not too small, but definitely more expensive (both screws and driver)

Edit: fohat makes a very fair point. More seasoned owners can very easily change the screws.

[Wodan]

hmhm ... yeah well if you want to include the tools, i think TORX will be a pain to source!

I guess PH is the best option in any way. I have a lovely PH1 driver that I use daily ... that head is just super common and proven.

[Darkshado]

I prefer hex or Torx, I've had to deal with too many stripped Phillips heads. Robertson (i.e. square) is ok too, but seems non existent stateside, plus I don't think they come in small metric sizes.

[matt3o]

it really makes very little difference to me... but if I really had to pick one I'd say torx and hex heads are very nice

[derzemel]

Torx and hex are the ones for me too

[Ir0n]

I voted hex lol

[Khers]

I voted hex for purely aesthetic reasons. Torx would be even nicer, but I can understand the rationale for going Phillips.

[wcass]

Silly me, i was not subscribed to this thread.

About grounding the controller to back plate, i was thinking of just a short wire with ring connectors on both sides. One side of the wire connected to the back plate by one of the existing bolts (with nut). The other side connected to a nut-bolt on the controller. The controller's nut or bolt head provides stand-off from the back plate. And yes, some conformal material on the back of the through hole pins. For testing, just use a blob of hot glue; when you want to remove it - paint it with rubbing alcohol and it should peel off easily.

Also, i had thought to use only 6 header pins; the 4 pins to USB were to be wires to 4 pin female Dupont (ideal wire colors would be red, black, white, green).

[ohaimark]

I'd prefer Torx, but I'll vote Phillips.

[arkanoid]

I voted Philips for simplicity, but I would have voted Torx with security bits if it was in the option, so nobody can touch my MF.

[lot_lizard]

About grounding the controller to back plate, i was thinking of just a short wire with ring connectors on both sides. One side of the wire connected to the back plate by one of the existing bolts (with nut). The other side connected to a nut-bolt on the controller.

So I have new designs for both plates where we are going to ground the PCB to the backplate at 3 points for the SSK and 4 for the full-size. Do you even think we need to worry about grounding the controller separately? The reason I ask, I am getting rid of the need for bolts/nuts at all on the top ridge now. If you think we need one, I will add it back.


The ground of the PCB will be with standoffs that can still be used as tension bolts if needed (like below)

61H0dQb59kL._SL1100_.jpg (236.45 KiB) Viewed 4528 times

[Phenix]

voted for hex, but hex torx would be even nicer

[livingspeedbump]

people actually voted Philips?!

[matt3o]

people actually voted Philips?!

I know right? In what world do we live in?!

I guess they are just more practical...