The Lisa expansion card
Zero Insertion Force (ZIF) sockets have a yellow plastic alignment cam that can break.
The cam rotates out of the way of the slot when the ZIF socket is opened, allowing the expansion card to slide in. When the ZIF socket is closed, the cam rotates back to close the end of the slot.
If the expansion card is protruding (and so the contacts are not properly aligned), the cam jams against the card, preventing the ZIF socket from closing and alerting the attentive operator to the problem. The spring forces on the ZIF socket actuator are large enough that the stress from a cam jam can result in breaking the cam.
Once the cam is broken, expansion cards can be misaligned in the socket and damage to the expansion card or another part of the Lisa may result if powered on. See
Sun Remarketing SCSI card: A cautionary tale for an example.
Since my particular use sometimes involves frequent and repetitive operation of the ZIF sockets, I wanted a more durable replacement cam; when metal
Selective Laser Sintering (SLS) 3D Printing became more affordable, I thought it may be a good technology for this purpose.
I modeled the cam with the small modification of a chamfer on the edge that first engages the card, which seems to help to nudge a slightly misaligned card into place rather than jamming the cam. (Denoted by the blue line in this image:
https://lisalist2.com/index.php?action=dlattach;topic=573.0;attach=813;image)
My first experiment was to reproduce the cam design using SLS stainless steel via jlcpcb.com
The part was dimensionally accurate in general, but required some clean-up to fit properly. The steel alloy is quite hard, making clean-up a bit laborious using needle files and small carbide and diamond Dremel size bits.
For this and subsequent SLS parts, the clean-up required clearing out the rectangular hole for the sliding metal handle/actuator, and clearing out the thin slot in the cam where it fits over the fin on the end of the ZIF socket body.
I concluded that a softer material than the stainless steel alloy could be more appropriate for easier clean-up.
My second experiment was to reproduce the cam design using SLS glass fibre reinforced Nylon via pcbway.com
The part was dimensionally accurate, but the part had protruding fibres making it a bit furry. Clean-up was easy with a sharp knife, but the health hazard of the resulting loose fibres is a concern.
I noticed that there was quite a bit of flex to the nylon part during operation of the ZIF, and realized the design of the cam includes a feature that provides a detent when the ZIF is fully closed. (Denoted by the pink ovals in this image:
https://lisalist2.com/index.php?action=dlattach;topic=573.0;attach=811;image) As the cam is rotated from the closed position, this detent feature flexes the cam. The nylon is quite a bit more flexible than the original yellow plastic so the effect is more obvious.
I suspect this flexing is part of the cause of cam stress and eventual breakage, although closing the ZIF with the expansion card not fully inserted is probably the main cause (along with aging).
When testing the stainless steel part, there was not enough flex to actually operate the ZIF, so I had removed the detent bump from its cam.
With the detent removed, the spring forces in the socket are adequate to provide confirmation that the socket is closed, so eliminating the detent feature does not seem detrimental to me.
Even though the nylon seems suitably strong, the loose fibre nuisance is a concern, and after observing its flexibility, I suspect a metal part will prove more durable and provide a longer lifespan in my type of usage.
My third experiment was to reproduce the cam design without the detent bump using SLS aluminum alloy via pcbway.com
The SLS aluminum parts were the same price as the SLS glass fibre nylon and SLS stainless steel.
The aluminum parts were generally dimensionally accurate, but the thin flange that engages a slot in the ZIF body was more distorted/warped on the SLS aluminum parts than it was on the stainless steel and nylon parts. Having ordered only these 3 SLS prints, I can't say if that is a characteristic of the SLS aluminum, the fabricator, part orientation, random chance, or something else. The fabricators do warn that part details as thin as this may be distorted.
Clean-up of the softer aluminum parts was easier than the stainless steel, but in a way more complex due to the warped flange. I ended up using a machine tool to clean out the thin slot in the cam as attempting to do so manually was not resulting in parallel sides.
Given the expense of metal SLS parts and the labour of clean-up, I suspect that for many and perhaps nearly all Lisa owners, a more economical
Fused Deposition Modeling (FDM) 3D printed part will be sufficiently durable for infrequent use (and cheap enough to have numerous spares on hand in case one breaks).
My fourth experiment was to reproduce the cam design without the detent bump on my own 3D printer, with a design modification so it can be FDM printed without supports. The "no supports" modification also makes the cam a bit stronger in the place that it typically breaks, so is useful for SLS too (although it does make the slot a bit longer so clean-up may be increased a bit).
I printed a couple of those in PLA (an inexpensive type of plastic commonly used for 3D printing), and they worked fine without cleanup. I suppose one could use something stronger than PLA or anneal the PLA. Experience is necessary to see if that is required or desired. I suspect not for infrequent use, especially if one has spares on hand.
For frequent use, I think the metal part is still my preference.
Below are files for 3D printing:
1) recommended: The cam without the detent bump and with modification for 3D printing without supports (suitable for FDM and SLS printing)
2) not recommended: The cam in it's original configuration (with detent bump and without modification for 3D printing without supports)
These were generated from FreeCAD files; if you would like the FreeCAD files to make further modifications, please contact me to discuss.
Author
Topic: Replacing a broken yellow cam on a Lisa expansion card ZIF socket with 3D print (Read 2189 times)